aresty rutgers undergraduate research journal, volume i, issue ii corrections ✵ in the “foreword to first issue of the aresty undergraduate research journal”, frederric kelada’s name was misspelled. andrew dietrich was a contributing author to the article, “chronic corticosterone shifts effort-related choice behavior in y-maze” in vol. 1, issue 1. adriana scanteianu and xiangyue wang were listed in “about the authors” for vol. 1, issue 1. their paper appears in this issue. aresty rutgers undergraduate research journal, volume i, issue iii foreword to volume i, issue iii aresty rutgers undergraduate research journal ✵ thank you for reading the third issue of the aresty rutgers undergraduate research journal! i hope you enjoy exploring the variety and richness of the undergraduate research happening at rutgers, from the analysis of 19th-century female writers to the binding of the vitamin d receptor, and everything in between. the aresty rurj was established to uplift, connect, and showcase the incredible undergraduate research community at rutgers. undergraduate students are vital to the backbone of the research happening at this university and we are excited to provide a platform for students to share their work with the rutgers community and beyond. this issue, and the past year and a half, has been undeniably marked by the covid19 pandemic. our authors and reviewers have gone above and beyond to continue their incredible work with our journal even throughout the challenges they faced during the lockdown and pandemic. our authors had to adapt to conducting remote research and many wrote their manuscripts based on combinations of remote and in-person research findings. our undergraduate reviewers participated in the fully virtual reviewer program and put their skills to work with their outstanding reviews on submitted manuscripts. our graduate student and faculty reviewers were remarkably flexible and generous with their time and were instrumental in the growth of our authors and their manuscripts. i would also like to acknowledge the faculty advisors and principal investigators who served as mentors for our authors in their journey in research. please continue to encourage your students to share their hard work and support them in their endeavors! finally, this issue would not have been possible without the key contributions and fantastic work by the editorial team (including the reviewer program director, managing editor, creative director, and chief copy editor), copy editors, aresty research center administrators, and the rutgers libraries team. thank you everyone for your hard work! we hope the aresty rurj continues to grow and cement itself within the rutgers research community and that our readers continue to follow our journey. aresty rurj editor-in-chief, tanvi banota microsoft word forewordtofirstissueofthearestyundergraduateresearchjournalrevmarker.docx foreword  to  first  issue  of  the  aresty  undergraduate  research  journal       a  unique  resource  for  the  rutgers  university  community,  the  aresty  research  center  promotes   the  integral  value  of  research  in  undergraduate  education.     this  inaugural  issue  of  the  journal  expands  the  scope  of  research  activities  the  center  offers  to   rutgers  undergraduates  to  include  the  peer-­‐reviewed  publication  process  –  a  crucial  element  of   any  structured  research  activity.  students  can  engage  with  the  journal  in  a  variety  of  roles  that   all  professional  researchers  take  at  different  times  –  those  of  the  authors  of  scholarly   publications,  those  of  peer  reviewers  who  ensure  the  quality  and  soundness  of  the  published   work,  and  those  of  editors  who  coordinate  the  review  process.       it  is  gratifying  to  see  this  innovative  concept  come  to  fruition  in  the  fall  of  2020.   seeded  by  the  concept  developed  by  a  group  of  students  and  driven  forward  by  their   enthusiasm;  shepherded  through  the  stages  of    development  through  trials,  and  some   inevitable  errors,  by  the  aresty  center  staff;  enthusiastically  supported  by  research  and   academic  administrators;  and  made  possible  in  its  final  form  by  the  investment  of  personnel   and  resources  by  rutgers  university  libraries,  the  journal  is  an  excellent  example  of  what  is   possible  at  this  great  public  research  university.     tireless  efforts  by  students  prachi  srivastava  and  frederic  kelada,  patient  guidance  by  the   aresty  center  director  tamiah  brevard,  open-­‐minded  support  by  dean  susan  lawrence,   creative  input  from  the  learning  centers  director  stacey  blackwell,  and  deep  understanding  of   the  publishing  process  both  intellectual  and  practical  by  librarians  lily  todorinova  and  rhonda   marker  were  all  crucial  for  making  this  new  aresty  center  initiative  a  success.     congratulations,  and  with  great  hopes  for  the  future  of  this  endeavor,     vadim  levin   faculty  director  of  the  aresty  research  center   professor,  earth  and  planetary  sciences,   rutgers  university-­‐new  brunswick   aresty rutgers undergraduate research journal, volume i, issue iv foreword to volume i, issue iv aresty rutgers undergraduate research journal ✵ welcome to the fourth issue of the aresty rutgers undergraduate research journal! this issue exhibits the accomplishments of a diverse set of bright minds emerging from the rutgers undergraduate community. their work focuses on tackling interdisciplinary challenges and contributing to dialogues at the frontiers of research in the sciences and humanities. tracking the dynamic carbonate chemistry of our oceans, using in silico techniques to examine the pharmacological potential of specific medicinal plants, and assessing the social value of the american shopping mall all of these topics and more are explored within this issue. the aresty rurj is incredibly proud to showcase these students’ achievements through their writing and is eager to see what they achieve next. this issue marks the aresty rurj’s first iteration with a significant change in student leadership; all students involved with the journal at its inception in 2020 have moved on. just as our submissions receive rigorous peer review, the aresty rurj is undergoing its own version of peer review as new minds on the editorial team evaluate and revise the journal. we are constantly striving to streamline our standard practices and procedures while seeking and developing new initiatives for the aresty rurj to grow. we are confident that the future of the aresty rurj is abundant in these opportunities. i would like to acknowledge and thank the many influential people whose efforts and commitment to the aresty rurj were paramount in the making of this issue. thank you to the editorial team (reviewer program director, managing director, chief copy editor, and marketing director) for exceeding expectations in taking on and managing various responsibilities involved in running the aresty rurj. thank you to the faculty and staff at the aresty research center and rutgers libraries for their helpful mentoring and guidance throughout our transition. lastly, thank you to all of our reviewers and copy editors who devoted their time to reading and polishing the manuscripts presented in this issue— undergraduate students, graduate students, and faculty alike. each and every reviewer is fundamental to the success of the peer review process. we hope you follow our journey as we continue to capture a glimpse into the exciting and expansive world of undergraduate research hosted across rutgers. thank you for reading! aresty rurj editor-in-chief, jason kawalec aresty rutgers undergraduate research journal, volume i, issue ii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. about the authors julian bowne is a senior at rutgers university studying aerospace engineering. he is a member of the rutgers rocket propulsion laboratory, and he has been doing research relating to space exploration. julian’s research involves control systems, robotics, and aerospace engineering. his current project has been an extension of a three-year long research project involving a traction control system for a martian rover. first, the project was a simple abs control scheme using pacejka’s magic tire formula. then, through nasa’s space grant consortium, the project developed into using a more complex terrain model, called the deformable-soil-rigid-tire model. next, under the aresty program, the project focused more on designing a controller using the sliding mode control method. currently, the project’s main focus is data-driven control, which is an advanced control system that uses data instead of a model. this project continues to evolve, revealing better solutions using more advanced techniques. contact: jkb144@scarletmail.rutgers.edu robert buckelew is a senior at rutgers university studying aerospace engineering. his academic and research interests focus on the modeling, simulation, and control of dynamical systems. his most recent collaborative project gained inspiration from a darpa initiative entitled "robotic servicing of geosynchronous satellites" and sought to produce a useful simulation tool for a complex satellite-mounted robotic arm system. robert hopes to apply the critical thinking skills developed throughout his time in the rutgers undergraduate research environment to solve problems of importance in the chemical processing industry following graduation. contact: robert.buckelew@rutgers.edu ethan catalanello is an undergraduate senior in mechanical engineering. his research is primarily in the field of nonlinear controller design and robotics. he is particularly focused on employing numerical models and computational analysis to create simulations of coupled dynamical systems, such as a freefloating satellite with mounted robotic manipulators. these simulations greatly reduce the cost and risk associated with refining spacecraft design since they do not require a physical launch. recently, he worked to design a controller that could potentially be utilized in the darpa rsgs project. ethan hopes that his research contributes to the proliferation of human presence in space by helping to encourage private organizations to allocate more resources to developing the next era of spacecraft. contact: ethan.catalanello@rutgers.edu aresty rutgers undergraduate research journal, volume i, issue ii yuexing hao is a rutgers undergraduate student majoring in computer science. through her undergraduate time, she spent time doing research in intelligent visual interface (ivi) lab and publishing papers in different leading journals such as bioinformatics, iet image processing. yuexing is interested in computer vision and machine learning. she is currently a part-time lecturer of rutgers sas signature course data 101. in the future, yuexing is going to pursue her dream as a ph.d. student. contact: yh599@scarletmail.rutgers.edu shane le compte is currently a senior at rutgers university studying mechanical engineering and mathematics. he loves creating things, solving problems in new ways, and is always open to a challenge. aside from academics, shane enjoys classical music and playing games with friends. the broader context of this research fits within the jj slade program at rutgers where this project is currently being continued. the end goal is to form a baseline from which real applications to space exploration can be realized within the near future. contact: stl61@scarletmail.rutgers.edu adriana scanteianu is a fourth-year student in the honors college and a member of the douglass residential college studying mathematics with a minor in urban studies. she is a rutgers presidential scholar, a lloyd c. gardner fellow, and a recipient of the katherine hazard prize in mathematics. adriana has also been recently accepted into the 3+1+1 program at the bloustein school, which culminates in a master of public policy. adriana’s research experience began in high school when she conducted and published computational biology research at the icahn school of medicine at mount sinai. as a rutgers student, she traveled to rwanda under the western washington university nsf reu program to study rwanda’s family planning system from both quantitative and qualitative perspectives. last summer, adriana began conducting research in quantum mechanics through the rutgers dimacs nsf reu program under prof. shadi tahvildar-zadeh, alongside rutgers student xiangyue wang. this research continued throughout the academic year and was accepted for presentation at a contributed paper session at the american mathematical society spring eastern sectional meeting. adriana and xiangyue have continued to work on their research remotely and will be working with a new rutgers dimacs reu cohort this summer. contact: adriana.scanteianu@gmail.com aresty rutgers undergraduate research journal, volume i, issue ii hannah varkey: for a phenomenon that affects nearly two-thirds of the world’s population, we have still yet to learn about the monsoonal system, particularly as its predictability wavers in the midst of the climate crisis. living in india for a significant part of my life, i’ve come face-to-face with the effects of the indian monsoon, impacting aspects that range from agriculture and economy to water supply and travel. my project started the summer after my first year, with my pi dr. richard mortlock, through the aresty summer science program 2018 internship. to me, my motivations for this project are driven by personal experience—from being able to discern whether the weather is really getting quite worse or unbearable (as my grandparents often tell me) to being able to provide solutions that can change lives for the better. at rutgers, i’m learning cognitive science and the geological sciences, exploring both worlds of the visceral and innate to the external and ever-changing. this perspective allows me to look at issues of climate change from being a neurological and public health concern to being able to appreciate to it in through a more in-depth and paleoclimatological perspective into the monsoon trends of southern asia. contact: hannah.g.varkey@rutgers.edu xiangyue wang is a senior in the honors program studying physics, mathematics, and philosophy. he strives to find innovative and equitable solutions to climate change. he is a rutgers scarlet scholar, an anthony d. kurtz scholar, and a recipient of the nj chinese chamber of commerce special recognition award. furthermore, he represents more than 30,000 undergraduate students in the rutgers senate. xiangyue’s effort to combat climate change began in high school when he was selected by the at&t young science achievers’ program to research the impact of ocean acidification on phytoplankton. later, he searched for sustainable and reusable concrete in the njit material dynamics lab. at rutgers, xiangyue uses physics-based machine learning models to improve wind energy efficiency under prof. ahmed aziz ezzat. xiangyue began conducting research in quantum mechanics through the rutgers dimacs nsf reu program under prof. shadi tahvildar-zadeh, alongside rutgers student adriana scanteianu. this research continued throughout the academic year and was accepted for presentation at a contributed paper session at the american mathematical society spring eastern sectional meeting. xiangyue and adriana have continued to work on their research remotely and will be working with a new rutgers dimacs reu cohort this summer. contact: maxwangmodc@gmail.com mailto:hannah.g.varkey@rutgers.edu aresty undergraduate research journal at rutgers university, vol. 1, issue 1, spring 2020 1 about the authors kayla castellitto conducted this study as a part of her george h. cook honors thesis research, while she was an undergraduate student at rutgers, the state university of new jersey, school of environmental and biological sciences. during her senior year at rutgers, she served as the vice president of the rutgers (omicron beta alpha) chapter of kappa omicron nu (kon) honor society and coordinated the other kon student members’ assistance in the study. upon her graduation in spring 2018, she completed her dietetic internship at rutgers university, school of health professions. after completion of more than 1200 hours of supervised practice in the clinical and community settings, she successfully passed the commission on dietetic registration exam and became a registered dietitian nutritionist (rdn). kayla castellitto is currently a clinical dietitian at robert wood johnson university hospital in new brunswick. dr. nurgul fitzgerald, phd, rdn, is an associate professor / extension specialist at the department of nutritional sciences at rutgers university. dr. fitzgerald served as the faculty advisor to kayla castellitto for her g. h. cook honors thesis research and to the kon honor society students, who were involved in this study. dr. fitzgerald’s research focuses on cultural, socioeconomic, and environmental effects on food access and intake patterns, and community-based interventions for health promotion and type 2 diabetes prevention especially in low-income or minority populations. amanpreet kaur graduated from the school of arts and sciences honors program at rutgers university in january 2020. having majored in biological sciences, she plans on attending medical school. in her junior year, she studied abroad in london, and she remains an avid member of the rutgers global community. in her spare time, she enjoys volunteering with elderly and terminally ill patients, travelling, and drawing. in spring 2018, amanpreet’s interest in the genetics and biology underlying diseases like alzheimer’s and parkinson’s led her to join dr. rongo’s lab in the waksman institute. she subsequently completed her research on the role of eicosanoids in regulating the ubiquitin proteasome system (ups) for credit. anusha patil is a senior at the school of arts and sciences, majoring in cell biology and neuroscience with minors in psychology and health & society. in high school, anusha volunteered in a brain trauma unit at a local hospital. here, she interacted with patients who all suffered brain injury yet exhibited varying levels of ability. she was inspired by the resilience these patients demonstrated on their neurorehabilitation journey, and hearing their lived experiences drove her to conduct research in neurotrauma. this research project unites her prior research experiences studying spinal cord injury at the w.m. keck center and epigenetics aresty undergraduate research journal at rutgers university, vol. 1, issue 1, spring 2020 2 at clef laboratory. anusha hopes to pursue a career in medicine, where she can continue to focus on conducting translational research. yelizaveta rassadkina graduated from rutgers university with honors in the spring of 2019 with a b.s. in microbiology. this opportunity has inspired her to continue doing research and apply to graduate school to get a ph.d. in microbiology. while at rutgers, she did her undergraduate research under her professor, dr. barkay, who studies the microbial transformations of metals in the environment. yelizaveta joined her lab during her junior year and continued to work there until she graduated. during her time she worked on a project with samples from hot springs in yellowstone national park. lots of microorganisms live in harsh environments that contain many toxic metals, such as mercury, and they have developed strategies to overcome the toxicity. her research project included characterizing the microbial community that was present in such extreme environments and then using that information to identify the ways microorganisms are able to resist mercury thus allowing them to grow where most living organisms would not be able to. ryan rodriguez is a rising junior at rutgers university, currently pursuing a major in english and a minor in education. taking a break from the likes of milton, dickens, cavendish, and eliot, and after spending two semesters as an intern at the plangere writing center on campus, ryan has gained substantial interest in measures educators can implement in restructuring pedagogy that offer a more student-driven approach. originally an assignment given in his internship class, the research he explores soon expanded into a study of how learning environments that stand adjacent to the classroom may provide more desirable spaces conducive to learning than traditional ones, and as a student himself— learning of and understanding the struggles new students face as they transition to the college setting— wanted to present an actualized alternative route to teaching that empowered the party it most affects. adriana scanteianu is a rising fourth year student in the honors college and a member of the douglass residential college studying mathematics with a minor in urban studies. she is a rutgers presidential scholar, a lloyd c. gardner fellow, and a recipient of the katherine hazard prize in mathematics. adriana has also been recently accepted into the 3+1+1 program at the bloustein school, which culminates in a master of public policy. adriana’s research experience began in high school when she conducted and published computational biology research at the icahn school of medicine at mount sinai. as a rutgers student, she traveled to rwanda under the western washington university nsf reu program to study rwanda’s family planning system from both quantitative and qualitative perspectives. last summer, adriana began conducting research in quantum mechanics through the rutgers dimacs nsf reu program under prof. shadi tahvildar-zadeh, alongside rutgers student xiangyue wang. this research continued throughout the academic year and was accepted for presentation at a contributed paper session at the american mathematical society spring eastern sectional meeting. adriana and xiangyue have continued to work on their research remotely and will be working with a new rutgers dimacs reu cohort this summer. aresty undergraduate research journal at rutgers university, vol. 1, issue 1, spring 2020 3 prachi srivastava is a senior in the school of arts and sciences at rutgers university-new brunswick. she is majoring in biological sciences and psychology with minors in music and health & society. in her spare time, she enjoys reading, volunteering, biking, and musicking. prachi’s interest in the intersection of neuroscience and psychology led her to join dr. benjamin samuels’ behavioral neuroscience lab in her freshman year. during her time with the lab, she has given two poster presentations, been granted the cooper surf award, and contributed to two published manuscripts. her hope is to continue working in neuroscience research with a focus in neurodegenerative diseases. xiangyue wang is a rising senior in the honors program studying physics, mathematics, and philosophy. he strives to find innovative and equitable solutions to climate change. he is a rutgers scarlet scholar, an anthony d. kurtz scholar, and a recipient of the nj chinese chamber of commerce special recognition award. furthermore, he represents more than 30,000 undergraduate students in the rutgers senate. xiangyue’s effort to combat climate change began in high school when he was selected by the at&t young science achievers’ program to research the impact of ocean acidification on phytoplankton. later, he searched for sustainable and reusable concrete in the njit material dynamics lab. at rutgers, xiangyue uses physics-based machine learning models to improve wind energy efficiency under prof. ahmed aziz ezzat. xiangyue began conducting research in quantum mechanics through the rutgers dimacs nsf reu program under prof. shadi tahvildar-zadeh, alongside rutgers student adriana scanteianu. this research continued throughout the academic year and was accepted for presentation at a contributed paper session at the american mathematical society spring eastern sectional meeting. xiangyue and adriana have continued to work on their research remotely and will be working with a new rutgers dimacs reu cohort this summer. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. chronic corticosterone shifts effort-related choice behavior in y-maze prachi srivastava written under the supervision of dr. benjamin samuels abstract major depressive disorder (mdd) affects more than 16.1 million american adults (about 6.7% of the u.s. population age 18+) in a given year (adaa). understanding how chronic stress impacts decision-making may allow us to help people suffering from mdd receive the most suitable antidepressant treatment given their behavioral tendencies in reward and motivational processing. in these experiments, our objective is to characterize effort-related choice tasks using chronic stressors in mice. to study this topic, we take advantage of the well-validated homology between corticosterone (cort) in rodents and cortisol in humans to induce mood disorders such as mdd and chronic stress in mice. effort-related choice tasks have been characterized using stressors in rats, but the stressors have not been chronic and have not been characterized in mice. the results of these experiments would be a better foundation for research involving antidepressant treatment experiments on mice. the experimental group was administered cort in their drinking water throughout all experiments. then, both the control and experimental groups were tested in a y-maze barrier task to demonstrate the behavioral effects of cort. the high reward (hr) arm of the y-maze contained a reward of four food pellets, which required high effort to obtain. the low reward (lr) arm of the y-maze contained a reward of two food pellets, which required less effort to obtain animals chronically exposed to cort displayed a stronger preference for low-effort, low-reward choices than control subjects. the results suggest that chronic cort may reduce motivation to work for a highly rewarding reinforcer when a less rewarding reinforcer is available. introduction motivation is a critical aspect of reward-related behaviors and is affected by mood disorders such as mdd (schrader 1997). an important component of motivation is that it initiates and maintains goal-directed behaviors requiring persistent effort. this activating component of motivation can be specifically assessed in preclinical experimental models of depression using effort-related choice tasks (salamone, et al. 1994). studies conducted by belzung, et. al (2011) outline the criteria for animal models of psychiatric disorders, particularly in anxiety and depressive disorders, which are the focus of this study. these criteria include mechanistic validity, face validity, and predictive validity. mechanistic validity indicates the extent to which the cognitive or biological mechanisms https://creativecommons.org/licenses/by-nc-sa/4.0/ https://creativecommons.org/licenses/by-nc-sa/4.0/ aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 2 underlying stress disorders (such as stress hormone dysregulation) are similar in humans and the model organism. face validity indicates the extent to which either the observable behavior (such as anhedonia) or the biological outcomes (such as elevated corticosterone, or cort) are analogous in humans and the model organism. finally, predictive validity indicates the extent to which a triggering factor will induce similar disorder-related outcomes in humans and the model organism, and whether the effects of treatments on the model organism mirror those on humans. the relevance of this framework is later discussed regarding various animal models of depression. effort-related choice behaviors are behaviors in which a subject chooses between a highly rewarding reinforcer that requires high effort to obtain and a less rewarding reinforcer that requires significantly less effort to obtain. the behavioral test discussed in this study is a yshaped barrier maze task. in the y-maze barrier task, the rodents may choose to climb over a wire-mesh barrier to acquire four food pellets in the hr arm of the y-maze, versus traversing the lr arm with no barrier to acquire two food pellets (yohn, et al. 2015). this behavioral task is used to determine whether a rodent is motivated to exert effort for a highly rewarding reinforcer or if the rodent will shift its preference to the reward that requires significantly less work to consume. the effort-related choice tasks in rodents can be seen as analogous to effort-expenditure tasks in humans (belzung and lemoine 2011). due to their comparability, effort-related choice tasks provide a valid and translatable behavioral assessment for studying motivational processes in rodents. salamone et al. (2003) suggest that drugs that interfere with dopamine transmission such as cort alter the rewarding aspects of primary reinforcers such as food. cort is a stress hormone analogous to human cortisol and is the major output of the hypothalamuspituitary-adrenal axis, which is hyperactive in both humans with mdd and in rodent models of chronic stress. chronic cort has been shown to induce negative valence behaviors such as anxiety and impair positive valence behaviors such as reward processing (dieterich 2019). therefore, cort was used to ensure mechanistic, face, and predictive validity. the effects of chronic cort administration on effort-related choice behavior has not been characterized. since stress-induced mood disorders such as depression involve deficits in reward processing and motivation, further study of how chronic stress influences effortrelated choice behavior is needed. we hypothesized chronic cort would reduce motivation for the high effort/high reward choices and shift responding to the less effort/less reward choices. to test this hypothesis, mice were chronically administered cort and then tested in a y-maze barrier task. methodology control and experimental populations there were two groups of male mice: the control group (cg) mice (n=6) and the experimental group (eg) mice (n=8). the cg and eg mice were treated with β-cyclodextrin aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 3 dissolved in their drinking water at a dose of 24 mg/kg/day. β-cyclodextrin was used to promote the subjects’ drinking behaviors and increase consumption of the drinking water containing the dissolved cort (david et al., 2009). the eg mice were treated with a dose of 5 mg/kg/day of cort, which was chosen based on previous studies (david et al., 2009). eg mice were administered cort during the four weeks prior to the experiment as well as throughout the experiment. figure 1. y-maze setup. the y-maze consisted of a start box (arm 1) where mice were placed to begin each trial. when a plexiglas barrier was removed, the mouse could traverse the y-maze and enter either of the two arms. the high reward arm (arm 2) side contained four pellets, and the low reward arm (arm 3) contained two pellets. after training, progressively taller barriers were introduced into arm 2 to force an effort-related choice between arms. arm 2 and arm 3 were counterbalanced across mice such that for half of the subjects, arm 2 was opposite to what is indicated in this schematic. y-maze design a mouse version of the y-maze barrier task used with rats (yohn et al., 2015) was implemented to assess the effect of chronic cort on effort-related decision-making. to begin each trial, mice were placed in a start box (arm 1) behind a plexiglas barrier (fig 1). when the plexiglas barrier was removed, the mouse could traverse the y-maze. once it moved towards the center of the maze, the mouse could choose between entering arm 2 or arm 3 (fig 1). mouse training on y-maze after being given treated water for four weeks, mice were trained in the y-maze barrier choice task to determine if chronic cort alters effort-related choice behavior. mice were aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 4 first habituated to the maze over the course of two days in two 10-minute sessions with unlimited food pellets in arms 2 and 3 of the y-maze (fig 1). these initial habituation trials allow the mice to navigate through and become more familiar with the y-maze. mice then completed an additional five habituation trials over the course of five days, which ended after they had entered both arms and consumed all available food pellets. in the hr arm, arm 2, there was a food bowl containing four pellets. in the lr arm, arm 3, there was a food bowl containing two pellets. this experimental setup is outlined in figure 1. the mice had free access to the entire maze during the habituation sessions, and no barriers were present in the arms. this was done to allow the mice to become acclimated to the maze and the differences in reward availability. after habituation, mice completed two days of forced-choice sessions. these sessions included ten forced-choice trials where each arm of the maze was blocked off in alternating trials. these trials further acclimate the animals to the differences in reward between the hr and lr arms. mice then were trained with five days of free-choice sessions, which began with two forced-choice trials followed by 10 free-choice trials. the free-choice sessions began with two forced-choice trials to serve as a reminder of the y-maze’s layout at the start of each day. this accounts for any loss of learning in between trials. following the two forced-choice trials, there were 10 free-choice trials. each free-choice trial began with the mouse in the start box in arm 1 where it had free access to choose between the hr and lr arms. in these sessions, there was no barrier placed in the y-maze. once the mouse entered one of the two arms, the entrance of that arm was blocked off with a barrier. the mouse then had one minute to consume the pellets before it was removed and returned to its home cage. mice were trained in these free-choice sessions until they reached the criterion of 70% accuracy in choosing the hr arm. while the cg mice took 2-3 free-choice sessions to reach the criterion, the eg mice took all 5 days of free-choice sessions to reach the criterion. thus, both groups reached the criterion prior to progressing to the following sessions. in the following sessions, a barrier was added to the hr arm, requiring high effort since the mouse had to climb over a barrier before it could reach the food bowl. in arm 3, no barrier was added; thus, reaching the food bowl required significantly less effort. the barriers used were 10, 15, or 20 cm tall wire-mesh barriers. mice completed three days of testing with the 10 cm barrier in the hr arm, three days with the 15 cm barrier, and three days with the 20 cm barrier. upon completing all test sessions, mice underwent three control sessions where 10 cm barriers were present in both hr and lr arms. this was done to test if cort administration affected reward discrimination or the ability to climb over the barriers. data analysis a two-way anova was used to examine the effect of cort administration on hr arm selection in the multiple y-maze conditions described in the methods (fig. 2). aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 5 figure 2. experimental timeline. after four weeks of cort and β-cyclodextrin administration to the eg (n=8) or β-cyclodextrin only to the cg (n=6), mice were habituated to the maze over the course of two days with unlimited food pellets in the arms. mice then completed an additional five days of habituation, with one arm always containing four pellets (arm 2) and the other two pellets (arm 3). after habituation, mice completed two days of forced-choice sessions and then were trained in five days of free-choice sessions with the hr and lr. then, a barrier was added on arm 2. mice completed three days of testing with the 10 cm barrier in arm 2, followed by three days with a 15 cm barrier, followed by three days with a 20 cm barrier, and lastly by a discrimination test with a 10 cm barrier in both arms. results for the results to be significant the p-value must be < 0.001. using bonferroni’s multiple comparisons test (bonferroni 2013), results suggest that cort reduced hr arm selection when there was a 15 cm barrier (p < 0.001) or a 20 cm barrier (p < 0.001) placed in the hr arm. importantly, there was no effect of cort on hr arm selection without a barrier in the hr arm (p = 0.149), with a 10 cm barrier in the hr arm (p = 0.779), and when there were 10 cm barriers in both arms of the maze (p > 0.999). therefore, chronic cort shifts preference from hr to lr when greater effort is required to obtain the high reward, as hr arm selection was reduced with the 15 and 20 cm high barriers. cort did not influence arm selection in trials without barriers or with 10 cm barriers in both arms compared to the cg. the results suggest that chronic cort reduced hr arm choice in the y-maze when more effort was required to obtain the four food pellets, shifting preference to the lr arm where only two pellets were available. discussion we used a y-maze barrier choice task to assess the effect of chronic cort on effort-related choice behavior in mice. effort-related choice behaviors have not been well-characterized in mice in other research (cagniard et al. 2012). also, no studies have assessed the effect of chronic cort on effort-related decision-making. many previous studies using chronic cort have modeled negative valence behavior tests that induce stress, anger, or fear. these tests include forced swim tests that force rodents to avoid drowning. negative valence tests such as the forced swim test are not well translatable to humans. however, the use of chronic cort in positive valence behavior tests have been less well-characterized in animal models of depression. positive valence behavior tests are tasks that induce pleasure or happiness. these include effort related choice tests that assess rodents’ preference for higher rewards that require high effort versus lesser rewards that require less effort. these tests are more aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 6 effective for testing behaviors using chronic cort because the animal models for these behaviors are more translatable to human models of positive valence behaviors. figure 3. effect of cort on hr arm selection in multiple y-maze conditions. a two-way repeated measures anova indicates a significant main effects of y-maze session, (f(4, 48) = 76, p < 0.001), and cort administration, (f(1, 12) = 15, p = 0.002), as well as a significant interaction, (f(4, 48) = 6.7, p < 0.001). comparisons indicate that cort administration reduced hr arm selection relative to control administration for 15 cm barrier (p < 0.001) and 20 cm barrier (p < 0.001) test sessions. thus, chronic cort shifts preference from hr to lr when greater effort is required to obtain the high reward. cort did not influence arm selection compared to the cg in trials without barriers or with 10 cm barriers in both arms. the results suggest that chronic cort reduced hr arm preference in the ymaze when more effort was required to obtain the high reward and shifted choice to the lr arm with the lesser reward. the results of our experiments suggest that chronic cort alters effort-related choice behavior in the y-maze barrier task in mice. chronic cort administration may reduce motivation to work for and obtain a highly rewarding reinforcer when a lesser reinforcer is freely available. this is the first study to examine the effect of chronic stress on effortrelated choice behaviors in mice. future studies might compare behavioral changes in male and female mice to obtain more comprehensive results. additionally, since chronic cort administration may impair spatial memory—which could impact selection of the hr or lr arm independently of the effort-related choice—the relationship between cort administration and spatial memory is another area for future research. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 7 since chronic mood disorders such as depression are highly prevalent, we hope that this research will ultimately be useful for identifying novel treatments, determining whether patients will respond to treatments, and developing personalized treatment strategies. references aberman je, salamone jd. nucleus accumbens dopamine depletions make rats more sensitive to high ratio requirements but do not impair primary food reinforcement. neuroscience. 1999; 92:545–552. anxiety and depression association of america (adaa). (n.d.). retrieved from https://adaa.org/about-adaa/press-room/facts-statistics belzung, c., & lemoine, m. (2011). criteria of validity for animal models of psychiatric disorders: focus on anxiety disorders and depression. biology of mood & anxiety disorders, 1(1), 9. doi:10.1186/2045-5380-1-9 bonferroni's method. nist/sematech e-handbook of statistical methods, 30 oct. 2013, www.itl.nist.gov/div898/handbook/prc/section4/prc473.htm. cagniard, b., p. d. balsam, d. brunner and x. zhuang (2006). "mice with chronically elevated dopamine exhibit enhanced motivation, but not learning, for a food reward." neuropsychopharmacology 31(7): 1362-1370. david, d. j., b. a. samuels, q. rainer, j.-w. wang, d. marsteller, i. mendez, m. drew, d. a. craig, b. p. guiard and j.-p. j. n. guilloux (2009). "neurogenesis-dependent andindependent effects of fluoxetine in an animal model of anxiety/depression." 62(4): 479-493. dieterich, a., srivastava, p., sharif, a. et al. chronic corticosterone administration induces negative valence and impairs positive valence behaviors in mice. transl psychiatry 9, 337 (2019). https://doi.org/10.1038/s41398-019-0674-4 eerde, wendelien. (2015). motivation and reward systems. 10.1002/9781118785317.weom060146. salamone, j. d., m. correa, s. mingote, s. j. j. o. p. weber and e. therapeutics (2003). "nucleus accumbens dopamine and the regulation of effort in food-seeking behavior: implications for studies of natural motivation, psychiatry, and drug abuse." 305(1): 1-8. salamone, j. d., m. s. cousins and s. bucher (1994). "anhedonia or anergia? effects of haloperidol and nucleus accumbens dopamine depletion on instrumental response selection in a t-maze cost/benefit procedure." behav brain res 65(2): 221-229. schrader, g. j. c. p. (1997). "does anhedonia correlate with depression severity in chronic depression?" 38(5): 260-263. yohn, s. e., c. thompson, p. a. randall, c. a. lee, c. e. muller, y. baqi, m. correa and j. d. salamone (2015). "the vmat-2 inhibitor tetrabenazine alters effort-related decision making as measured by the t-maze barrier choice task: reversal with the adenosine a2a antagonist msx-3 and the catecholamine uptake blocker bupropion." psychopharmacology (berl) 232(7): 1313-1323. https://doi.org/10.1038/s41398-019-0674-4 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 8 acknowledgements this work was funded by nimh r01 mh112861 (bas). aresty rutgers undergraduate research journal, volume i, issue ii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. models for predicting global plastic waste yuexing hao, glenn shafer (faculty advisor) ✵ abstract for more than half a century, plastic products have been a part of people’s lives. when plastic waste is thrown into nature, it can cause a sequence of dangerous effects. previous researchers estimated that global plastic waste in 2020 will be more than 400 million tons. to reduce plastic waste, they built scientific models to analyze the sources of plastic and provided solutions for regenerating these plastic wastes. however, their models are static and inaccurate, which may cause some false predictions. in this paper, we first observe the distribution of the real-world plastic waste data. then, we build simple exponential growth model and logistics model to match these data. by testing different models on our plots, we discover that the self-adaptive model is the best to describe and correctly predict our future plastic waste production, as this model combines the benefits of simple exponential growth model and the logistic model. the self-adaptive model has the potential to minimize the error rate and make the predictions more accurate. based on this model, we can develop more accurate and informative solutions for the real-world plastic problems. 1 introduction plastic waste became a serious problem in the 21st century. according to data from 2017, scientists estimate that human beings produced around 8300 million metric tons (mt) of virgin plastics. as of 2015, approximately 6300 mt of plastic waste had been generated. of this waste, only around 9% had been recycled[1]. most plastic products are sent directly to landfills or make their way to the ocean. it may take one hundred to thousands of years for plastic products to decompose, and the available places for landfill waste are becoming fewer and fewer[2,3]. therefore, we should prevent this global crisis from becoming more serious. in this paper, we will introduce three models and find the model best suited for better prediction making. based on the previous data, we will consider more factors and make the model more adaptive to the real-world situations. there are some researchers who modeled past plastic consumption and predicted the future. however, our model will be more accurate and have better data visualization. 2 methodology to estimate the maximum levels of singleuse or disposable plastic product waste, we should first know the constraints and extent. the constraints are the efficiency of the natural environment or the waste disposal companies at processing plastic pollution. if their efficiency of disposing plastic products is higher than the rate of people producing plastic waste, there will be no plastic pollution at all. therefore, we consider the constraint to be the availability of processing plastic waste. the extent is the range of our plastic waste, since our plastic waste cannot increase to infinity. we should set an upper bound and lower bound for the weight of our waste. the hard part of this problem is that if nations and governments do not accurately know their nation’s total production of plastic waste, it increases the difficulty of managing single-use plastic waste comprehensively. the composition of modern plastic waste is complex because the waste may contain both single-use and multi-use plastics. however, we can estimate the disposable plastics product usage by finding the characteristics of plastics waste’s historical change. then, based on this trend, we can predict the quantity change of plastic waste in the future. in addition, there emerges another difficult aspect of this problem: there are many influencing factors regarding the generation of plastic waste. aresty rutgers undergraduate research journal, volume i, issue ii there is no simple pattern or law that can be determined by looking at the data. we should, therefore, develop a model which does not need a typical data distribution law and use less theoretical calculations that can still make an accurate prediction. that is to say, we may not need to add much mathematical functions or calculations inside the model, but the model can adaptive the different situations by itself. therefore, in this paper, we employ the self-adaptive model to estimate and analyze the data on plastic waste production over the past 5 to 10 years, and then predict its production in the next 30 years. it is called a self-adaptive model because this model does not have specific relationships among factors in a system. we cannot clearly indicate the relationships between the factors in the data, such as the influences of countries’ rules, population density, levels of development. thus, it is a good choice to use the self-adaptive model for this problem. 3 building plastic pollution models in this part, we first define plastic pollution into seven kinds of waste: textiles, transportation, packaging, electronic, consumer products, industrial machinery, building & construction. these different kinds of waste will be considered as different variables. currently, plastic packaging is the greatest source of primary plastic production. because of the rapid development of online shopping and delivery services, the demand for plastic has increased rapidly in just a few years. thus, we can find some connections between these variables and the total amount of plastic waste[2]. to estimate the maximum mitigation levels for plastic waste without destroying the environment, we take the seven kinds of waste from different kinds of single-use or disposable plastic waste into account. correspondingly, we consider each plastic waste factor’s weights, the resources to process, the total budget, and the cost of the environmental damage it causes. we also set the damage levels for the environment to make our models more accurate. after comparing each factor, we find the best combination to maximally mitigate the dangerous effects that plastic waste brings about. i. the simple exponential growth model before the 21st century, from the 1960s to the 1990s, the simple exponential growth model (also called as malthusian model) is a great model for describing the exponential growth of plastic waste[5]. the model could perfectly match with the data and creates great visualization. however, after fitting the data, we discovered that plastic waste’s rate of increase is not as fast after 2000. this is because nations and governments, as intervention variables, have paid more attention on this global crisis and already taken some measures to prevent more plastic pollution. to show the intervention variables, we decided to switch our models to be more accurate. figure 1: global plastic production per year. as shown in the figure, plastic waste was largely static from 1950-1960. after this, plastic products became prevalent in humans’ lives, and plastic waste increased rapidly from 1960 onward. there were some significant increases between the 1970s and 2000s[4]. to estimate the maximum mitigation levels for plastic waste without destroying the environment, we take the seven kinds of waste from different kinds of single-use or disposable plastic waste into account. aresty rutgers undergraduate research journal, volume i, issue ii ii. the logistic model the logistic model was developed by belgian mathematician pierre verhulst[8]. he came up with this model’s idea by thinking about the rate of population increased may be limited. it may be affected by many other factors, which may not always increase. we agree with the thoughts of varhulst: the global plastic waste may not increase after reaching some points. therefore, we used logistic model here. however, after 2015, we observe that the model’s coefficient is not a constant anymore. the curve’s rate of increase will change because of intervention variables. that is to say, the model’s coefficient should be a flexible number. to make the coefficient adjustable, based on our logistic model’s differential equation, we created another self-adaptive differential equation. iii. the self-adaptive model based on the differential equation from last section, we develop a more advanced model called a self-adaptation model, for which the coefficients are flexible and easy to change. some researchers call this kind of model a grey model, since it can be flexible between black model or white model. under this model, we create a self-adaptation differential equation, which can be easily adapted in different conditions. by employing this model, our predictions are close to reality and the error is small. since the simple exponential growth model’s predictions between 1950-1970 and logistic model’s predictions before 2010 are accurate, we will employ both models in our new self-adaptive model. the coefficient of plastic waste before 2010 is nearly static[4]. however, after 2010, when more countries and nations began to ban or limit the use of single-use or disposable plastic products, the coefficients of the plastic waste curve began to decrease. after 5-10 years, the self-adaptive model will change the coefficients, which makes it more adaptable to real-world data. after comparison of these models, we chose to employ a self-adaptive model as our main model to predict future plastic waste production. 4 selection of the models the criteria of model’s selection are based on the accuracy of the models. (see appendix 1)  firstly, we started by building a simple exponential growth model and trying to find a valuable result. before the 21st century, around 1960s to 1990s, the malthusian model is considered a great model for describing the exponential growth since the function is proportional to the speed to which the function grows.  however, after fitting the data, we discovered that the plastic waste’s increasing rate is not that fast anymore. this is due to the fact that nations and governments, as intervention variables, paid more attention on this global crisis and already took some measurements to prevent more plastic pollutions. to show the intervention variables, we decided to switch our models to a more accurate model. figure 2: the logistic model for fitting the real-world data. this plot shows that our logistic model is perfectly suited to the real data, where the x-axis represents the year, and the y-axis represents plastic production. compared to our previous model, the logistic model is closer to the realword data from 1960 to 2015. aresty rutgers undergraduate research journal, volume i, issue ii  secondly, we built a logistics model. the logistics model matches with our real data from 19602015. however, after 2015, we found out that the model’s coefficient is not a constant anymore. the increasing rate will also change because the intervention variables. that is to say, the model’s coefficient should be a flexible number. to make the coefficient adjustable, based on our logistics model’s differential equation, we created another self-adaptive differential equation.  lastly, we move our eyes to the self-adaptive model, in which the coefficients are flexible and easy to change. using this model, we created a self-adaptation differential equation. by employing this model, our predictions and the reality are similar, and the error is small. therefore, after comparisons, we chose to employ the self-adaptive model as our main model to predict the future plastic waste productions. 5 results to examine the accuracy and efficiency of our models, we compared each year’s plastic waste and the data calculated from each model. all these three models' error rates are low in the beginning before the 21st century[4]. then, depending on the different functions, each model has some deviation from the real-world data. by graphing all the models (figure 3), we can easily see the differences between the models and the best model for future prediction. therefore, we concluded that the self-adaptive model is the most accurate one, performing much better than the simple exponential growth model or logistic model. the strength of our model is that we take three models and use them to continue to improve the accuracy of our model. we built up the first simple exponential growth model, which gave us a general idea about the trend of plastic waste. based on the prediction model, we came up with another logistic model. by employing the logistic model, the results perfectly match our data. we then tried to use logistic model to predict our future data; however, the results deviated from the prediction data. we figured out that the problem might be the coefficients of our logistic model, which should not be a constant. since the coefficients are flexibly changing, we switched our model to a self-adaptive model. 6 discussion there are a number of plastic waste prevention techniques, which can commonly be summarized as the 4 rs: reduce, reuse, recycle, and recover. before, people summarized this technique as 3rs, figure 3: the prediction of different models. in this figure, we can visually compare the prediction results of these three models. blue dots represent real-world data[4], and while yellow dots represent prediction data made by previous researchers[5]. we can clearly see that a blue linear line of a trend. this line just shows a trend of the future plastic waste but has no meaning to the predictions. the simple exponential growth model (red) provides a great example of the maximum level of plastic pollution that the system can handle. it could perfectly suit current data, but not future data. the logistics model (purple) shows the exponential growth of the prediction. after 2030, the logistics model deviates with the real-world data. the selfadaptive model (green) displays the flexible curve which suits not only the current data, but also future data. by using a self-adaptive model, our predictions can be more detailed and precise. aresty rutgers undergraduate research journal, volume i, issue ii which did not include recovery[6]. however, more and more people are realizing that it is not enough to recycle. they also need to reduce and reuse recycled plastics. the total population of the above 15 countries are around 5 billion, including the development and developed countries. from figure 4, we find out that developing countries usually produce less plastic waste than developed countries. all nations are responsible for managing their waste production. for example, the countries above the trend should make more actions to prevent increasing plastic waste further. governments can design laws to prevent highly plastic polluting companies from producing single-use or disposable plastics. they can also reduce taxes for those companies that have plastic recycling technology and produce multipleuse or biodegrade plastic products. these actions could limit single-use or disposable plastic production. moreover, they would provide a cleaner and healthier plastic cycle. since plastic pollution has become a global crisis, governments, companies, and human beings have realized the urgency of recycling plastic products and reducing single-use or disposable plastic. in our self-adaptive model, we predict that the rate of increase of plastic waste in 2030 and 2040 will be much slower than the previous year’s rates∎ figure 4: the relationship between gdp and plastic waste. we choose 15 countries from different regions and development levels[4]. the sizes of the circles represent the total plastic waste of that country. we can conclude from the graph that when gdp goes up, plastic waste per year also increases. in the other word, the total plastic wastes have some connections with the country’s degree of development. aresty rutgers undergraduate research journal, volume i, issue ii 7 reference [1] geyer, r., jambeck, j. r., & law, k. l. (2017). production, use, and fate of all plastics ever made. science advances, 3(7), e1700782. [2] jambeck, j. r., geyer, r., wilcox, c., siegler, t. r., perryman, m., andrady, a., … & law, k. l. (2015). science, 347(6223), 768-771. [3] li, w. c., tse, h. f., & fok, l. (2016). plastic waste in the marine environment: a review of sources, occurrence and effects. science of the total environment, 566, 333-349. [4] hannah ritchie and max roser (2020). plastic pollution. published online at ourworldindata.org. [5] eriksen, m., lebreton, l. c., carson, h. s., thiel, m., moore, c. j., borerro, j. c., ... & reisser, j. (2014). plastic pollution in the world's oceans: more than 5 trillion plastic pieces weighing over 250,000 tons afloat at sea. plos one, 9(12), e111913. [6] lebreton, l., andrady, a. (2019). future scenarios of global plastic waste generation and disposal. palgrave commun 5. [7] cramer, j.s. (2002) the origins of logistics regression. chapter 9 of logit models from economics and other fields, cambridge university press. aresty rutgers undergraduate research journal, volume i, issue ii 8 supplementary tables 9 see also https://prezi.com/view/bjdexjkypkmav4zxq2tz/ year plastic wastes (tonnes) simple exponential growth model error (%) logistics model error (%) self-adaptive model error (%) 1960 1970 1980 1990 1995 2000 2005 8,000 35,000 70,027 120,383 152,068 213,209 263,002 35,110 70,069 120,443 151,099 213,028 263,799 0.3143 0.0599 0.0498 -0.6372 -0.0849 0.0303 35,069 70,090 120,409 152,430 213,790 262,997 0.1971 0.0899 0.0216 0.2381 0.2725 -0.0019 34,980 70,431 120,011 151,995 213,563 263,503 -0.0571 0.0769 -0.0390 -0.0480 0.1717 0.1905 2010 2011 2012 2013 … 2017 2018 2019 313,000 325,089 338,797 352,642 … 368,782 398,603 442,488 320,263 329,994 380,863 423,569 … 457,592 482,905 540,000 2.3204 1.5088 12.4163 20.1130 … 24.0820 21.1494 22.0372 320,067 324,989 339,002 353,059 … 369,728 399,766 444,050 0.2257 -0.0308 0.0605 0.1183 … 0.2565 0.2918 0.3530 313,067 325,209 338,997 352,368 … 368,597 398,645 442,766 0.2578 0.03690.0 590 -0.0777 … -0.0501 0.0105 0.0628 2020 2021 2022 2023 2024 2025 … 486,323 510,360 546,085 584,311 625,213 668,978 … 589,670 600,678 634,335 669,895 700,000 724,330 … 21.2507 17.6970 16.1605 14.6470 11.9618 8.2741 … 487,982 513,480 546,999 582,994 624,785 670,012 … 0.3411 0.6113 0.1673 -0.0543 -0.6846 0.1546 … 486,373 510,299 546,327 583,994 625,869 668,321 … 0.0461 0.0623 0.0443 -0.0542 0.0105 -0.0982 … 2030 715,806 768,450 7.3545 715,949 0.1997 624,330 0.0529 2040 876,893 1000,785 14.1285 879,323 0.2771 823,780 0.0947 2050 1,275,997 1,868,544 46.4379 1,276,544 0.4264 1,276,544 0.0426 supplemental table 1: the comparison of total plastic wastes and its simulation based on three models https://prezi.com/view/bjdexjkypkmav4zxq2tz/ aresty rutgers undergraduate research journal, volume i, issue ii [a] school of earth and environmental sciences, queens college, cuny, flushing, ny [b] department of earth, atmospheric, and planetary sciences, mit, cambridge, ma this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. investigating the strength of the indian monsoons during climate extremes with stable isotope records in corals hannah varkey, richard mortlock (faculty advisor), cecilia m. mchugh[a], dhiman r. mondal[b] ✵ abstract the indian monsoon affects the lives of over a billion inhabitants living in southern asia via the hydrological cycle. agriculture on land and freshwater discharge into the ocean. this discharge and nutrient cycling are tied with the monsoon cycles that directly impact society and the economy. previous studies have demonstrated a strong connection between the strength of the indian monsoon and the cooling of the north atlantic during climate extremes, such as during the last glacial period 20,000 years ago, and the little ice age (~1300-1870 a.d.). in our study, we compare the relative strength of the monsoon during two different climate states: the little ice age (lia) and the modern (2015) with proxy measurements obtained in surface corals from saint martin’s island, southeast bangladesh. we used the oxygen-isotope 18o/16o ratio (δ18oc) of coralline aragonite (caco3) to reconstruct changes in the δ18o of seawater (δ18ow) attributed to freshening from monsoon rains. during both climate states, corals recorded large variations in δ18oc (up to 2 parts per thousand or ‰). we attribute these changes, in part, to local salinity changes which are reflected by variability in δ18ow from local riverine discharge. while our records only represent 5-year snapshots and may not be representative of the average climate state, this data does not support that the monsoon was substantially weaker during the lia compared to the modern. in this study, the coral records indicate subtle patterns of isotopic composition as a function of precipitation and temperature variability, serving as a preliminary for further study through longer records lasting a century. beyond this, it would better our understanding of interactions between extremes in temperature and climate systems. 1 introduction saint martin’s island, bangladesh lies in the heart of the indian monsoon (figure 1) where seasonal shifts in wind direction bring torrential rainfall. the summer monsoons, seasonal wind, and rains brought about by differential heating of the land and water, last from june to september every year. during the winter, the winds reverse towards the southwest direction, and precipitation is reduced. the seasonal cycle of the monsoon influences local seawater hydrography in two important ways. during the summer, sea surface temperature (sst) increases, bringing about an increase in rainfall and local riverine input which serves to lower salinity. the freshening of seawater leads to a decrease in δ18o in seawater (δ18ow) since precipitation has a much lower δ18o as a result of the raleigh distillation process[3]. oxygen isotopes undergo fractionation processes, where water containing the lighter 16o isotope is more likely to get evaporated to form a gaseous or vapor state and then precipitated as liquid in rain. hence, the freshening of seawater via precipitation results in a positive correlation between δ18ow and salinity, as both are lowered. aresty rutgers undergraduate research journal, volume i, issue ii reliable instrumental records of sst and salinity are scarce beyond the early 20th century. corals, however, provide a means for reconstructing surface water conditions beyond instrumental records because the δ18oc recorded by the corals depends upon both the δ18o of the surrounding seawater and its temperature[7]. fractionation of 16o and 18o increases with decreasing temperature so that higher δ18oc is associated with lower ssts and lower δ18oc is associated with higher ssts. many corals display annual banding patterns that reflect changes in the density of the skeletal material. when x-rayed, the corals exhibit couplets of light and dark bands. each dark-light couplet represents a one-year growth. with high resolution micro-milling, one can often obtain samples at monthly resolution for reconstructing past changes in salinity and sst. here, we compare the strength of the monsoon during two different climate states: the little ice age (lia, ~1300-1870 a.d.) and the modern (died in 2015) by comparison of their stable isotope and trace metal chemistry. it is hypothesized that the monsoon should have been weaker during the lia. satellite data suggests st. martin’s island does not experience large seasonal variations in sst (range of less than 3°c). therefore, large changes in δ18oc should be driven by changes in δ18ow with lower δ18ow recorded during the summer monsoon. seasonal variation in sst can also modify δ18oc. in order to constrain this seasonal variability in sst, we present measurements of coral sr/ca, which has been shown to be a reliable recorder of temperature[3]. finally, we use the carbon-isotope ratio of 13c to 12c (δ13cc) as an indicator of coral feeding strategy (photosynthesis vs heterotrophy) and δ13c of the dissolved inorganic carbon pool (dic) in seawater. 2 methodology both lia and modern corals in this study belong to the porites species, which are stony corals with small polyps. porites are important in paleoclimatology studies, frequently used as recorders of past marine conditions. the corals tend to be grey-brown to white in color and form hemispherical mounds or ‘microatolls’ in intertidal zones in the indo-pacific waters[4]. they can have greenish tints to the outer walls, due to their symbiotic relationship with single-celled zooxanthellae within the tissues, or more specifically, corallites. corallites, skeletal cups formed from each polyp, are composed of calcium carbonate and precipitated as the mineral aragonite. the corals were collected during sampling expeditions to st martin’s island in 2015/2016. at present, they are stored in airtight containers and drilled using a micro-mill to collect samples along a transect. figure 1: saint martin’s island: the location of lia coral d09-01corresponds to sm-d09. the modern coral location corresponds to that of sm-q07[4] aresty rutgers undergraduate research journal, volume i, issue ii to identify the banding patterns and to guide micro-milling, x-rays of the slabs were taken at the radiology lab at robert wood johnson university hospital, new brunswick. we micro-sampled the slabs at 0.5 to 1 mm spacings parallel to the growth axis using a manual drill. annual growth bands are ~1 cm wide and suggest our sampling resolution is monthly. approximately 80 μg of each powdered sample was analyzed by stable isotopic mass spectrometry in the stable isotope facility in the department of earth & planetary sciences at rutgers university. isotope data was reported relative to pdb (pee dee belemnite, the standard established for δ18o and δ13c) in the standard per mil (‰) notation (equation 1). measurement precision (1 sd) is 0.08‰ for δ18o and 0.05‰ for δ13c, respectively. for the strontium-calcium (sr/ ca) analysis, an adjoining transect of d09-01 was micro-milled and sampled at a similar spacing. ~70 μg of coral powder was acidified to 400 microliters of 3% nitric acid and analyzed by an inductively coupled plasma atomic emission spectrophotometer or icp-oes for sr/ca isotopic ratios (at the dept. of marine and coastal sciences), similar to methods in schrag, 1999[6]. sr/ca ratios were converted to sst via equation 2, demonstrating a temperature sensitivity of about 2 °c for a change of 0.1 in sr/ca. above: figure 2: sampled d09-01 porites lobata coral below: figure 3: sampled living porites lutea coral equation 1: δ18o is the ratio of stable isotopes oxygen-18 (18o) to oxygen-16 (16o), in a sample relative to the ratio in a standard. it is defined in “per mil” (‰, parts per thousand) equation 2: porites sr/ca = 10.790 (±0.043)– 0.068 (±0.002) x sst (°c), where the sr/ca ratio is expressed in mm/m units.[5] (1) (2) aresty rutgers undergraduate research journal, volume i, issue ii 3 results figure 4 and figure 5 show stable isotope results in the two corals— fossil d09-01 (u-th dated to 1762) and modern (collected in 2015), respectively. oxygen isotopic ratios, δ18o and carbon isotopic ratios δ13c were primary indicators of past sea conditions. the δ18o is a function of temperature and salinity, while δ13c measures productivity and increased photosynthesis from vegetation and organisms that have a symbiotic relationship with corals. this relationship correlates with the amounts of sunlight received over time. repeating patterns of high and low δ18o and δ13c are associated with the banding patterns in both the lia and modern corals (figure 4 and figure 5). these patterns suggest that the changes in isotopic above: figure 4: lia coral d09-01. note that δ18o and δ13c variations are aligned with the high-density and low-density banding displayed in the x-ray image. the red lines plot the δ18o variations and the blue lines, δ13c. below: figure 5: δ18o and δ13c variations shown alongside an x-ray image in the modern coral. aresty rutgers undergraduate research journal, volume i, issue ii values are driven by seasonal changes. the banding patterns suggest both coral records represent 5 years growth. both the mean and range in δ18o in the modern coral is similar to that of the lia coral (~ 1.2‰). the mean lia coral δ18o is higher, by 0.2‰ compared to the modern sample, although differences in the mean values are not statistically significant (table 1 and table 2). the range in δ13c of the living coral are similar to those in the lia coral, about 3‰. the modern coral average, however, is about 1.5‰ lower compared to the lia coral. paleotemperature equations based on δ18oc in porites indicate a temperature increase of 1°c for a δ18o decrease of 0.22‰[7]. therefore, an amplitude change of 1.2‰ would reflect a ~5.5 °c range in sst, or about twice what is observed in instrumental records. sr/ca ratios are converted to sst suggest temperatures ranging between 25 and 30°c although a number of sr/ca values yield unreasonably high ssts (figure 6 and figure 7) and there is no obvious pattern to suggest a seasonally related signal. high sr/ca ratios (low sst) do not correlate with low δ18oc as would be predicted if changes in sst were the dominate control on both proxies. 4 discussion the observation that changes in δ18oc covary with changes in δ13cc suggests that they are driven by a common mechanism— the strength of the indian monsoonal rainfall. since δ13c is not sensitive to changes in temperature, we conclude the most likely explanation is that isotopic changes in both the lia and modern corals are due to changes in δ18o and δ13c in the local seawater during increased riverine discharge (freshening) resulting from the monsoon rains. since δ18o in seawater averages about 0‰ and river as well as rain values are in the range of -5 to 10‰, mixing these two endmembers provides a first order explanation to our coral results. δ18ow in the bay of bengal strongly correlates with salinity changes in the region, rather than purely sst[1]. from the equation derived from 18ow and salinity data in the bay of bengal (equation 3) a change of 1.2 ‰ in 18ow would represent a change in salinity of about 6 p.s.u. (figure 6), this range would likely be beyond the tolerance limits for a coral and its symbionts. we cannot, however, discount that some of the seasonal variability in δ18oc is due to changes in sst, as suggested by the sr/ca data. we therefore conclude that about 50% of the 1.2‰ amplitude change in δ18oc is due to seasonal changes in sst (3°c) and 50% due to salinity changes of about 3 p.s.u.— the effects are additive. higher ssts during summer warming serve to lower δ18oc due to a decrease in isotopic fraction with increasing temperatures. increased precipitation and riverine input during the summer monsoon both add water with lower δ18o and thus serves to lower δ18ow and hence δ18oc. lower δ13c is associated with lower δ18o and may indicate increased riverine input of low δ13c in total dissolved inorganic carbon (dic). seasonal variation of around 2‰ in δ13cc may also reflect changes in coral metabolism related to feeding strategy[2]. for example, during the winter months cloud cover is above: table 1: mean and standard deviation of the fossil coral isotope values below: table 2: mean and standard deviation of the living coral isotope values equation 3: δ18oseawater (‰) = 0.18 × sss(p.s.u) 5.9(‰/p.s.u) p.s.u – practical salinity unit (3) aresty rutgers undergraduate research journal, volume i, issue ii reduced and photosynthesis enhanced by the symbiotic zooxanthellae. during photosynthesis, 12c is favored over 13c so the pool of dic becomes enriched in 13c thereby increasing δ13cc during warmer, sunny months. during the summer monsoon cloud cover is increased and photosynthesis reduced. corals may rely more on heterotrophic feeding of zooplankton and δ13cc reflects incorporation of a pool of low 13c (e.g. -25‰). the offset between 13cc in the modern versus the lia coral is significant (1.3‰) and may reflect changes brought about by increased and/or changing agriculture. specifically, increased rice production in this region during the 20th century would have added a pool of lower 13c (~-15‰) to the riverine total dic pool. certain values of sr/ca corresponding to unrealistic temperatures may be due to local effects, such as changes in the sr/ca delivered to the study area by rivers or perhaps due to changes in the rates of calcification in the coral. 5 conclusions stable isotope records obtained in both lia and modern corals at st. martin’s island, bangladesh display changes at monthly resolution, driven by the seasonal monsoon. we estimate variability in 18oc is split equally between changes in sst and changes in 18ow. the average δ18oc in the d09-01 fossil coral is only 0.2‰ higher compared to the living and leads top: figure 6: sr/ca isotope ratios from lia sample (transect 3). several values yield extreme values in sr/ca (e.g. 7 to 8) middle: figure 7: temperatures derived from the sr/ca ratios using equation 2 in the lia coral (transect 3). note: very low sr/ca (figure 4) translate to high and extreme and unreasonable estimates of ssts (e.g. > 32°c). bottom: figure 8: note the linear and positive correlation between δ18o of seawater with salinity, obtained from the bay of bengal. the slope suggests a change of about 0.2‰ per one unit change in salinity[1]. aresty rutgers undergraduate research journal, volume i, issue ii us to conclude the monsoons were only slightly weaker during the little ice age. that is, a combination of lower ssts and/or increased salinity (decreased freshening) could explain the differences between the living and lia coral δ18o records. longer records in both 18oc and sr/ca will be needed to make a statistically significant comparison between the two climate states. our pilot study shows the potential for generating century-long historical records of the monsoon, via records of stable isotopes and trace metals. this would provide much needed spatial and temporal resolution for climate models and climate forecasting. indeed, monsoons arise in any location where a strong land-sea contrast is present, from south asia and northern australia to west africa and southwestern north america—bringing to light how paleoclimate research on the monsoons would better quantify as well as benefit our understanding of the world’s climate and its impact. further work should include obtaining seawater samples for measurements of δ18o, salinity, sr/ca, and sr-isotope data at st. martin’s island along with measurements of trace metal concentrations in porites corals that might serve as a “fingerprint” for identifying riverine discharge∎ 6 acknowledgements i would like to express my gratitude to advisor, dr. mortlock for all his support and encouragement. thank you to the aresty research center; dr. james wright; reference librarian, maria ortiz-myers; dr. kaixuan bu; jennifer nemes, jenn pereira, chloe and the team of expert techs of the radiology dept. in rwjuh; and mark yu for all your help as well. samples provided through funding from nsf grants to dr. mchugh: oise 09-68354, onr n00014-11-1 0683. sr/ ca analyses were funded by aresty fellowship award (20182019). 7 references [1] delaygue, g., bard, e., rollion, c., jouzel, j., stiévenard, m., duplessy, j.-c., and ganssen, g., 2001, oxygen isotope/salinity relationship in the northern indian ocean: journal of geophysical research: oceans, v. 106, no. c3, p. 4565-4574. [2] grottoli, a. g., and wellington, g. m., 1999, effect of light and zooplankton on skeletal δ13c values in the eastern pacific corals pavona clavus and pavona gigantea: coral reefs, v.18, no. 1, p. 29-41. [3] isotope tracers in catchment hydrology (1998), c. kendall and j.j. mcdonnell (eds.). elsevier science b.v., amsterdam. pp. 51-86. [4] mondal, d. r., mchugh, c. m., mortlock, r. a., steckler, m. s., mustaque, s., and akhter, s. h., 2018, microatolls document the 1762 and prior earthquakes along the southeast coast of bangladesh: tectonophysics, v. 745, p. 196-213. [5] ramos, r. d., goodkin, n. f., siringan, f. p., and hughen, k. a., 2017, diploastrea heliopore sr/ca and δ18o records from northeast luzon, philippines: an assessment of interspecies coral proxy calibrations and climate controls of sea surface temperature and salinity: paleoceanography, v. 32, no. 4, p. 424-438. [6] schrag, d. p., 1999, rapid analysis of high-precision sr/ca ratios in corals and other marine carbonates: paleoceanography, v. 14, no. 2, p. 97-102. [7] weber, j. n., and woodhead, p. m. j., 1972, temperature dependence of oxygen-18 concentration in reef coral carbonates: journal of geophysical research, v. 77, no. 3, p. 463-473. aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. a systematic analysis of compounds present in ocimum tenuiflorum (tulsi) regarding its anti-inflammatory properties using in-silico techniques jinay patel, sonia arora (faculty advisor) ✵ abstract the objective of this study was to gather data, create a database of the compounds present in ocimum tenuiflorum (o. tenuiflorum), and gather related literature on the compounds found. a thorough literature search was performed to gather information on compounds present in o. tenuiflorum, including chemical structures, relative abundance, presence in different plant parts, and availability from chemical supply vendors. the compounds’ chemical structures were refined using discovery studio visualizer and chimera software for future insilico docking studies. the structures with cleaned structural geometry were obtained through d.s. visualizer for docking in the future. from the literature search of previously presented articles, it was found that methyl eugenol had the greatest percent composition in o. tenuiflorum. after searching the protein data bank, cox-1, cox-2, and nf kappa b were found to be the main protein targets of o. ten uiflorum compounds in the arachidonic acid inflame matory pathway. thus, the anti-inflammatory properties of o. tenuiflorum have been analyzed in this article for future in silico docking. 1 introduction tulsi is a plant of the species ocimum, scientifically known as ocimum tenuiflorum l. and more commonly known as “the holy basil.”[5] it is native to india and parts of north and eastern africa, china, hainan island, and taiwan, where it is also titled as "the elixir of life" or "the queen of herbs."[5] numerous parts of o. tenuiflorum, such as the leaves, stems, and flower spikes, are traditionally used in ayurveda and siddha medicine for treating conditions such as coughs, bronchitis, fever, and bile disturbances.[5] ayurveda and siddha medicines are two of the most ancient medicinal branches of india based on herbal and mineral compounds. o. tenuiflorum has also been vividly known for its anti-inflammatory, antiseptic, and other numerous organ protective properties.[5] based on its prominent medicinal uses, the ocimum tenuiflorum was chosen for further investigation. in this study, the focus was on inflammation because it is a well-known symptom of numerous infectious and non-infectious diseases. various in vitro and in vivo studies have documented the anti-inflammatory effects of o. tenuiflorum, and it has been suggested that o. tenuiflorum has many bioactive secondary metabolites that help inhibit certain inflammatory pathways synergistically or alone.[6] arachidonic acid is the major polyunsaturated fatty acid present in mammalian systems, which is oxygenated by three important pathways— the cyclooxygenase (cox), the lipoxygenase (lox), and the epoxygenase pathways.[10] the inhibition of cox and lox pathways in arachidonic acid metabolism in conjunction with major o. tenuiflorum compounds like eugenol and linoleic acid contributes to the anti-inflammatory action of o. tenuiflorum, observed in both acute and chronic inflammatory models in animals.[4] this observation suggests that certain compounds present within o. tenuiflorum are responsible for anti-inflammatory responses in ani aresty rutgers undergraduate research journal, volume i, issue iii mal models. this implication can be further explored to determine which compounds in o. tenuiflorum may induce anti-inflammatory activity in humans. as a result, o. tenuiflorum has demonstrated anti-inflammatory effects normally observed in nonsteroidal anti-inflammatory drugs (nsaid) such as phenylbutazone, ibuprofen, naproxen, aspirin, and indomethacin.[4] the inhibition of the cyclooxygenase enzymes cox-1 and cox-2 is what drives the nsaid mechanism.[1] although nsaids are effective, they have a few common side effects, such as nausea, dyspepsia, vomiting, and skin reactions. on the contrary, herbal medications are devoid of such side effects, which illustrates the importance of conducting a study of the compounds found in o. tenuiflorum: it may open the path for herbal nsaids without these side effects.[1] o. tenuiflorum has various transcription factors, one of which is the nuclear factor-kappa b (nf kappa b). nf kappa b is one of the major regulators of inflammation, cellular transformation, tumor cell survival, proliferation, invasion, angiogenesis, and metastasis as compared to the other transcription factors.[11] plants of the ocimum species are known to have an abundance of terpenes.[7] additionally, plant isolates containing terpenoids (a modified class of terpenes) have been found to suppress nf kappa b signaling, a protein complex linked to the pathogenesis of inflammatory diseases, cancer, viral infection, and autoimmune diseases.[7] thus, the study of o. tenuiflorum for anti-inflammatory properties is vital. there are numerous compounds present in the ocimum species amongst which methyl eugenol, β-selinene, γ-murolene, rosmarinic acid (phenolic), ursolic acid, and camphene are some major ones based on their percentage of makeup. these and other compounds from o. tenuiflorum have been characterized and analyzed in this study to form an extensive database. this database is used to determine which of the compounds identified would interact with desirable molecular targets and decrease inflammation in future studies. therefore, in this study, an in-silico approach was used to refine the compounds present in o. tenuiflorum and to develop a highly detailed database to be used for future purposes, such as investigating how an individual compound from o. tenuiflorum extract will interact with these molecular targets. this interaction would be detailed using the in-silico docking mechanism, which is a molecular modeling technique that is used to predict how a protein (enzyme) interacts with small molecules (ligands) to deduce their mechanism of action on targets found in future dry labs. 2 methodology first, a literature search was conducted through pubchem to create a list of well-known compounds present in o. tenuiflorum and their percent composition in the part of the plant they are found in.[8] the vendors that distribute the compound in the usa were identified from pubchem as well (table 1).[8] next, the collection of the virtual structures of all the compounds was created and the files were saved in the sds (2d) formats (figure 1). then, within the structural data, the pharmacology and biochemistry of the compounds were reviewed for each compound from pubchem in order to get an idea of the mechanism of action and the aspects of human interaction associated with the individual compounds.[8] this was an adaptation process, so new compounds were added to the list when and if they were discovered throughout the study. later, software such as discovery studio (d.s.) visualizer (biovia dasasult systemes, san diego, ca, usa) and chimera were downloaded.[9] chimera was used to convert the mol format files into sybyl mol 2 format files for compatibility within d.s. visualizer. the sybyl mol 2 format files were then opened in d.s. visualizer, wherein hydrogen atoms were added to the structure. the geometry of the hydrogen atoms was cleaned and 3d coordinates were assigned to them. the three-dimensional structure was then saved as sybyl mol 2 format files (figure 2). the formats sybyl mol 2 and sds were employed to have a better approximation of the structure of the compounds, and to notice the 2d and 3d differences in the molecules, which would visually assist in the molecular docking. and finally, another literary search was conducted on the protein data bank for the inflammatory molecular pathways and o. tenuiflorum within aresty rutgers undergraduate research journal, volume i, issue iii compounds percentage found vendors part of plant 1. 1,8-cineole trace(t) all leaves 2. 1,10 di-epi-cubenol 1.8 n/a leaves 3. 3,4-dimethoxycinnamic acid (phenolic) n/a all n/a 4. 3.4.5-trimethoxycinnamic acid n/a 1,2,4 n/a 5. 4.4′-methylene-bis (2-methyl aniline) n/a n/a n/a 6. α-elemene 0.5 2 flower spikes 7. α-humulene 0.2 4 leaves 8. α-pinene 0.2 2 leaves 9. α-terpineol trace(t) 1,2,3 leaves 10. alloaromadendrene 1.16 n/a n/a 11. apigenin n/a 2 leaves 12. β-bourbonene 0.2 n/a leaves 13. β-cubebene 0.1 2,4 flower spikes 14. β-pinene 0.1 2,3 leaves 15. β-selinene 3.3 2,4 leaves 16. βsesquiphellandrene 0.2 n/a leaves, flower spikes 17. baicalin (lavonoids) n/a 2,4 n/a 18. benzaldehyde 0.44 2,4 leaves 19. bicyclogermacrene trace(t) n/a leaves 20. caffeic acid (phenolic) n/a all leaves 21. camphene 0.1 1,2,3 leaves 22. camphor 0.1 2,3 leaves 23. carnosic acid n/a 2,4 n/a 24. chrysoeriol (flavon) n/a 1,2,4 n/a 25. ci α-copaene 1.9 n/a leaves, flower spikes 26. ci α-cubebene trace(t) 4 flower spikes 27. ci β-caryophyllene 4.1 2,3,4 leaves 28. ci β-gurjunene trace(t) n/a leaves 29. ci borneol 2.4 2 leaves 30. ci terpin-4-ol 0.1 2,3 leaves, stem 31. cubebol 0.3 n/a leaves 32. di (ethylhexyl) phthalate n/a 1,2 n/a 33. di-n-butyl phthalate, dibutyl phthalate n/a 1,2,4 leaves table 1: this table shows various compounds found in specific parts of o. tenuiflorum such as the leaves, stem, or flower spikes. it also includes their percentage composition, the vendors who supply them in the usa, and their pubchem sid and purchasable chemical id. some of the compounds have trace(t) in their “percentage found” column, which implies that the compound is present in a quantity too small to be measured. pubchem[8] aresty rutgers undergraduate research journal, volume i, issue iii compounds percentage found vendors part of plant 34. diosmetin (flavone glycoside) n/a 1,2,4 n/a 35. (e)-α-bergamotene 0.72 2,4 leaves, flower spikes 36. (e)-β-ocimene 1.9 2,4 leaves 37. e-methyl cinnamate 1.5 all leaves, stem 38. epi-α-cadinol 1.03 n/a leaves 39. estragol n/a 1,2,4 leaves, flower spikes 40. eugenol 0.9 all leaves 41. γ-murolene 5.82 2 leaves 42. germacrene a 0.7 n/a leaves 43. germacrene d 2.3 2,4 leaves 44. geraniol trace(t) all leaves 45. globulol 1.05 2,4 n/a 46. isosakuranetin (lavanone) n/a 4 n/a 47. kaempferol n/a all n/a 48. linalool 0.5 3 leaves 49. limonene 0.2 1,2 leaves 50. luteolin (flavonoid) 0.5 n/a leaves 51. methyl chavicol trace(t) 1,2,3 leaves 52. methyl eugenol 82.9 2,3,4 leaves 53. myrtenal n/a 2 leaves 54. nevadensin (flavones, glycosides) n/a 2,4 n/a 55. p-cymene trace(t) 1,2,3 flower spikes 56. pedunculin n/a n/a n/a 57. permethrin n/a 1,2 n/a 58. peinidin (anthocyanidins) n/a n/a n/a 59. rosmarinic acid (phenolic) 0.27 2,4 leaves, stems 60. sabinene trace(t) 2 leaves 61. δ-cadinene 1.1 2,4 leaves 62. (trans)-β-farnesene 0.12 2,4 leaves 63. (trans)-β-guaiene 0.29 n/a leaves 64. ursolic acid 2.5 1,2,4 leaves 65. xanthomicrol n/a 2 stem 66. (z)-3-hexanol 1.8 1,2 leaves vendors 1) tim tec 2) musechem 3) acros organics 4) zinc table 1 continued aresty rutgers undergraduate research journal, volume i, issue iii the context of inflammation and its molecular targets concerning humans.[2] this search was conducted to determine which molecular target should be further used to analyze the docking compatibilities of the compiled o. tenuiflorum compounds. 3 results a detailed database was created in table 1 to compile an extensive list of 66 compounds in o. tenuiflorum. table 1 also displays information about the percentage of a particular compound present in certain parts of o. tenuiflorum, such as the leaves, stem, flower spikes, etc. in table 1, it is evident that most of the compounds were available from the list of usbased vendors like tim tec, musechem, acros organics, and zinc. the pubchem sid and purchasable chemical id were also available for all of the compounds. the highest percentage of composition for a compound in o. tenuiflorum was for methyl eugenol, about 82.9% within the leaves, which was available from musechem, acros organics, and zinc (table 1). overall, a total of 66 compounds were identified to be present in o. tenuiflorum. out of those compounds, many were in trace amounts and some did not have any data on the percentage of their presence (table 1). next, 2d structures in sds format files were obtained from pubchem in the form of 2d chemdraw files (figure 1). figure 1 shows the chemical structure of a few major compounds found in o. tenuiflorum. the compounds seen were not ready for docking since they had not been refined yet to allow for further in-silico processes due to their structural ambiguity. it would be difficult to dock them; therefore, further work on each compound was needed, such as adding hydrogens and removing unwanted ions (figure 1). next, the refined structures were obtained from d.s. visualizer in order to transfer the coordinates to proceed to the in-silico docking step (figure 2). the refined structure of several o. tenuiflorum compounds are portrayed in figure 2. finally, the literature search on the protein data bank yielded that cox-1, cox-2, and nf kappa b were the anti-inflammatory target proteins that interact with certain compounds found in o. tenuiflorum. however, within the protein data bank the cox-2 protein was the most prevalent target found in humans and cox-1 had the least human entries. these targets will be explored in future studies with the collected o. tenuiflorum database compounds. figure 1: the 2d structures of compounds present in o. tenuiflorum are shown here, such as ursolic acid, methyl eugenol, γ-murolene, rosmarinic acid, camphene, and β-selinene. some of the figures that were downloaded from pubchem portray the oxygen and hydroxide atoms in red fonts. figure 2: 3d structures of a few of the major compounds present in o. tenuiflorum, such as camphene, methyl eugenol, rosmarinic acid, β-selinene and γ-murolene with hydrogen atoms added and their geometry cleaned. aresty rutgers undergraduate research journal, volume i, issue iii 4 discussion & conclusion for this study, a thorough database was created for various compounds present in different parts of o. tenuiflorum. this database is being stored on the computers in the arora lab to be used for future studies. during the literary search of these structures, it was found that methyl eugenol was the most abundant compound and γ-murolene was the second most abundant compound. both compounds are present in the leaves of o. tenuiflorum. these compounds are significant because they are present in such high abundance, which suggests that they might have an important role in the anti-inflammatory effect of o. tenuiflorum. to test this statement, these structures will have to be run through insilico docking modules to determine whether the abundance has any correlation with the anti-inflammatory effect. moreover, the purpose of attaining the 2d files was to make the structures available before refinement so that they could be reviewed when required. the refined structures with cleaned geometry were obtained so that their docking sites were clear of any hindrance and to ensure that docking could be done properly in the future. also, the target proteins were identified from the protein data bank so that they could be used along with their ligands to dock. thus, the future goal is to use in-silico approaches to discover how compounds will interact with various molecular targets and inflammatory pathways, specifically cox-2, as it is the most prevalent target yielded in most searches related to human query within the protein data bank. the database gathered in this study will be crucial for future studies to identify which compounds present in o. tenuiflorum are attributed to its anti-inflammatory properties. this information could in turn allow for the development of herbal anti-inflammatory medication with minimal side effects∎ 5 acknowledgements i would like to thank my mentor and supervisor dr. sonia arora for her invaluable support and knowledge behind this paper and research. her passion for the subject, along with attention to detail has been greatly inspiring, allowing me to diligently work on this research project. her ingenious remarks and constructive criticism for various drafts of this paper has allowed for this research paper to be successfully completed. i will always be extremely grateful to dr. arora for imparting me with the knowledge and wisdom about in-silico approaches, which can be used to create comprehensive databases in any research work. 6 references [1] ahmad a., abuzinadah m.f., alkreathy h.m., banaganapalli b., & mujeeb m. (2018). ursolic acid rich ocimum sanctum l leaf extract loaded nanostructured lipid carriers ameliorate adjuvant induced arthritis in rats by inhibition of cox-1, cox-2, tnf-α and il-1: pharmacological and docking studies. plos one, 13(3), e0193451. [2] berman, h. m., westbrook, j., feng, z., gilliland, g., bhat, t. n., weissig, h., shindyalov, i. n., & bourne, p. e. (2020). the protein data bank. nucleic acids research, 28(1), 235–242. [3] biovia, dassault systemes, discovery studio visualizer, version 20.1, san diego: dassault systèmes, (2020). [4] cohen m. m. (2014). tulsi—ocimum sanctum: a herb for all reasons. journal of ayurveda and integrative medicine, 5(4), 251–259. [5] flegkas a., milosević ifantis t., barda c., samara p., tsitsilonis o., & skaltsa h. (2019). antiproliferative activity of (-)rabdosiin isolated from ocimum sanctum l. medicines, 6(1), 37. [6] jamshidi, n., & cohen, m. m. (2017). the clinical efficacy and safety of tulsi in humans: a systematic review of the literature. evidence-based complementary and alternative medicine. evidence-based complimentary and alternative medicine: ecam, 2017, 9217567. [7] kapewangolo, p., omolo, j. j., bruwer, r., fonteh, p., & meyer, d. (2015). antioxidant and anti-inflammatory activity of ocimum labiatum extract and isolated labdane diterpenoid. journal of inflammation (london, england), 12, 4. [8] kim, s., chen, j., cheng, t., gindulyte, a., he, j., he, s., li, q., shoemaker, b. a., thiessen, p. a., yu, b., zaslavsky, l., zhang, j., & bolton, e. e. (2019). pubchem 2019 update: improved access to chemical data. nucleic acids research, 47(d1), d1102–d1109. [9] pettersen ef, goddard td, huang cc, couch gs, greenblatt dm, meng ec, ferrin te. (2004). ucsf chimera—a visualization system for exploratory research and analysis. j comput chem. 25(13):1605-1612. [10] reddy, k. k., vidya rajan, v. k., gupta, a., aparoy, p., & reddanna, p. (2015). exploration of binding site pattern in arachidonic acid metabolizing enzymes, cyclooxygenases and lipoxygenases. bmc research notes, 8, 152. [11] yadav, v. r., prasad, s., sung, b., kannappan, r., & aggarwal, b. b. (2010). targeting inflammatory pathways by triterpenoids for prevention and treatment of cancer. toxins, 2(10), 2428–2466. aresty rutgers undergraduate research journal, volume i, issue iii jinay patel is rutgers university graduate, having b.s degree in biotechnology and plant science from school of environmental and biological science. his research was on o. tenuiflorum to gather the data on the compounds present within it and study its anti-inflammatory properties using insilico techniques. he is currently working as a qc analyst at roche molecular systems within enzymatic department and in the future, he would like to attend graduate school and study drug development. education in drug development for him would entail discovering new medicines and treating diseases to support better patient care. the following research on o. tenuiflorum is one of the crucial steps for him in attaining this goal. aresty rutgers undergraduate research journal, volume i, issue ii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. control of satellites with onboard robotic manipulators robert buckelew, ethan catalanello, annalisa scacchioli (faculty advisor) ✵ abstract free-floating satellites with onboard robotic manipulators are subjected to widely varying loads resulting from the motion of the robotic manipulators. as there are no fixed supports in space, these loads will cause the satellite to move. by modelling the motion of the onboard robotic arms, determining the necessary reaction loads (which must be supplied by the satellite to keep the arm fixed), and simulating the resulting satellite dynamics, we designed a model of a satellite-arm system. we found that a proportional-integral-derivative (pid) control scheme, with disturbance-estimating capabilities, was effective in maintaining satellite position and orientation during the operation of the onboard robotic manipulator. the matlab-based simulink modeling environment was used to perform the simulations of satellite dynamics and control. 1 introduction the operational success of several fields, including communication technology and space exploration, is dependent on a functional space satellite infrastructure. an advanced satellite network, such as a constellation of communications satellites, will require routine maintenance to maximize its functional lifetime[1]. the defense advanced research projects agency (darpa) has proposed technology to perform robotic servicing of on-orbit satellites[2]. the project is titled robotic servicing of geosynchronous satellites (rsgs) and involves satellitemounted robotic arms which can perform operations on other satellites. potential missions for the “nurse” satellite include inspection, installation, and repair of on-orbit satellites as well as their relocation to new orbits. in general, the motion of robotic manipulators and the body to which they are mounted are coupled—that is, the motion of one tends to influence the motion of the other. the magnitude of this effect depends on the inertial properties of the bodies and the interaction forces involved. in a microgravity environment such as low earth orbit (leo), no dissipative reaction forces exist unlike those present in operations performed on and near the earth’s surface. therefore, the motion of satellite-mounted robotic manipulators will have significant effects on satellite attitude (the direction the satellite is pointing) [3]. it is desirable to reduce the effects of robotic manipulator disturbances on satellite position and attitude using an attitude control system (acs). an acs measures satellite position and orientation with respect to reference values. it uses mechanical actuators (e.g., reaction wheels, gas jet thrusters) to reject disturbances. in this project, we model the dynamics of the frend mark ii robotic arm and simulate the motion of a base satellite subjected to disturbances from the robotic arm motion. additionally, we designed a control scheme using a basic pid (proportional-integral-derivative) controller. pid controllers feature three gains—proportional, integral, and derivative—which determine the control input signal based on the magnitude, time integration, and time rate of change of the error signal respectively. the goal is to reduce the perturbations arising from operation of the robotic arm, resulting in a satellite that maintains its position and attitude. while extensive work has been performed in this area utilizing complex, nonlinear controls, we wanted to evaluate the effectiveness of a much simpler, linear control scheme, as it was assumed the perturbations arising from the robotic arm would be very small. we believed that these small perturbations would be well approximated by linear control theory, and thus that linear controllers would be sufficient at stabilizing the system. this belief/idea, combined with the aresty rutgers undergraduate research journal, volume i, issue ii development of the modular simulation for control design, serves as the novel contribution of our project. we have created an environment that allows for control refinement without requiring rigorous mathematical proofs. 2 methodology the robotic arm proposed by darpa for the robotic servicing of geosynchronous satellites (rsgs) program is the frend mark ii (hereafter referred to as the frend), a 7-joint 3d manipulator which will be responsible for performing operations on the “patient” satellite. an illustration of the frend arm is shown in figure 1. the robotics, vision, and control toolbox, a third-party matlab-based toolbox for robotics applications (hereafter referred to as “the toolbox”) was used to develop a model of the frend and to perform dynamics calculations. the results of these dynamics calculations, based on the frend model, were used as inputs to the satellite dynamics simulation. these results effectively modeled the disturbances to the satellite arising from the motion of the robotic arm. frend arm modelling the toolbox is capable of modelling and performing calculations on any robotic arm so long as the arm’s relevant parameters are provided. the relevant parameters are inertial parameters, including the mass and inertia matrix and denavit-hartenberg parameters (which describe arm geometry). the denavit-hartenberg parameters used in the frend model were defined according to the definitions shown in figure 2 and measurements taken from figure 3. the denavit-hartenberg parameters are shown in supplemental table s1. figure 1: darpa rendering of the frend mark ii arm[4] it is desirable to reduce the effects of robotic manipulator disturbances on satellite position and attitude using an attitude control system (acs). aresty rutgers undergraduate research journal, volume i, issue ii none of the required parameters for the frend were publicly accessible, so the measurements were approximated using pixel measurements from schematics and images that were publicly available. knowing this information, an image editing software can be used to analyze the schematic in figure 3 in order to define a pixel scale which relates real-world dimensions to pixel dimensions. the geometry of the arm was then calculated from the schematic in figure 3 and used to compute the inertia tensor. each link is assumed to be a cylinder of solid aluminum having a density of 2700 𝑘𝑘𝑘𝑘 𝑚𝑚3 and a center of mass located at the geometric center. this assumption of material composition represents a “worst case scenario” as the actual arm is not solid aluminum[4]. the computed dynamic parameters are shown in supplemental table s2 and then inertia matrices are shown in supplemental table s3. above: figure 2: graphic definition of denavit-hartenberg parameters[5] right: figure 3: schematic of the frend mark ii showing the relative orientations of link reference frames[4] aresty rutgers undergraduate research journal, volume i, issue ii [a] gravity was neglected in our simulation because the satellite is in space. the toolbox features a function that can numerically evaluate the robot dynamics equation (equation (1)). this is a system of coupled differential equations that relates the joints’ angular kinematics (𝑞𝑞, �̇�𝑞, �̈�𝑞) of the robotic arm joints to the joint torques (𝑄𝑄) and other external loads (end effector wrench 𝑊𝑊 and gravity torque 𝐺𝐺(𝑞𝑞)[a]) acting on the arm. solving these differential equations yields the joint torques and the reaction loads at the base of the arm. 𝑄𝑄(𝑞𝑞, �̇�𝑞, �̈�𝑞) = 𝑀𝑀(𝑞𝑞)�̈�𝑞 + 𝐶𝐶(𝑞𝑞, �̇�𝑞)�̇�𝑞 + 𝐹𝐹(�̇�𝑞) + 𝐺𝐺(𝑞𝑞) + 𝐽𝐽(𝑞𝑞)𝑇𝑇𝑊𝑊 the parameters of equation (1) are defined in table 1 below. the joint torques 𝑄𝑄(𝑞𝑞, �̇�𝑞, �̈�𝑞) obtained through the solution of the differential equation can be utilized to solve for the reaction forces as well as the moments the arm will induce on the satellite. thus, these loads will be utilized in the simulation as the external disturbance the satellite is subjected to. satellite modelling to model the effects of robotic arm motion on satellite position and attitude, we utilized the aerospace blockset’s “6dof” (six degree-of-freedom) block in the matlab-based simulink modeling environment. this block implements the equations of motion described in equations (2) and (3), which are the differential equations describing translational motion and angular motion, respectively. the parameters of these equations are defined explicitly in supplemental table s5. �̈�𝑥 = ∑ 𝐹𝐹𝑗𝑗(𝑡𝑡)𝑁𝑁𝑗𝑗=1 𝑚𝑚 �̈�𝜃(𝑡𝑡) = 𝐼𝐼−1 ���𝑀𝑀(𝑡𝑡) + 𝑟𝑟𝑗𝑗 × 𝐹𝐹𝑗𝑗(𝑡𝑡)� − �̇�𝜃(𝑡𝑡) × �𝐼𝐼�̇�𝜃(𝑡𝑡)� 𝑁𝑁 𝑗𝑗=1 � 𝑞𝑞, �̇�𝑞, �̈�𝑞 joint angular position, velocity, and acceleration respectively 𝑄𝑄 vector of generalized joint torques – torques required at each joint to achieve the motion described by (𝑞𝑞, �̇�𝑞, �̈�𝑞) 𝑀𝑀(𝑞𝑞) joint-space inertia matrix – describes the moment of inertia of each link in matrix form 𝐶𝐶(𝑞𝑞, �̇�𝑞) coriolis and centripetal coupling matrix – describes centripetal forces due to rotation of the links 𝐹𝐹(�̇�𝑞) friction matrix – describes torques from friction present in the joints 𝐺𝐺(𝑞𝑞) gravity load – describes torques originating from gravity acting on the links 𝐽𝐽(𝑞𝑞) manipulator jacobian matrix – describes the sensitivity of the end effector to the motion of each joint 𝑊𝑊 end effector wrench – describes the load on the end effector table 1: definition of terms in equation (1) (1) (2) (3) aresty rutgers undergraduate research journal, volume i, issue ii since there is no satellite design specified for the rsgs program, we assumed a generic small satellite configuration. we take the satellite to be a rectangular prism of 2𝑚𝑚 × 1𝑚𝑚 × 1𝑚𝑚 with two thin rectangular solar arrays fixed at opposing ends 3𝑚𝑚 × 1.5𝑚𝑚 × .0055𝑚𝑚 located . 5𝑚𝑚 from the satellite body. the density of the satellite was assumed to be 200 𝑘𝑘𝑘𝑘 𝑚𝑚3 , which is an approximate mass density for spacecraft of this size[9] and the solar panels were assumed to be identical to those manufactured by spectrolab, a manufacturer of satellites for use in space, and have a planar density of 2.06 𝑘𝑘𝑘𝑘 𝑚𝑚2 [10]. this means that the entire satellite has a mass of 𝑚𝑚 = 411.9305362785 𝑘𝑘𝑘𝑘, and the inertia tensor is shown below: 𝐼𝐼 = (303.02687565826 0 0 0 170.14297228667 0 0 0 199.55068127826 )𝑘𝑘𝑘𝑘 ∙ 𝑚𝑚2 controller modelling recall that our objective is to maintain the translational position as well as the angular position (or attitude) of the satellite throughout operation of the on-board robotic arm. to accomplish satellite position and attitude control, pid controllers, which represents a basic function for the control of linear systems, were implemented for each of the six degrees of freedom, namely the three translational coordinates (𝑥𝑥, 𝑦𝑦, 𝑧𝑧) of satellite position and the three euler angles (𝜙𝜙, 𝜃𝜃, 𝜓𝜓) defining satellite roll, pitch, and yaw, respectively. the optimal values of the pid controller gains, 𝐾𝐾𝑃𝑃, 𝐾𝐾𝐼𝐼, and 𝐾𝐾𝐷𝐷 are determined automatically using the matlab pid tuner tool. the optimal gains for the rsgs model satellite are shown in supplemental table s4. 3 results as a result of this project, we have developed a multiple component model in which a user can define a robotic arm configuration and trajectory as well as the inertial parameters of a satellite. the user can then simulate the motion response of the satellite to disturbances arising from motion of the arm and employ a predictive control scheme to maintain a reference satellite position and attitude. a simplified simulink block diagram demonstrating the fundamental logic behind the model is shown in figure 4. as shown in the figure, the controller (which is a pid controller in this case) receives the translational position and angular position offset as well as an estimation of the incoming arm loads which it then uses to calculate the necessary control force/torque. 𝑥𝑥(𝑡𝑡) vector of satellite positional coordinates 𝜃𝜃(𝑡𝑡) vector of satellite angular position coordinates 𝐹𝐹𝑗𝑗(𝑡𝑡) vector of translational force caused by robotic arm 𝑗𝑗 𝑀𝑀𝑗𝑗(𝑡𝑡) vector of moments caused by robotic arm 𝑗𝑗 𝑚𝑚 satellite mass 𝐼𝐼 satellite inertia tensor 𝑟𝑟𝑗𝑗 vector defining the base of arm 𝑗𝑗 relative to the satellite com table 2: definition of terms in equations (2) and (3) aresty rutgers undergraduate research journal, volume i, issue ii the objective of this work is to demonstrate the use of a pid controller with load estimative capacity in the control of position and attitude of satellites with onboard robotic manipulators. the simulation was run for multiple scenarios: without control, with non-predictive control, and with the load estimating controller. for runs with the controller, it is assumed that the satellite has an ideal control actuator onboard that can instantly induce the required torques or forces determined by the pid controller. it is important to note that this is a preliminary assumption used to evaluate the effectiveness of the control scheme rather than to evaluate its performance in a realistic application. the position and at left: figure 4: simplified block diagram describing the logical relations between the disturbance load, satellite, and controller below: figure 5: disturbance forces and moments acting on the satellite throughout the arm’s trajectory the forces and moments of the satellite disturbance load are plotted against time in figure 5. aresty rutgers undergraduate research journal, volume i, issue ii titude responses of the satellite both with and without control are compared in figure 6. the left side of figure 6 shows the uncontrolled case: notice that the angular and translational positions diverge from the reference (initial condition). the controlled case is shown on the right side of figure 6. notice how the satellite’s deviation from the reference was drastically reduced compared to the uncontrolled case and that the satellite returned to the initial reference state (position of (0, 0, 0) and angular position of (0, 0, 0)) after the disturbance loads ceased. it is important to note that this simulation represents a “worst case scenario” for the satellite, as several assumptions were made that negatively impact the control effectiveness. despite these assumptions, the pid controller was very effective at stabilizing the system. thus, we have succeeded in designing an effective controller for the satellite. 4 discussion of simulation results the use of the proportional-integral-derivative control scheme with load prediction resulted in a very accurate satellite response. the maximum deviations from reference position and attitude were . 07 𝑚𝑚𝑚𝑚 and 1.26° respectively. without the use of the controller, the satellite motion would continue indefinitely following a disturbance. these results provide evidence that such a control scheme could effectively maintain satellite position and attitude during robotic servicing operations. the use of calculated robotic arm disturbances as load predictions is a reasonable assumption when considering the actual implementation of robotic servicing missions. if a space organization were to implement satellitemounted robotic arms for maintenance purposes, figure 6: comparison of system response between uncontrolled (left) and controlled (right) satellite position (top) and angular position (bottom) aresty rutgers undergraduate research journal, volume i, issue ii the simulation of the robot dynamics prior to the execution of the mission would result in a more accurately controlled system during operation, as the result of the simulation would be used as the load estimation. in addition to the load prediction, the pid controllers add robustness to the scheme, adjusting for additional perturbations arising from the operational environment. 5 conclusions and future work the goal of this project was to create a dynamics simulation of a robotic manipulator and satellite in order to develop a model-based control algorithm for maintaining satellite position and orientation. even with a very basic pid feedback control scheme, the results of the simulation were very promising. the arm load estimation technique allowed the controller to effectively eliminate translational and angular displacement. we believe that this research offers a framework for future investigation into more advanced control methods, including nonlinear approaches such at lyapunov theory, which incorporates the nonlinear system model into the controls equation. we also hope to develop a more accurate satellite model that includes realistic actuators, such as reaction wheels and gas thrusters, to better evaluate the effectiveness of the control scheme. with this, we also hope to investigate various satellite/arm designs and to determine which configuration is most effective. this was inspired by dapra’s robotic servicing of geosynchronous satellites (rsgs)[4] proposal, which aims to extend the operational lives of satellites located in geostationary orbit through issue diagnosis, repair, and upgrades. the implications of this proposal include lowering the cost of operation for geosynchronous satellites, thereby encouraging more private companies to contribute to space infrastructure. this, in turn, aligns closely with the nasa space technology mission directorate[18], as contributing to this earth-based infrastructure will facilitate future missions carried out to the moon any beyond. we hope future work on this project will incentivize more private companies to begin developing autonomous, robotic satellite systems to further humanities reach beyond planet earth∎ 6 acknowledgements the authors would like to acknowledge the help of others who contributed to the success of this project. we thank our aresty research advisor, dr. annalisa scacchioli, for her continuous guidance and encouragement throughout the duration of the project. we thank dr. aaron mazzeo for introducing us to the robotics toolbox. we thank dr. laurent burlion for his discussion about control methodologies and satellite attitude dynamics. we thank mike vivar for his recommendation of the robotics toolbox. finally, we thank our undergraduate peers, especially michael higgins and aaron pfister, for their continued interest in our project. 7 references [1] f. garcia, l. peret, g. verfaillie, deployment and maintenance of a constellation of satellites: a benchmark, 2003. [2] defense advanced research projects agency. robotic servicing of geosynchronous satellites (rsgs). https://www.darpa.mil/program/robotic-servicing-of-geosynchronous-satellites [3] a. antonello, p. tsiotras, a. valverde, free-flying spacecraftmounted manipulators: a tool for simulating dynamics and control, 2019. [4] defense advanced research projects agency. (2019, may 22). robotic servicing of geosynchronous satellites proposers day. [5] p.i. corke, “robotics, vision & control”, springer 2017, isbn 978-3-319-54413-7. [6] mathworks. 6dof (euler angles). https://www.mathworks.com/help/aeroblks/6dofeulerangles.html [7] p.c. hughes, “spacecraft attitude dynamics”, dover publications 2004, isbn 0-486-43925-9 [8] northrop grumman. space logistics services. https://www.northropgrumman.com/space/space-logistics-services/ [9] phipps, claude & albrecht, g. & friedman, h. & gavel, d. & george, e. & murray, j. & ho, chunching & priedhorsky, w. & michaelis, m.m. & reilly, j.p.. (1996). orion: clearing near-earth space debris using a 20-kw, 530-nm, earth-based, repetitively pulsed laser. march 1996. [10] spectrolab. space solar panels. https://www.spectrolab.com/datasheets/panel/panels.pdf [11] mathworks. open pid tuner. https://www.mathworks.com/help/slcontrol/ug/designing-controllers-with-the-pid-tuner.html [12] g.f. franklin, j.d. powell, a. emami-naeini, “feedback control of dynamic systems”, pearson 2015, isbn 978-0-13349659-8 [13] mathworks. sgolayfilt. https://www.mathworks.com/help/signal/ref/sgolayfilt.html https://www.darpa.mil/program/robotic-servicing-of-geosynchronous-satellites https://www.darpa.mil/program/robotic-servicing-of-geosynchronous-satellites https://www.mathworks.com/help/aeroblks/6dofeulerangles.html https://www.northropgrumman.com/space/space-logistics-services/ https://www.northropgrumman.com/space/space-logistics-services/ https://www.spectrolab.com/datasheets/panel/panels.pdf https://www.mathworks.com/help/slcontrol/ug/designing-controllers-with-the-pid-tuner.html https://www.mathworks.com/help/slcontrol/ug/designing-controllers-with-the-pid-tuner.html https://www.mathworks.com/help/signal/ref/sgolayfilt.html aresty rutgers undergraduate research journal, volume i, issue ii [14] c.j. damaren, a.h.j. de ruiter, j.r. forbes, “spacecraft dynamics and control: an introduction”, wiley 2013, isbn 9781-118-34236-7 [15] s.a. rawashdeh, attitude analysis of small satellites using model-based simulation, international journal of aerospace engineering, 2019. [16] r. hernandez-alvarado, l.g. garcia-valdovinos, et. al., neural network-based self-tuning pid control for underwater vehicles, sensors, 2016. [17] h.e.soken, c. hajiyev, s. sakai, robust kalman filtering for small satellite attitude estimation in the presence of measurement faults, european journal of control, 2013. [18] nasa. space technology mission directorate. https://www.nasa.gov/directorates/spacetech/about_us/index.html https://www.nasa.gov/directorates/spacetech/about_us/index.html aresty rutgers undergraduate research journal, volume i, issue ii 8 supplementary tables joint 𝑑𝑑 𝑎𝑎 𝛼𝛼 offset 1 0.3469785 0 − 𝜋𝜋 2 0 2 0.2517295 0 − 𝜋𝜋 2 0 3 0.8776515 0 𝜋𝜋 2 0 4 0.2122692 0 − 𝜋𝜋 2 0 5 0.1592019 0 � 127 449 � 0 5a 0.6349236746 0 −� 127 449 � 0 5b 0.1074953 0 𝜋𝜋 2 0 6 0.1728089 0 − 𝜋𝜋 2 0 7 0.2000229 0 0 0 link 𝑚𝑚 (𝑘𝑘𝑘𝑘) 𝑟𝑟 (𝑚𝑚) 𝐿𝐿 (𝑚𝑚) 1 2.45581930238976 0.0530673 0.3469785 2 1.78167282722394 0.0530673 0.2517295 3 8.64176380422475 0.0625922 0.8776515 4 0.802308387477559 0.03877995 0.2122692 5 0.540058000435033 0.0367389 0.1592019 5a 2.15384119253858 0.0367389 0.6349236746 5b 0.364654547302287 0.0367389 0.1074953 6 0.425389148502536 0.0312961 0.1728089 7 0.536125954214708 0.0326568 0.2000229 supplemental table s1: denavit-hartenberg parameters for the frend mark ii arm model. supplemental table s2: mass, radius, and length of each cylindrical link object used in the definition of the frend arm model aresty rutgers undergraduate research journal, volume i, issue ii link 𝐼𝐼 (𝑘𝑘𝑘𝑘 ∙ 𝑚𝑚2) 1 (0.00778041772567878 0 0 0 0.0116706265885182 0 0 0 0.00778041772567878 ) 2 (0.0056446167813748 0 0 0 0.0084669251720622 0 0 0 0.0056446167813748 ) 3 (0.0380886296066415 0 0 0 0.0571329444099623 0 0 0 0.0380886296066415 ) 4 (0.00135740156152532 0 0 0 0.00203610234228798 0 0 0 0.00135740156152532 ) 5 (0.00082005923636261 0 0 0 0.00085043340488162 0.00010738583988220 0 0.00010738583988220 0.001199714686025 ) 5a (0.003270532724590 0 0 0 0.003391669964095 − 0.00042827260265995 0 − 0.00042827260265995 0.004784661847380 ) 5b (0.000553715210877318 0 0 0 0.000830572816315976 0 0 0 0.000553715210877318 ) 6 (0.000468726352717391 0 0 0 0.000703089529076087 0 0 0 0.000468726352717391 ) 7 (0.000643230468209275 0 0 0 0.000643230468209275 0 0 0 0.000964845702313913 ) 𝐾𝐾𝑃𝑃 𝐾𝐾𝐼𝐼 𝐾𝐾𝐷𝐷 force 617.061827129926 35.6866537699081 2667.42085820047 moment 102.843637854979 5.94777562634487 444.570143033412 supplemental table s3: inertia tensors for each link of the frend mk. ii arm model supplemental table s3: proportional, integral, and derivative gains for the force and moment pid controllers aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. targeting the epigenetic factors of neuroinflammation anusha patil written under the direction of dr. david p. crockett abstract currently, there are no effective treatments for traumatic brain injury (tbi). this is because the mechanisms behind post-injury neuroinflammation are not well understood. this project studies a novel signaling pathway responsible for the activation of microglia post-tbi. its goal is to identify epigenetic factors of neuroinflammation that may be targeted with future therapies. we are examining the possibility that class iia histone deacetylases (hdacs), particularly hdac7, are responsible for initiating the inflammatory response after tbi. in addition, we plan to explore what its upstream regulation factors may be. one possible upstream regulation factor explored is regulating kinase 2, also known as par1b (par1b/mark2), since prior research indicates that a deficiency in par1b/mark2 increased the inflammatory response of microglia in a mouse model of tbi. in our experiments, we examined brains from sham mice (i.e., without head injury) and mice subjected to closed head injury (chi). our experiments made use of wild-type mice and mice deficient in par1b/mark2. qualitative analyses were conducted using fluorescent microscopy imaging of immunohistochemistry. using cell-specific markers of inflammation, we found an increase in astrocytic marker gfap (glial fibrillary acidic protein) and microglial protein iba1 (ionized calcium binding protein) expression in the cortex of the mice after chi. these increases were dramatic ipsilaterally (same side) to the injury, but only moderate contralaterally (opposite side). in the control brains (sham operates), little to no increase in these markers were detected. in our experiments, we observed increased expression of hdac7 in post-tbi microglia as well as in reactive gfap-expressing astrocytes. others have observed that par1b/mark2 may negatively regulate hdac7 activity and there is evidence that hdac7 is an inhibitor of antiinflammatory genes. https://creativecommons.org/licenses/by-nc-sa/4.0/ https://creativecommons.org/licenses/by-nc-sa/4.0/ 2 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 introduction traumatic brain injury (tbi) has been dubbed the “silent epidemic.” it is estimated that 69 million people worldwide suffer from death and disability caused by tbi, which is more than any other traumatic insult (dewan et al 2019). in the u.s. alone, approximately 1.7 million people sustain a tbi each year, which leads to at least 52,000 deaths annually. the financial burden on society in both direct and indirect costs adds up to an astounding $60 billion, not to mention the cost of the individual’s life or independence. unfortunately, there is currently no effective treatment for tbi. one major reason for this is the body’s inflammatory response that follows damage to the central nervous system (e.g., loane and byrnes 2010; faden and loane, 2015). neuroinflammation is a double-edged sword because it has both beneficial and detrimental effects post-tbi, and general immunosuppression has not proven helpful (russo and mcgavern, 2016; loane and byrnes 2010). neuroinflammation has also been implicated in many neurodegenerative diseases such as parkinson’s and alzheimer’s disease (faden and loane, 2015). fully understanding the mechanisms behind the body’s neuroinflammatory response will not only allow for the development of targeted treatments for tbi, but also will likely allow for better treatments for those with neurodegenerative diseases (ransohoff, 2016) and other diseases that may have a neuroinflammatory component such as diabetes (wang et al 2014). the epigenetic mechanisms that control gene expression of neuroinflammation have been studied, specifically the mechanism of histone deacetylases (hdacs). hdacs are a class of enzymes that modify chromatin conformation by removing acetyl groups from the histones of chromatin. this condenses chromatin, making it more difficult for the dna to be accessed by transcriptional machinery and therefore repressing transcription (allfrey et al 1964). in recent years, there has been a focus on hdacs as a therapeutic target to treat a variety of diseases including cancer, cardiovascular disease, stroke, injury and diabetes (yoon and eom, 2016; wang et al, 2014; mihylova et al, 2011; baltan et al, 2011). of the hdac enzymes, class iia hdacs receive particular attention given that they have higher expression in the brain and are involved in neuronal activity, degeneration, and apoptosis. class iia hdacs have been implicated as important factors in regulating the immune response (altman and kong, 2011; poralla et al, 2015; shakespear et al, 2013; wang et al, 2014; gupta et al, 2016). for example, hdac5 and hdac7 have been detected in macrophages, immortalized microglia known as bv2 cells, and microglia (gupta et al 2016; shakespear et al 2013; crockett et al 2017). hdac7 serves as a negative regulator of the anti-inflammatory gene, nur77 (dequiedt et al 2003; hanna et al 2011; liu et al 2017), and there is preliminary evidence of increased expression of hdac7 in microglia following tbi (crockett et al 2017). it has been hypothesized that class iia hdacs may play an important role in initiating the neuroinflammatory response. because of this, it is important to determine possible upstream factors that regulate class iia hdac function. following this, it is important to examine the possibility that polarity proteins par1 (partitioning defective 1), also known as mark 3 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 (microtubule affinity regulating kinase), act as functional regulators of hdac iia. this is because mark2/par-1b has been observed to play a vital role in neuronal migration and axon formation. it was previously observed that animals deficient in par-1b/mark2 seemed to be “primed” for an inflammatory response; par-1b/mark2 deficient mice that had sustained tbi displayed a heightened inflammatory response (dibona et al 2019). it was therefore proposed that one of mark2’s downstream functions is to regulate class iia hdacs within microglia and their response to injury. by determining whether such a signaling pathway exists within microglia, it is possible that this pathway could be therapeutically targeted in the future with the hope of regulating the runaway inflammatory response associated with brain injury. materials and methods to ensure the rigor and reproducibility of the studies, the following measures were taken. first, data collection and analyses were completely blinded to the experimenter. second, rigorous control experiments were performed to ensure the validity of the conclusions—all antibodies were positive and negative control tested with immunohistochemistry. our tbi studies were conducted using a controlled cortical impact (cci) injury model in mice over the course of several years (e.g., dibona et al 2019). sham mice (those who just had a craniotomy performed without significant injury) as well as cci-injured mice consisting of both wild-type (wt) and par1b/mark2 knockout (ko) (−/−) mice (b6.129x1-mark2tm1hpw/j; stock 009365) were used. knockout mice (without par1b/mark2) were used to demonstrate an in vivo model of changes in morphology and density of astrocytes and microglia. all animal procedures were approved by the rutgers-rwjms institutional animal care and use committee prior to the start of experimentation. all animals had free access to food and water and were housed in a clean, temperature-regulated facility that maintained a strict 12-hour light/dark cycle. animals were anesthetized prior to being sacrificed by cardiac perfusion. once the mice were phosphate-buffered saline (pbs) cleared and fixed with paraformaldehyde (pfa), the brains were extracted and submerged in pfa for 1–2 hours post-fixing. each brain was washed with pbs and transferred through sucrose gradients (10%, 20%, and 30%) for one day each and stored in a 4 °c refrigerator. brains from these mice were mounted with o.c.t. compound embedding medium, and 60-μm sections were obtained with a frozen sliding microtome. brains were pierced on the right side (injury was made on the left) to serve as a landmark and maintain orientation throughout the later steps of the process. sections were stored in cryoprotectant solutions (25% glycerol, 25% ethylene glycol, 30% sucrose in 0.1 m pbs) until immunohistochemistry was performed. for immunohistochemistry, two intact hippocampal sections per animal were taken. cells were stained with dapi, gfap, and iba1. dapi is a stain for gross cell morphology and is used to label nuclei. gfap is a stain used to determine astrocyte activation, as astrocytes are specialized cells 4 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 that function to support the central nervous system (cns). iba1 is a stain used to determine microglia activation, as microglia are specialized macrophages of the cns. for the cci-tbi experiment, tissue was washed 3 times for 5 minutes each with 1× pbs to remove cryoprotectant residue. following thorough washing, tissue was permeabilized in blocking solution (2 ml donkey serum, 50 ml pbs-t) for 30 minutes at room temperature. after blocking solution was removed from each tissue, the unrinsed sham-operated control and cci-tbi tissues were then incubated with 0.200 ml of donkey-anti-mouse fab fragment igg (jackson immunoresearch, west grove, pa, 1:60 dilution) for at least 1 hour at room temperature. this solution was removed from each tissue, and primary antibodies were then applied at the desired dilution to incubate overnight at room temperature. the following day, tissue was washed again using the process above, secondary antibody at desired dilution was applied, and the tissues were incubated at room temperature for at least 2 hours. the tissue was then washed and mounted on slides, before being sealed in with dpx. to minimize bias in imaging acquisition and quantification, all slides were blinded prior to imaging, and unblinding occurred after image quantification. results these figures are fluorescent microscope images taken after immunohistochemistry was performed on brain sections (containing the hippocampus) taken from different mice. figure 1 acts as a summary of findings in the ipsilesional cortex taken from brain of a mouse with closed head injury (chi) performed at 21 days and euthanization performed at 28 days. in figure 1a, the dapi stain demonstrates that there are no obvious changes to the structure of the cortex or hippocampus other than the injury site, which negates the possibility of increased expression being due to structural changes. however, in figure 1 b and c, the results qualitatively indicate an increase in expression of gfap (figure 1 b) and iba1 (figure 1 c). this is made more evident when looking at figure 2, which serves as a negative control and shows the baseline levels of expression for both gfap and iba1 in chi092319-1, a mark2 wild-type sham mouse (only craniotomy) operated on at 20 days and euthanized at 28 days. 5 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 figure 3 shows the differences in expression levels on each side of the cortex: the same side as the injury (ipsilesional) and the opposite side of the injury (contralesional). there is increased expression of gfap in both the ipsilesional (figure 1 b) and contralesional (figure 1 c) sides of the cortex; however, the ipsilesional side has a much greater increase in expression in gfap, especially in the area immediately surrounding the injury site. there is also increased expression of iba1 in both the ipsilesional (figure 1 e) and contralesional (figure 1 f) sides of the cortex; however, the ipsilesional side has a much greater increase in expression in iba1, especially in the area immediately surrounding the injury site. 6 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 discussion purpose the purpose of this project was to determine the role that class iia hdacs may play in initiating the inflammatory response after tbi. now that it is understood that class iia hdacs, particularly hdac7, do in fact play a role in initiation of the inflammatory response after tbi, upstream factors can be explored to unearth the molecular mechanism behind the subsequent neuroinflammation after tbi. gaining knowledge of the pathways involved will facilitate attempts to discover therapy or even cures for tbi and spinal cord injury (sci). though it took several tries to obtain intact sections and perform operational ihc staining, the experiment was successful since the high expression of gfap and iba1 supported the hypothesis that hdac plays a role in mediating neuroinflammation. this is evident through the ihc staining, which involved mixing hdac7 (a class iia hdac) with gfap and iba1 secondary antibodies. when comparing injury in figure 1 vs sham in figure 2 results, there is a higher expression of gfap and iba1 in the closed head injury mouse compared to the baseline levels in the sham operation mouse. the results also suggest something that had no clearly defined expectations, which was a global increase in expression of gfap and iba1. this is evident in figure 3, when comparing the contralesional and ipsilesional sides of the cortex. not only was there increased expression of gfap and iba1 in the same side of the cortex as the injury, but also on the opposite side of the brain (albeit in lower concentration). this indicates that the inflammation doesn’t just take place at the injury site, but spreads throughout the rest of the brain at the molecular level. the reason this increased staining is important is because of the roles these cells play in the cns. if there is disease or injury (such as sci or tbi) of the cns, one would expect an activation in astrocytes (given their role of support) and microglia (given their role of maintaining cns health by phagocytizing infection and damaged neurons). 7 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 future directions the next stage of this experiment is to compare the expression level differences between the contralesional and ipsilesional sides of the cortex in a mark2-deficient mouse with chi, as well as expression levels in a chi mark2-deficient mouse vs sham mark2-deficient mouse. when these molecular results are obtained and compared to the results from these wildtype (wt) mice, it can be determined whether par1b/mark2 is a possible upstream factor for class iia hdacs. after this is completed, the next steps will be to test possible therapies from both behavioral and molecular perspectives. the potential therapies for tbi that will be focused on are the diabetic drug metformin and class iia hdac inhibitors. the diabetic drug metformin has been 8 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 identified as a potential candidate because of its anti-oxidative, anti-inflammatory, and antiischemic properties, which may lend itself to neuroprotective effects. there is evidence that metformin can improve the cognitive outcome of mice following tbi (dibona et al, under review). these studies suggested that metformin’s pleiotropic effects (bridgeman et al, 2018) may affect the neuroinflammatory response through stimulation of par1b. even a 50% decrease in par1b/mark2 can lead to an apparent increase in neuroinflammation (dibona et al 2019). metformin treatment has shown positive effects in several animal studies dealing with spinal cord injury and brain trauma (wang et al 2016; zhang et a 2016; tao et al 2018; dibona et al 2019 under review) and in human clinical trials (taheri et al 2019). in addition, metformin has been shown to influence hdac7 functioning (mihaylova et al 2011; bridgeman et al; 2018). metformin has been shown to alter microglial functioning in an animal model of stroke (ji et al, 2018). although the mechanism by which metformin creates positive effects in injury models is not fully understood, it has been hypothesized that metformin dampens the inflammatory response by decreasing the availability of nuclear hdac7 by enhancing par1b/mark2 activity in microglia. possible flaws one of the potential problems with this protocol is the sham operation animal as a negative control. according to our experimental setup, the sham mouse receives a craniotomy in place of a closed head injury. though the craniotomy did function as a negative control in this experiment, we cannot be sure that a craniotomy procedure does not cause unique trauma different from that of chi and accompanying unique neuroinflammation. as a result, this may not be the best representation of baseline expression levels of gfap and iba1. to prevent such confounding data disrupting the results of other experiments, a future experiment could be done comparing an animal only under anesthesia with the conventional negative control of sham operation. this experiment would demonstrate whether the sham is an accurate representation of baseline expression levels without conferring distinct damage and, therefore, altered expression levels. acknowledgements special thanks to dr. david crockett for serving as a wonderful mentor and providing many great learning opportunities. in addition, many thanks to the aresty reseach center for funding this research project and to robert wood johnson medical school for providing resources. this paper was a progress report submitted to the department of cell biology and neuroscience rutgers university. 9 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 references allfrey, v. g., faulkner, r., & mirsky, a. e. (1964) acetylation & methylation of histones & their possible role in the regulation of rna synthesis. proceedings of the national academy of sciences. (5):51786--794. altman, amnon & kong, kok-fai (2011) to kill, you have to duck an hdac. nature immunology. :12279 ep -. baltan, selva, bachleda, amelia, morrison, richard s, & murphy, sean p (2011) expression of histone deacetylases in cellular compartments of the mouse brain & the effects of ischemia. transl stroke res. (3):2411-23. bridgeman, stephanie claire, ellison, gaewyn colleen, melton, phillip edward, newsholme, philip, & mamotte, cyril desire sylvain (2018) epigenetic effects of metformin: from molecular mechanisms to clinical implications. diabetes, obesity & metabolism. (7):201553--1562. das gupta, kaustav, shakespear, melanie r, iyer, abishek, fairlie, david p, & sweet, matthew j (2016) histone deacetylases in monocyte/macrophage development, activation & metabolism: refining hdac targets for inflammatory & infectious diseases. clinical & translational immunology. (1):5e62. dequiedt, f. , martin, m. , von blume, j. , vertommen, d. , lecomte, e. , mari, n. , heinen, m.-f. , bachmann, m. , twizere, j.-c. , huang, m. c. , & et al. (2006)new role for hpar-1 kinases emk & c-tak1 in regulating localization & activity of class iia histone deacetylases. molecular & cellular biology. (19):267086--7102. dequiedt, franck , kasler, herbert , fischle, wolfgang , kiermer, veronique , weinstein, marc , herndier, brian g , & verdin, eric (2003)hdac7, a thymus-specific class ii histone deacetylase, regulates nur77 transcription & tcr-mediated apoptosis. immunity. (5):18687-98. dewan, michael c , rattani, abbas , gupta, saksham , baticulon, ronnie e , hung, ya-ching , punchak, maria , agrawal, amit , adeleye, amos o , shrime, mark g , rubiano, andrés m , rosenfeld, jeffrey v , & park, kee b (2018)estimating the global incidence of traumatic brain injury. j neurosurg. :1-18. dibona, victoria l , zhu, wenxin , shah, mihir k , rafalia, aditi , ben cheikh, hajer , crockett, david p , & zhang, huaye (2019)loss of par1b/mark2 primes microglia during brain development & enhances their sensitivity to injury. j neuroinflammation. (1):1611. faden, alan i. & loane, david j. (2014) chronic neurodegeneration after traumatic brain injury: alzheimer disease, chronic traumatic encephalopathy, or persistent neuroinflammation? neurotherapeutics. (1):12143--150. hanna, richard n. , shaked, iftach , hubbeling, harper g. , punt, jennifer a. , wu, runpei , herrley, erica , zaugg, claudia , pei, hong , geissmann, frederic , ley, klaus , & et al. (2012)nr4a1 (nur77) deletion polarizes macrophages toward an inflammatory phenotype & increases atherosclerosis. circulation research. (3):110416--427. 10 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 ji, juan , xue, teng-fei , guo, xu-dong , yang, jin , guo, ruo-bing , wang, juan , huang, ji-ye , zhao, xiao-jie , & sun, xiu-lan (2018)antagonizing peroxisome proliferator-activated receptor γ facilitates m1-to-m2 shift of microglia by enhancing autophagy via the lkb1ampk signaling pathway. aging cell. (4):17e12774. liu, tian-ya , yang, xiao-ying , zheng, long-tai , wang, guang-hui , & zhen, xue-chu (2016)activation of nur77 in microglia attenuates proinflammatory mediators production & protects dopaminergic neurons from inflammation-induced cell death. journal of neurochemistry. (4):140589--604. loane, david j & byrnes, kimberly r (2010) role of microglia in neurotrauma. neurotherapeutics. (4):7366-77. mihaylova, maria m. , vasquez, debbie s. , ravnskjaer, kim , denechaud, pierre-damien , yu, ruth t. , alvarez, jacqueline g. , downes, michael , evans, ronald m. , montminy, marc , & shaw, reuben j. (2011)class iia histone deacetylases are hormone-activated regulators of foxo & mammalian glucose homeostasis. cell. (4):145607--621. poralla, lukas, stroh, thorsten, erben, ulrike, sittig, marie, liebig, sven, siegmund, britta, & glauben, rainer (2015) histone deacetylase 5 regulates the inflammatory response of macrophages. journal of cellular & molecular medicine. (9):192162--2171. ransohoff, richard m. (2016) how neuroinflammation contributes to neurodegeneration. science. (6301):353777--783. russo, matthew v. & mcgavern, dorian b. (2016) inflammatory neuroprotection following traumatic brain injury. science. (6301):353783--785. shakespear, melanie r. , hohenhaus, daniel m. , kelly, greg m. , kamal, nabilah a. , gupta, praveer , labzin, larisa i. , schroder, kate , garceau, valerie , barbero, sheila , iyer, abishek , & et al. (2013)histone deacetylase 7 promotes toll-like receptor 4-dependent proinflammatory gene expression in macrophages. journal of biological chemistry. (35):28825362--25374. taheri, ali , emami, mahdi , asadipour, erfan , kasirzadeh, sara , rouini, mohammad-reza , najafi, atabak , heshmat, ramin , abdollahi, mohammad , & mojtahedzadeh, mojtaba (2019)a randomized controlled trial on the efficacy, safety, & pharmacokinetics of metformin in severe traumatic brain injury. journal of neurology. (8):2661988--1997. wang, chen , liu, chang , gao, kai , zhao, haosen , zhou, zipeng , shen, zhaoliang , guo, yue , li, zhuo , yao, tianchen , & mei, xifan & (2016)metformin preconditioning provide neuroprotection through enhancement of autophagy & suppression of inflammation & apoptosis after spinal cord injury. biochemical & biophysical research communications. (4):477534--540. xu, ge, ge, yinggang, tao, xiaohong, gao, qing, & liang, xiaoyan (2017) mark2 inhibits the growth of hela cells through ampk & reverses epithelial-mesenchymal transition. oncology reports. (1):38237--244. yoon, somy & eom, gwang hyeon (2016) hdac & hdac inhibitor: from cancer to cardiovascular diseases. chonnam medical journal. (1):521--11. zhang, di , xuan, jun , zheng, bin-bin , zhou, yu-long , lin, yan , wu, yao-sen , zhou, yi-fei , huang, yi-xing , wang, quan , shen, li-yan , & et al. (2016) metformin improves 11 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 functional recovery after spinal cord injury via autophagy flux stimulation. molecular neurobiology. (5):543327--3341. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. privacy paradox or privacy apathy? exploring the relationship between social media usage and public opinion on government usage of data collection programs robert wargaski ✵ abstract the prominence of social media as a mechanism for global communication has raised questions regarding its integrity and security of personal information identifiers such as name, address, and location history. the rise of government surveillance programs, such as those edward snowden exposed in 2013, are a case study in mass collection of identifying personal information without the consent of the american public. this paper looks to determine if there is a causal relationship between social media usage and negative opinions regarding mass personal identifying information government collection programs within the united states. using data compiled by the pew research center, i found that there was no statistically significant relationship at all. this has powerful policy implications such as the normalization of the surveillance state. further research is needed to address concerns regarding the broad variables used as part of this paper. 1 introduction social media is mainly associated with the development of facebook in 2004 and websites such as twitter in 2006. in a recent united states supreme court ruling, it was argued that social media is “the modern public square” and vital for information dissemination (packingham v. north carolina, 2017). as of september 2021, a little under 50% of americans stated that they get news regularly from sites such as twitter. among young people aged 18-24, social media is one of the primary mediums by which news is obtained (walker and matsa, 2021). social media has provided a platform for anyone to voice their opinions, which has led to bipartisan growth in political activism, freelance journalism, and other various activities (jurrat, 2011). in recent years, congressional hearings have forced social media executives to answer questions regarding breaches revealing the personal data of its users (facebook, social. 2018). public controversies have raised questions about what steps these companies use to secure users’ data. the types of data collected by social media companies can include location information, health records, text messages, and emails (levinson-waldman et al., 2022). in addition, questions regarding data collection have not been solely relegated to the private sector. high-profile cases such as the edward snowden leaks in 2013 revealed governmentbacked mass data collection programs. the issue has galvanized public interest in government data collection programs (cassidy, 2013). in this paper, i will determine if there is a connection between social media usage and concern about government collection of personal information. i hypothesize that social media usage encourages apathy regarding government collection of personal information. as mentioned previously, high-profile cases such as snowden’s expose many citizens to the reality that the government has the tools and the influence to collect private citizens' information without them knowing (levinson-waldman et al., 2022). if social media usage does indeed persuade american citizens to remain unconcerned with government collection of personal information, aresty rutgers undergraduate research journal, volume i, issue iv those results could be indicative of broader policy implications regarding the normalization of the socalled “surveillance state” (“defining the surveillance,” 2013). the privacy paradox: prone to apathy? when looking at the literature surrounding social media usage and general views on privacy, multiple public opinion polls show that americans believe they lack oversight over their own information once it is online. according to a pew research poll, around 91% of americans believe they have zero control over what information social media companies collect about them in a given day (madden and rainie, 2015). despite this, a separate pew research poll found that 90% of americans deemed controlling what information was collected about them as “very important.” to many scholars, this is known as the “privacy paradox.” as defined by barth and de jong (2017), the privacy paradox explains the difference between how an individual intends to protect their data on the internet and how they actually behave online. this idea is corroborated by arguments that users of social media oftentimes voluntarily disclose personal information in order to develop a “self-identity” within the platform, disregarding their intentions to keep their personal information private (wu, 2018). in practice, this means that even though individuals are against social media companies having access to their personal information, they knowingly give up such information to these same companies. in interviews, individuals state that they engage in this behavior with a sense of apathy, arguing that once information is on the internet, they have no control over it (hargiattai and marwick, 2016). this behavior is most commonly associated with younger individuals, primarily because they rely on social media more than their older counterparts and therefore feel as though they have no option to opt-out (blank et al.). government mass collection programs: privacy over security? when looking specifically at american attitudes towards government collection of personal information, there is staunch opposition. according to a pew research poll published shortly after the snowden leaks, more than half of all americans disapprove of the government’s metadata collection programs. when accounting for differences in political ideology, opposition to government surveillance remains surprisingly bipartisan with a majority of democrats, republicans, and independents all in opposition to enhanced surveillance measures (gao, 2015). when looking more specifically at data collection programs such as those enacted by the nsa after 9/11, public opposition to these types of programs is mixed (reddick et al., 2015). half of americans believe the metadata collection programs are justified. however, 56% of americans also believe there are not enough limits on the collection programs (dimock et al., 2013). does context matter? views on privacy and government data collection are often context-dependent and can shift depending on the stated goals of the data collection program. in a survey taken during the covid-19 crisis, most americans stated that they would be likely to engage in contact-tracing programs monitoring who they had been in contact with for the purposes of viral transmission mapping. americans who were in favor of the practice argued that the benefits of public health outweighed the costs to personal freedoms (mcclaine, 2020). conversely, many americans were against private entities such as facebook collecting their data for use in advertisements (gramlich, 2021). it seems that in cases where vital infrastructure such as hospital capacity is at stake, individuals are much more willing to knowingly provide social media companies with their personal information. conversely, advertisements in most cases are deemed a nuisance and individuals are therefore less willing to provide their personal information for the express purpose of advertisements. immediately following the 9/11 attacks, many americans called for practices such as forcing everyone to carry a national i.d. card with them (maniam, 2021). however, a decade later, 54% of americans aresty rutgers undergraduate research journal, volume i, issue iv stated that curbing civil liberties would not stop terrorist attacks. in 2015, americans called government collection of electronic data and communications “very concerning” (maniam, 2021). in other words, it seems that context matters in the privacy vs. security debate. if personal information is being used for situations that are deemed “good,” individuals are more likely to go along with the policies in place. the opposite is true for situations deemed a nuisance or otherwise. hypothesis in a comparison of individuals, those who use social media are more likely to have favorable views towards government collection of personal information when compared to those who do not use social media. 3 methods the dataset i used involved results from research that was conducted by the pew research center concerning privacy and security issues in the modern age, specifically in terms of growing reliance on social media. the survey, titled “american trends panel wave 49,” was fielded online from june 3rd to june 17th, 2019, with approximately 4,272 individuals responding. survey demographics showed that around 63% of respondents were white, 12% were black, and 15% were hispanic. no other races were surveyed. around 48% of respondents were male and 52% were female. in terms of age distribution for the survey, around 20% were 18-29, 33% were 30-49, 26% were 50-64, and 20% were 65 and older. around 43% leaned republican, while around 52% leaned democrat. the remaining 5% did not fall into either republican or democrat party affiliation. all data was self-reported on an ordinal scale. the independent variable used was titled “snsuser_w49.” as seen in figure 1, the independent variable asked respondents if they used social media or not, placing the results into one of two categories. in terms of what constituted “social media usage,” any social media usage irrespective of frequency, so long as it is self-reported by the survey participant as “yes,” is flagged as “social media use.” figure 1: graph of dependent variable. out of the 4,272 respondents recorded, 3,500 of them stated that they use social media in some form, forming 81.9% of the responses. in contrast, 18.1% of respondents stated that they do not use social media. the dependent variable i chose was “concerngov_w49” which asked participants “how concerned are you, if at all, about how the government uses the data it collects about you?” on an ordinal scale. as seen in figure 2, respondents selected answers ranging from “very concerned,” “somewhat concerned,” “not too concerned,” and “not at all concerned.” of the total number of respondents who took the survey in its entirety, 2,134 responses were missing and only 2,138 responses were valid. figure 2: graph of dependent variable. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. figure 3: graph of crosstab. for the analysis, a crosstab was used along with the corresponding graph to determine a general relationship between the independent and dependent variables. the crosstab can show a general pattern which helps predict any potential statistical significance with other statistical analytic techniques. building on that, the crosstab was then controlled for age to determine if age made a difference to the results of the original. each individual crosstab represented a specific age range. cramer’s v was used to determine the strength of association between the variables, along with a traditional chisquare test. finally, a correlation and subsequent regression analysis were done to determine if the results of the dependent variable were indeed different from the independent variable. 3 analysis and results first, a crosstab was run as seen in figure 3. overall, i found that there was no significant difference in views concerning government data collection between social media users and non-users. these patterns hold even when controlling for age with the controlled crosstab showing similar results as seen in figure 5. in the 18-29 age category, there is still only a 2-3% difference in opinion across all concern levels. between those who are 30-49, the difference is again within 2-3%, except for the “not concerned at all” category with a little over 5% difference. for those who were 50 and over, there was again a difference between 2-5%. there seems to be no significant difference in opinion between those who use social media and those who don’t, even when controlling for age. the cramer’s v result was 0.091, suggesting a very weak association between the independent and dependent variables. a chi-squared test found that the p-value was 0.001, suggesting i can reject the null hypothesis. the pearson chi-square test resulted in 17.880, which suggests that only around 17.880% of cases would have the same connection if the null hypothesis were correct. finally, for my regression analysis, i found that the r-value is 0.022, the r squared value is 0.000476, and the adjusted r squared value is 0.000008. for the adjusted r squared value specifically, 0.0008% of the variation within the dependent variable is responsible for outcomes of the independent variable. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. figure 4: table of regression analysis. figure 5: coefficient chart. 5 discussion based on the analysis, the results suggest that there is no statistically significant relationship between whether an individual uses social media or not and whether that is a good predictor of their views on government collection of personal information. specifically, when looking at both the cramer’s v and regression analysis, there is little indication that the independent variable impacts the dependent variable. while the correlation and crosstabs seemed to indicate some sort of relationship, they both ended up being too weak to conclude that there was a robust and viable relationship. therefore, based on the evidence, it is appropriate to accept the null hypothesis. in terms of why there seems to be no relationship between social media use and views on government collection of information, one of the reasons could be that it is accepted as normal by us citizens (goitein, 2013). the classic argument in favor of surveillance is that if someone is not doing anything wrong, they should not be concerned about what information their government has on them. even with the knowledge of high-profile cases such as the snowden leaks, it assumes that the public, at least to some extent, is okay with the practice so long as they live comfortably (cofone, 2020). additionally, because government collection of information includes email, phone calls, and gps logs, whether someone uses a social media platform may not have been a strong predictor to begin with. possibly, other factors such as political party affiliation are at play (gao). besides that, younger generations use social media at higher rates than older ones and often are not as worried about surveillance as older individuals, probably because they have grown up in what could be called a quasi-surveillance state in the wake of incidents such as 9/11 and view the practice with a sense of apathy (desilver, 2013). the previous analysis is highly speculative, and further research is needed to establish a connection. specifically, future researchers should focus on how individuals who use different social media platforms compare. i used a single variable compiling all individuals who either use social media or do not use social media for my research. however, that could have skewed the results since there was no specificity as to whether the type of websites used impacted an individual’s views on government aresty rutgers undergraduate research journal, volume i, issue iv information collection. for instance, people who use only facebook might have different opinions than people who simultaneously use facebook, twitter, and instagram. a future research design should account for this variation, as social media sites have gained an exponential following in recent years, giving researchers rich data pools to use. many of the issues present with the current findings could be due to the survey methodology used. for one, all the information compiled by the pew research center was based on self-reported survey responses. in addition, the survey data based on my parameters was incomplete. around half of the total respondents were unaccounted for in my research because they were not flagged as people who do or do not use social media. therefore, they had to be disregarded in the statistical analysis. overall, this is perhaps the greatest flaw in my research methodology and should be taken into account in future studies∎ 6 references [1] american trends panel wave 49. (2019, june 17). pew research center. retrieved from https://www.pewresearch.org/internet/dataset/american-trends-panel-wave-49/ [2] barth, s., & de jong, m. (2017). the privacy paradox: investigating discrepancies between expressed privacy concerns and actual online behavior a systematic literature review. telematics and informatics, 34(7), 1038-1058. https://www.sciencedirect.com/science/article/pii/s0736585317302022 [3] blank, g., bolsover, g., & dubois, e. (n.d.). a new privacy paradox: young people and privacy on social networking sites. social science research network. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2479938 [4] cassidy, j. (2013, august 20). snowden's legacy: a public debate about online privacy. the new yorker. retrieved from https://www.newyorker.com/news/john-cassidy/snowdens-legacy-a-public-debate-about-online-privacy [5] cofone, i. n. (2020). nothing to hide, but something to lose. university of toronto law journal, 70(1). https://muse.jhu.edu/article/743215 [6] defining the surveillance state. (2013, october 31). privacy international. https://privacyinternational.org/blog/1513/definingsurveillance-state [7] desilver, d. (2013, june 20). young americans and privacy: 'it's complicated'. pew research center. retrieved from https://www.pewresearch.org/facttank/2013/06/20/young-americans-and-privacy-its-complicated/ [8] dimock, m., doherty, c., tyson, a., & gewurz, d. (2013, july 26). few see adequate limits on nsa surveillance program. pew research center. retrieved from https://www.pewresearch.org/wp-content/uploads/sites/4/legacy-pdf/7-26-2013-nsa-release.pdf [9] facebook, social media privacy, and the use and abuse of data: hearings before the committee on the judiciary, 2018 leg. (d.c. apr. 10, 2018). https://www.judiciary.senate.gov/meetings/facebooksocial-media-privacy-and-the-use-and-abuse-of-data [10] gao, g. (2015, may 29). what americans think about nsa surveillance, national security and privacy. pew research center. retrieved from https://www.pewresearch.org/fact-tank/2015/05/29/what-americans-thinkabout-nsa-surveillance-national-security-and-privacy/ [11] goitein, e. (2013, july 31). the danger of american apathy on nsa surveillance. brennan center for justice. retrieved from https://www.brennancenter.org/our-work/analysisopinion/danger-american-apathy-nsa-surveillance [12] gramlich, j. (2021, june 1). 10 facts about americans and facebook. pew research center. retrieved from https://www.pewresearch.org/facttank/2021/06/01/facts-about-americans-and-facebook/ [13] hargittai, e., & marwick, a. (2016). "what can i really do?" explaining the privacy paradox with online apathy. international journal of communication, 10, 3737-3757. https://ijoc.org/index.php/ijoc/article/view/4655 [14] jurrat, n. (2011, april). mapping digital media: citizen journalism and the internet. open society foundations. retrieved from https://www.opensocietyfoundations.org/uploads/ce357846-eecf-4114-a3e4-500ce863fbe6/mappingdigital-media-citizen-journalism-and-internet20110712.pdf [15] lau, t. (2019, may 22). the government is expanding its social media surveillance capabilities. brennan center for justice. retrieved from https://www.brennancenter.org/our-work/analysis-opinion/government-expanding-its-social-media-surveillance-capabilities [16] levinson-waldman, r., panduranga, h., & patel, f. (2022, january 7). social media surveillance by the u.s. government. brennan center for justice. retrieved from https://www.brennancenter.org/our-work/researchreports/social-media-surveillance-us-government https://www.pewresearch.org/internet/dataset/american-trends-panel-wave-49/ https://www.pewresearch.org/internet/dataset/american-trends-panel-wave-49/ https://www.sciencedirect.com/science/article/pii/s0736585317302022 https://www.sciencedirect.com/science/article/pii/s0736585317302022 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2479938 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2479938 https://www.newyorker.com/news/john-cassidy/snowdens-legacy-a-public-debate-about-online-privacy https://www.newyorker.com/news/john-cassidy/snowdens-legacy-a-public-debate-about-online-privacy https://muse.jhu.edu/article/743215 https://privacyinternational.org/blog/1513/defining-surveillance-state https://privacyinternational.org/blog/1513/defining-surveillance-state https://www.pewresearch.org/fact-tank/2013/06/20/young-americans-and-privacy-its-complicated/ https://www.pewresearch.org/fact-tank/2013/06/20/young-americans-and-privacy-its-complicated/ https://www.pewresearch.org/fact-tank/2013/06/20/young-americans-and-privacy-its-complicated/ https://www.pewresearch.org/wp-content/uploads/sites/4/legacy-pdf/7-26-2013-nsa-release.pdf https://www.pewresearch.org/wp-content/uploads/sites/4/legacy-pdf/7-26-2013-nsa-release.pdf https://www.judiciary.senate.gov/meetings/facebook-social-media-privacy-and-the-use-and-abuse-of-data https://www.judiciary.senate.gov/meetings/facebook-social-media-privacy-and-the-use-and-abuse-of-data https://www.pewresearch.org/fact-tank/2015/05/29/what-americans-think-about-nsa-surveillance-national-security-and-privacy/ https://www.pewresearch.org/fact-tank/2015/05/29/what-americans-think-about-nsa-surveillance-national-security-and-privacy/ https://www.pewresearch.org/fact-tank/2015/05/29/what-americans-think-about-nsa-surveillance-national-security-and-privacy/ https://www.brennancenter.org/our-work/analysis-opinion/danger-american-apathy-nsa-surveillance https://www.brennancenter.org/our-work/analysis-opinion/danger-american-apathy-nsa-surveillance https://www.pewresearch.org/fact-tank/2021/06/01/facts-about-americans-and-facebook/ https://www.pewresearch.org/fact-tank/2021/06/01/facts-about-americans-and-facebook/ https://ijoc.org/index.php/ijoc/article/view/4655 https://www.opensocietyfoundations.org/uploads/ce357846-eecf-4114-a3e4-500ce863fbe6/mapping-digital-media-citizen-journalism-and-internet-20110712.pdf https://www.opensocietyfoundations.org/uploads/ce357846-eecf-4114-a3e4-500ce863fbe6/mapping-digital-media-citizen-journalism-and-internet-20110712.pdf https://www.opensocietyfoundations.org/uploads/ce357846-eecf-4114-a3e4-500ce863fbe6/mapping-digital-media-citizen-journalism-and-internet-20110712.pdf https://www.opensocietyfoundations.org/uploads/ce357846-eecf-4114-a3e4-500ce863fbe6/mapping-digital-media-citizen-journalism-and-internet-20110712.pdf https://www.brennancenter.org/our-work/analysis-opinion/government-expanding-its-social-media-surveillance-capabilities https://www.brennancenter.org/our-work/analysis-opinion/government-expanding-its-social-media-surveillance-capabilities https://www.brennancenter.org/our-work/analysis-opinion/government-expanding-its-social-media-surveillance-capabilities https://www.brennancenter.org/our-work/analysis-opinion/government-expanding-its-social-media-surveillance-capabilities https://www.brennancenter.org/our-work/research-reports/social-media-surveillance-us-government https://www.brennancenter.org/our-work/research-reports/social-media-surveillance-us-government aresty rutgers undergraduate research journal, volume i, issue iv [17] madden, m., & rainie, l. (2015, may 20). americans' attitudes about privacy, security and surveillance. pew research center. retrieved from https://www.pewresearch.org/internet/2015/05/20/americans-attitudesabout-privacy-security-and-surveillance/ [18] maniam, s. (2021, june 1). americans feel the tensions between privacy and security concerns. pew research center. retrieved from https://www.pewresearch.org/facttank/2016/02/19/americans-feel-the-tensions-betweenprivacy-and-security-concerns/ [19] mcclaine, c. (2020, october 30). americans' views on covid-19 contract tracing. pew research center. retrieved from https://www.pewresearch.org/fact-tank/2020/10/30/keyfindings-about-americans-views-on-covid-19-contacttracing/ [20] nam, t. (2019). what determines the acceptance of government surveillance? examining the influence of information privacy correlates. the social science journal, 56(4), 530-544. https://doi.org/10.1016/j.soscij.2018.10.001 [21] packingham v. north carolina, no. 15-1194 (june 19, 2017). https://www.supremecourt.gov/opinions/16pdf/151194_08l1.pdf [22] reddick, c. g., chatfield, a. t., & jaramillo, p. a. (2015). public opinion on national security agency surveillance programs: a multi-method approach. government information quarterly, 32(2), 129-141. https://www.sciencedirect.com/science/article/pii/s0740624x15000246 [23] walker, m., & matsa, k. e. (2021, september 20). news consumption across social media in 2021. pew research center. retrieved from https://www.pewresearch.org/journalism/2021/09/20/news-consumption-across-social-mediain-2021/ [24] wu, p. f. (2018). the privacy paradox in the context of online social networking: a self-identity perspective. journal of the association for information science and technology, 70(3), 207-217. https://doi.org/10.1002/asi.24113 robert wargaski is a junior in the school of arts and sciences. he is majoring in political science with a focus on counterterrorism and national security policy. in the future, robert plans to go to graduate school for political science and hopes to someday work in the federal intelligence community. https://www.pewresearch.org/internet/2015/05/20/americans-attitudes-about-privacy-security-and-surveillance/ https://www.pewresearch.org/internet/2015/05/20/americans-attitudes-about-privacy-security-and-surveillance/ https://www.pewresearch.org/internet/2015/05/20/americans-attitudes-about-privacy-security-and-surveillance/ https://www.pewresearch.org/fact-tank/2016/02/19/americans-feel-the-tensions-between-privacy-and-security-concerns/ https://www.pewresearch.org/fact-tank/2016/02/19/americans-feel-the-tensions-between-privacy-and-security-concerns/ https://www.pewresearch.org/fact-tank/2016/02/19/americans-feel-the-tensions-between-privacy-and-security-concerns/ https://www.pewresearch.org/fact-tank/2020/10/30/key-findings-about-americans-views-on-covid-19-contact-tracing/ https://www.pewresearch.org/fact-tank/2020/10/30/key-findings-about-americans-views-on-covid-19-contact-tracing/ https://www.pewresearch.org/fact-tank/2020/10/30/key-findings-about-americans-views-on-covid-19-contact-tracing/ https://doi.org/10.1016/j.soscij.2018.10.001 https://www.supremecourt.gov/opinions/16pdf/15-1194_08l1.pdf https://www.supremecourt.gov/opinions/16pdf/15-1194_08l1.pdf https://www.sciencedirect.com/science/article/pii/s0740624x15000246 https://www.sciencedirect.com/science/article/pii/s0740624x15000246 https://www.pewresearch.org/journalism/2021/09/20/news-consumption-across-social-media-in-2021/ https://www.pewresearch.org/journalism/2021/09/20/news-consumption-across-social-media-in-2021/ https://www.pewresearch.org/journalism/2021/09/20/news-consumption-across-social-media-in-2021/ https://doi.org/10.1002/asi.24113 aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. the role of autophagy in phosphatidyl glycerol facilitated cholesterol clearance from the endolysosomal system of npc-1 deficient cells tamara allada, olga ilnytska phd, judith storch (faculty advisor) ✵ abstract niemann pick type c (npc) disease is a rare lysosomal storage disorder in which one of the genes that codes for either the npc-1 or npc-2 protein is mutated, causing cell lysosomes to accumulate cholesterol and lipids. previous studies discovered that a unique late endosomal/lysosomal phospholipid, lysobisphosphatidic acid (lpba), is involved in cholesterol clearance from late endosomes. it has also been shown that exogenous treatment of the npc-1 deficient cells with lbpa’s precursor, phosphatidylglycerol (pg), leads to lbpa enrichment and subsequent endolysosomal cholesterol clearance. autophagy is a mechanism of cellular clearance in the endolysomal system and we are interested to see if it is a partial route in cholesterol clearance during pg treatment of npc-1 deficient cells. to do so, we silenced the gene that codes for an essential protein in the autophagy pathway, making the cells autophagy deficient. we then treated the cells with pg, measured the amount of cholesterol clearance in those cells, and compared it to cells with normal autophagy. we found significantly less cholesterol clearance by pg in cells with defective autophagy, confirming that autophagy is involved as a partial route in cholesterol clearance during pg treatment, but not enough of a difference to conclude that it is a major underlying mechanism. 1 introduction niemann pick type c (npc) disease results in cholesterol and lipids becoming trapped in cell lysosomes and unable to be trafficked through the cell to where they are needed.[19,26] cholesterol accumulates in the lysosome and exits via lysosomal membrane proteins, such as npc-1 and npc-2, in unaffected cells. the npc proteins are essential for transporting cholesterol through the endolysosomal system; therefore, nonfunctional npc proteins result in an accumulation of cholesterol and secondary lipids in the late endosomal and lysosomal compartments. an npc patient suffers from severe motor and neurological symptoms that worsen with time.[14] the exact method of how cholesterol is transported out of the endolysosomal system is unknown,[14] but our preliminary data suggests that autophagy may be involved. autophagy is a mechanism of cellular clearance in which damaged proteins, organelles, and membranes are transported to the endolysosomal system where they can be broken down into metabolites used by the cell.[28] it is essentially a cellular degradation and recycling process. in order for cells to perform essential biosynthetic and energy production pathways, they need basic monomers of larger molecules. cells can obtain these monomers through the break-down and recycling of damaged organelles, proteins, and cellular membranes. these molecular degradations occur in lysosomes, which are acidic vesicles containing numerous hydrolases and found within the cell. hydrolases are a class of proteins that break down large molecules into their monomers. macroautophagy involves sequestering the molecules that need to be degraded into vesicles called autophago aresty rutgers undergraduate research journal, volume i, issue iii somes and delivering them to lysosomes for degradation,[28] as shown in figure 1. the process of autophagy begins with the process shown in figure 2. several molecules come together to form the ulk1/2-atg13-rb1cc1 induction complex, which can develop at multiple sites throughout the cytoplasm.[5,10] once formed, the complex can either be activated or inactivated via a molecule called the mechanistic target of rapamycin complex 1 (mtorc1). when mtorc1 is associated with the induction complex, the complex is inactive and the process of autophagy stops. when mtorc1 dissociates from the induction complex, the complex activates, resuming autophagy.[9] to summarize, the process of autophagy is inhibited when mtorc1 is associated with the ulk1/2atg13-rb1cc1 induction complex and resumes when mtorc1 disassociates from the induction complex. in addition, figure 2 displays how the nutrient status of the cell is involved in autophagy regulation. during nutrient starvation, there is low energy in the cell. the nutrition status therefore determines the status of autophagy, as the cell needs the monomers produced by autophagy for the energy production pathways.[9] during nutrient enrichment, there is enough energy in the cell and autophagy is not needed. the next step of autophagy is nucleation, or formation of the membrane, which is an essential step in macroautophagy.[18] once the membrane begins to expand, it is called a phagophore. this structure is double membranous and expands to eventually create a closed, circular vesicle.[8] the process of phagophore elongation is composed of 2 conjugated systems. the first system results in the formation of the atg12-atg5-atg16l1 complex, as shown in figure 3, which is needed for the second system. once formed, the complex associates with the phagophore membrane.[15,16] the second system is shown in figure 4 and ultimately allows the phagophore membrane to elongate and move locations within the cell.[18] one step in this process is facilitated by the atg12-atg5-atg16l1 complex formed in the first system. as the phagophore membrane expands, it curves around its target molecules, figure 1: overview of macroautophagy in mammalian cells. figure obtained from [18]. figure 2: activation of the induction complex during a nutrient starved cellular state. figure obtained from [18]. figure 3: (left) atg12–atg5-atg16l1 conjugation complex. figure obtained from [18]. figure 4: (right) atg8/lc3 conjugation system. figure obtained from [18]. aresty rutgers undergraduate research journal, volume i, issue iii known as cargo, and keeps adjusting to wrap around the cargo.[18] when it has fully wrapped around the cargo, the ends connect, forming a closed double membranous vesicle known as an autophagosome.[18] the autophagosome will move toward the lysosome via the cell’s microtubules.[17] once at the lysosome, the autophagosome’s outer membrane will fuse with the lysosomal membrane, creating an autolysosome. the autophagosome may also fuse with endosomes, in both early and late stages, before fusing with the lysosome.[2,25] the last step of autophagy is referred to as autophagic flux where degradation within the autolysosome begins when the inner membrane is disrupted by atg15.[7,24] once the inner membrane is disrupted, the cargo that the autophagosome carried is exposed to the lysosome’s acidic environment and hydrolases that degrade molecules into their basic monomers. these metabolites are then exported from the lysosome into the cell via transporter proteins on the lysosomal membrane.[29] different metabolites are transported out by different receptors. npc-1 and npc-2 proteins are examples of these receptors used in the transport of cholesterol out of the lysosome.[20,27] niemann pick c (npc) disease is a type of lysosomal storage disorder, which is a category of disorders usually caused by a mutation in a gene that codes for a lysosomal transport protein.[21] several studies have shown that autophagy is blocked or dysfunctional in most lysosomal storage disorders because of impaired fusion between the autophagosome and the lysosome.[13,22,23] this block is thought to be caused by abnormal cholesterol accumulation in the lysosome.[1] therefore, in npc disease, since one of the npc proteins is nonfunctional, cholesterol will be trapped in the lysosomes and autophagy will be blocked due to an inability of the autophagosome to fuse with cholesterol-filled lysosomes. it was found that cholesterol clearance depends on the interaction of a region of the npc-2 protein with a lipid called lysobisphosphatidic acid, or lpba.[6] in patients with npc-1 disease, meaning their npc-1 protein is dysfunctional, cholesterol clearance is still affected. this is because the limiting membrane of a lysosome is covered in a sugar coat called glycocalyx and it is believed that the luminal domains of the npc-1 protein are required for cholesterol to be able to penetrate through the glycocalyx and exit the lysosome.[11,12] increasing the lbpa levels in npc-1 deficient cells have been found to result in a significant amount of cholesterol clearance because the lbpa can still interact with the npc-2 protein.[6] exogenous treatment of lbpa’s precursor phosphatidylglycerol, or pg, also results in cholesterol clearance because the pg can go through its biosynthetic pathway to make lbpa.[13] our preliminary data shows that pg treatment of npc-1 deficient cells facilitates autophagosome fusion with the lysosome and subsequent autophagic flux. since npc-1 deficient cells block autophagy and pg helps it resume, we are interested in seeing if autolysosome formation and autophagic flux are involved in cholesterol clearance from late endosomes when these cells are treated with pg. to determine whether pg/lbpa-induced autophagic flux and cholesterol clearance are directly linked, we blocked autophagy by silencing the gene that codes for the atg12 protein. atg12 is essential to autophagy because it is part of the atg12-atg5-atg16l1 complex that binds atg8/lc3 to the membrane, an essential process in phagophore elongation. autophagy can be blocked by silencing the genes of a number of essential proteins to the process, but we will only focus on the gene that codes for atg12. after making the autophagy knockdown cell line and the control cell line, we treated the cells with pg and measured the subsequent cholesterol accumulation to see if the cells with blocked autophagy still cleared cholesterol to the same extent. if autophagy is involved in cholesterol clearance during pg treatment, the cells that received the pg treatment but had autophagy blocked would have less cholesterol clearance than cells that received the pg treatment and had functional autophagy. 2 methodology npc-1 deficient cells were seeded in 12-well plates on coverslips with antibiotic-free growth medium supplemented with fbs (fetal bovine serum), which contains nutrients that help the cells grow and divide. the cells were npc-1 deficient either be aresty rutgers undergraduate research journal, volume i, issue iii cause they were fibroblast cells taken from patients with npc-1 disease or because they were npc-1 knocked out hela cells, meaning the gene that codes for the npc-1 protein was silenced. these hela cells are true npc-1 -/-, meaning npc-1 will be 100% knocked-out. after 24 hours, the cells were at confluency, or 0.5 x 106 cells per plate, and we washed the cells with shrna transfection medium. we then made 2 cell lines: an autophagy knockout line and a control line. the cells for the autophagy knockout line were transduced with atg12 shrna lentiviral particles and the cells for the control line were transduced with scrambled control shrna lentiviral particles. the cells of both lines were transduced with shrna lentiviral particles in concentrations of 1x105 infectious units of virus (ifu). the atg12 shrna encoded for a sequence that degraded mrna in the cell that coded for the atg12 protein. autophagy was blocked because these cells could not make the atg12 protein. the scrambled control shrna encoded for a sequence that did not degrade any known cellular mrna. since these cells could express all of their normal proteins except npc-1, autophagy was not blocked. knockdown shrna did not get rid of 100% of the target mrna, and preliminary experiments showed that autophagy would be reduced but not completely blocked. after 48 hours of incubation, the old medium was removed and 2 µg/ml antibiotic puromycin medium was added to all well. cells transfected with the virus particles would have a gene that make the cells resistant to puromycin, while cells that were not transfected would be killed by the antibiotic. cells were then incubated for 3 days, after which all stable cells that remained alive were known to be the ones that were transfected. each cell line had a control group and a treatment group, with a total of 48 wells. we treated both treatment groups with 100μm pg and let the cells incubate. after 48 hours we washed all cells using phosphate buffer saline (pbs) and fixed them to the coverslips using a 4% paraformaldehyde solution. we then stained the cells on the coverslips with cholesterol-binding polyene antibiotic filipin iii and a nuclear dye (sytox green). each of the 48 coverslips contained about 50-100 fibroblast cells. images were taken with the epifluorescent revolve microscope at 40x magnification. fluorescent intensities of the subcellular structures stained with filipin were analyzed using image j software as described by.[11] 3 results fluorescent microscopy detects the intensity of filipin from the coverslips on which the cells were fixed (figure 5). filipin binds to cholesterol; the higher the filipin intensity, the higher the cholesterol levels in the cells. the results of the fluorescent microscopy are shown in figure 6. scr control is a control group that is npc-1 deficient only. scr+pg is a pg-treated group that is npc-1 deficient only. both of these groups came from the control cell line that received the scrambled control shrna. atg12 kd control is a control group that is both npc-1 deficient and has autophagy blocked. atg12 kd + pg is a pg-treated group that is npc-1 deficient and has autophagy blocked. both of these groups have autophagy blocked because they came from the autophagy knockout cell line that received the atg12 shrna. the untreated control groups provide the baseline filipin intensity for each cell line. figure 6 shows that there is a statistically significant difference in cholesterol content between the scr control and scr+pg groups, with scr+pg having less cholesterol. this is in agreement with the previous finding that pg supplementation increases lbpa levels and thus decreases cholesterol content.[14] the red bracket in figure 6 shows that atg12 kd+pg has significantly more cholesterol than scr+pg. this shows that autophagy is a partial route in cholesterol clearance during pg treatment, since the pg treatment was less effective when autophagy was blocked. lastly, figure 6 shows that atg12 kd+pg has significantly less cholesterol than atg kd control. this indicates that autophagy is not the major underlying mechanism for pg/lbpa-induced cholesterol efflux since pg treatment had an effect even when autophagy was blocked. this is also suggested by the fact that there is only 15% more cholesterol in atg12 kd+pg than in scr+pg; there would be closer to 100% more cholesterol in atg12 aresty rutgers undergraduate research journal, volume i, issue iii kd+pg if autophagy was a major route in cholesterol clearance. the statistical tests were t-tests using a significance level of 0.01 and all three tests produced a p value of less than 0.001, as indicated by the asterisks. 4 discussion the statistically significant results between scr+pg and atg12 kd+pg indicate that autophagy is blocked in npc disease due to an inability of the autophagosome to fuse with the lysosome and this impaired ability is restored by pg/lbpa enrichment in npc-1 cells, ince the inner-lysosomal portion of the npc-1 protein is thought to be required for cholesterol to penetrate through the glycocalyx that coats the limiting lysosomal membrane, enhanced autolysosome formation during pg/lbpa enrichment could potentially provide a route for cholesterol egress that bypasses the glycocalyx.[11,12,13,22,23] we hypothesize that the glycocalyx of the lysosome may become weakened when the autophagosome fuses with the lysosome, allowing cholesterol to more readily reach the limiting membrane of the lysosome, and therefore more readily exit the compartment. npc-1 disease shares many similarities with alzheimer’s disease.[3] increased endogenous lbpa has been reported in both npc-1[6] and alzheimer’s disease,[4] where such an increase might be a compensatory homeostatic cellular response to eliminate toxic deposits. therefore, pg/lbpa enrichment has great potential for the treatment of npc-1 disease and potentially other lysosomal storage disorders and neurological diseases with autophagy defects, such as parkinson’s or alzheimer’s disease. it is important to understand all processes involved in cholesterol clearance during the treatment, including autophagy. in future work, to understand the extent of autophagy’s role in cholesterol clearance during pg treatment, we will block autophagy by silencing other molecules, such as atg5. there are many ways that autophagy can be blocked because there are many essential proteins to the process that can be silenced. silencing some genes for autophagy proteins may result in a more complete knockdown than silencing other genes. figure 5: sample of images of npc-1 fibroblast cells in each condition, filipin stained. figure 6: quantification of filipin intensity in npc1-deficient fibroblasts stably transduced with lentiviral non silencing control shrna (scr) or shrna targeting atg12 and selected for stable shrna expression. 𝑁𝑁 > 500 cells/condition. data from 2 independent experiments. ∗∗∗ signifies a 𝑝𝑝 value < 0.001. aresty rutgers undergraduate research journal, volume i, issue iii patients with npc-1 disease develop detrimental neurological symptoms because their cells accumulate cholesterol in the endolysosomal compartment and are unable to clear the cholesterol from that compartment. lbpa clears cholesterol from the endolysosomal compartment, which indicates that potential therapies to prevent the neurological symptoms may be found in interventions that increase lbpa, such as pg treatment. uncovering the full process of pg-facilitated clearance of cholesterol will aid in manipulating lbpa levels for potential therapeutic benefits in patients with npc disease, other lysosomal storage disorders, and neurological diseases∎ 5 references [1] ballabio a, bonifacino js. lysosomes as dynamic regulators of cell and organismal homeostasis. nature reviews molecular cell biology. 2020;21(2):101-118. [2] berg to, fengsrud m, strømhaug pe, berg t, seglen po. isolation and characterization of rat liver amphisomes. evidence for fusion of autophagosomes with both early and late endosomes. the journal of biological chemistry. 1998;273(34):21883-. [3] borbon ia, erickson rp. interactions of npc1 and amyloid accumulation/deposition in the app/ps1 mouse model of alzheimer’s. journal of applied genetics. 2011;52(2):213-218. [4] chan rb, oliveira tg, cortes ep, et al. comparative lipidomic analysis of mouse and human brain with alzheimer disease. the journal of biological chemistry. 2012;287(4):2678-2688. [5] chen y, klionsky dj. the regulation of autophagy unanswered questions. journal of cell science. 2011;124(pt 2):161-170. [6] chevallier j, chamoun z, jiang g, et al. lysobisphosphatidic acid controls endosomal cholesterol levels. the journal of biological chemistry. 2008;283(41):27871-27880. [7] epple ud, suriapranata i, eskelinen el, thumm m. aut5/cvt17p, a putative lipase essential for disintegration of autophagic bodies inside the vacuole. journal of bacteriology. 2001;183(20):5942-5955. [8] he c, klionsky dj. regulation mechanisms and signaling pathways of autophagy. annual review of genetics. 2009;43(1):67-93. [9] hosokawa n, hara t, kaizuka t, et al. nutrient-dependent mtorc1 association with the ulk1-atg13-fip200 complex required for autophagy. molecular biology of the cell. 2009;20(7):1981-1991. [10] itakura e, mizushima n. characterization of autophagosome formation site by a hierarchical analysis of mammalian atg proteins. autophagy. 2010;6(6):764-776. [11] kwon hj, abi-mosleh l, wang ml, et al. structure of n-terminal domain of npc1 reveals distinct subdomains for binding and transfer of cholesterol. cell. 2009;137(7):12131224. [12] li x, saha p, li j, blobel g, pfeffer sr. clues to the mechanism of cholesterol transfer from the structure of npc1 middle lumenal domain bound to npc2. proceedings of the national academy of sciences pnas. 2016;113(36):1007910084. [13] lieberman ap, puertollano r, raben n, slaugenhaupt s, walkley su, ballabio a. autophagy in lysosomal storage disorders. autophagy. 2012;8(5):719-730. [14] mccauliff la, langan a, li r, et al. intracellular cholesterol trafficking is dependent upon npc2 interaction with lysobisphosphatidic acid. elife. 2019;8. [15] mizushima n, kuma a, kobayashi y, et al. mouse apg16l, a novel wd-repeat protein, targets to the autophagic isolation membrane with the apg12-apg5 conjugate. journal of cell science. 2003;116(pt 9):1679-1688. aresty rutgers undergraduate research journal, volume i, issue iii [16] mizushima n, yamamoto a, hatano m, et al. dissection of autophagosome formation using apg5-deficient mouse embryonic stem cells. the journal of cell biology. 2001;152(4):657-667. [17] monastyrska i, rieter e, klionsky dj, reggiori f. multiple roles of the cytoskeleton in autophagy. biological reviews of the cambridge philosophical society. 2009;84(3):431-448. [18] parzych kr, klionsky dj. an overview of autophagy: morphology, mechanism, and regulation. antioxid redox signal. 2014;20(3):460-473. [19] peake kb, vance je. defective cholesterol trafficking in niemann-pick c-deficient cells. febs letters. 2010;584(13):2731-2739. [20] pfeffer sr. npc intracellular cholesterol transporter 1 (npc1)-mediated cholesterol export from lysosomes. the journal of biological chemistry. 2019;294(5):1706-1709. [21] saftig p, klumperman j. lysosome biogenesis and lysosomal membrane proteins: trafficking meets function. nature reviews molecular cell biology. 2009;10(9):623-635. [22] seranova e, connolly kj, zatyka m, et al. dysregulation of autophagy as a common mechanism in lysosomal storage diseases. essays in biochemistry. 2017;61(6):733-749. [23] settembre c, fraldi a, jahreiss l, et al. a block of autophagy in lysosomal storage disorders. human molecular genetics. 2008;17(1):119-129. [24] teter sa, eggerton kp, scott sv, kim j, fischer am, klionsky dj. degradation of lipid vesicles in the yeast vacuole requires function of cvt17, a putative lipase. the journal of biological chemistry. 2001;276(3):2083-2087. [25] tooze j, hollinshead m, ludwig t, howell k, hoflack b, kern h. in exocrine pancreas, the basolateral endocytic pathway converges with the autophagic pathway immediately after the early endosome. j cell biol. 1990 aug;111(2):329-45. [26] vanier m, millat g. niemann-pick disease type c. clinical genetics. 2003;64(4):269-281. [27] yim ww-y, mizushima n. lysosome biology in autophagy. cell discovery. 2020;6(1):6-6. [28] yorimitsu t, klionsky dj. autophagy: molecular machinery for self-eating. cell death and differentiation. 2005;12(s2):1542-1552. [29] xu h, ren d. lysosomal physiology. annual review of physiology. 2015;77(1):57-80. my name is tamara allada and i am a 2021 rutgers graduate and the main author of this paper. i spent my final two years of college working with dr. judith storch and dr. olga ilnytska in the rutgers department of nutritional sciences. our lab's main area of study is the method of intracellular cholesterol transport, as the exact pathway in which cholesterol is trafficked through a cell is currently unknown. there are many lysosomal storage disorders, such as niemman pick c disease, that cause cholesterol to become trapped in a compartment of the cell and unable to be trafficked, resulting in severe neurological defects. knowing the exact pathway of intracellular cholesterol transport would aid greatly in the search for a treatment to such disorders. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. socializing at the shopping mall: a review of literature to reassess the social value of the american mall eliza k. peterson, andrew urban, phd (faculty advisor) ✵ abstract the shopping mall was not just a monument to consumption, as it is often framed, but also a place where american suburbanites could find community. the american shopping mall was pioneered in the 1950s, partially as a way to fix the social gaps that are present in a suburban environment and rose to its heyday in the 1980s and 1990s. during this time, the mall became a place for suburban residents to socialize, as illustrated by the phenomena of “mall rats” and mall walkers. despite this social significance, the mall was typically only seen as a place to fulfill consumer needs rather than social needs. i will survey the literature surrounding the social legacy of malls in an effort to reframe the legacy of the institution, in light of the benefits provided by hindsight. shopping malls should be recognized for their ability to facilitate social connections and community during the late 20th century in a space as atomized as the suburbs. 1 introduction the shopping mall looms large in the american historical imagination. in many ways, it emblematizes the prosperity and optimism of the late 20th century; the mall was a place with all the modern conveniences, serving the citizens of the newly exploded suburbs and their ever-growing consumer appetites. at the same time though, its image has been used as a shorthand for the vapidity of consumerism; the public often saw mall shoppers as mindless consumer zombies. these critiques, however, have not all stood the test of time. in hindsight, shoppers can see malls for what they really are – a public space that serves social needs as much as it serves consumer demand. fond memories of the evocative sensory experience of mall shopping, combined with the sense of community found in many of these spaces, has outweighed moral panics about the mall and its denizens. i will be arguing for the deeper social significance of the shopping mall through an exploration of the literature and philosophy surrounding the institution as the american shopping mall was an important space for socializing, particularly during the mall’s heyday in the 1980s and 1990s. 2 background the introduction of the mall made the acquisition of goods, and lots of them, easier at an unprecedented level. no longer did suburban consumers have to contend with the inconveniences of downtown like bad weather, short hours, or, worst of all, parallel parking. the enclosed shopping mall — which usually features one to four “anchor stores” as well as various other tenants for retail, service, and dining — was initially developed in the early 1950s in the midwestern suburbs before spreading across america followed by the rest of the world in the latter half of the 20th century. this piece will focus on american shopping malls, which became a major social and cultural force, reaching its height of power in the 1980s and 1990s. the american shopping mall, while often seen as artificial and shallow, in fact provided many citizens, and especially american suburban teenagers, with a place to gather and socialize. for these mall-frequenters, the mall was not simply a place to fulfill consumer desires — it was also a place to fulfill social desires. malls were not always seen as social spaces. for much of their history they were denigrated as no more than shrines to consumerism. the word “conaresty rutgers undergraduate research journal, volume i, issue iv sumerism” was first used in 1915, although its theoretical underpinnings were established in economist and sociologist thorstein veblen’s 1899 treatise, the theory of the leisure class, which introduced the concept of “conspicuous consumption.” in the theory of the leisure class, veblen argues that members of the upper class, particularly the noveau riche, will buy items excessive in quality and/or quantity to signal their elevated economic status.9 consumerism springs from this idea of “conspicuous consumption” though consumerism transcends class in some ways, especially in america. america is widely known for our robust consumer culture, with consumption being inextricably tied to identity and status, and the individualism pervading american culture rendering that tie even tighter. though people buy different tiers of items based on their class, americans, from blue collar laborers to the elite, are united by their voracious appetites for consumption. sociologists and economists, among other scholars, have taken issue with veblen’s theory of conspicuous consumption in the light of new developments in corporate power, the growing importance of consumerism to personal identity, and the moralistic nature of his arguments.4 consumerism can refer to policies that encourage consumption, but for the purposes of this paper, i will be focusing more on consumerism as a cultural rather than an economic phenomenon. since the early 19th century, consumption has taken on deeper meaning in american culture; not only can it elevate social status, but it can also serve as a confidence boost or a representation of patriotic citizenship for many americans. the mall was both a symbol of and a vector for the consumerism that characterizes postwar america in the nation’s historical memory. in her highly influential 2003 book a consumer’s republic, historian lizabeth cohen argues that american society was socially, economically, and physically restructured around mass-consumption in the aftermath of world war ii. she devotes an entire chapter to shopping malls, which “promoted themselves as the new civic centers of booming suburban towns,” as the process of suburbanization shifted the population from urban centers into these new neighborhoods.8 cohen convincingly argues that there was a direct link between suburbanization, the rising tide of consumerism, and the birth of the shopping mall. she points to the feelings of suburbanites — that there were not enough retail options in their new communities, that urban downtowns were not easily accessible by car – and retailers’ response to these feelings as the major reasons behind the growth of the shopping mall. in fact, cohen positions the regional shopping center as a direct result of the growing american consumer appetite, an appetite that was flourishing under the prosperity and security of the postwar era. the regional shopping center served a purpose beyond fulfilling the desire for consumption; it was borne of a desire to forge a community from the atomized residents of the suburbs. despite the flaws in its execution (such as the fact that many malls were de facto segregated due to the discriminatory housing practices of the suburbs in which the malls were situated) the mall was designed with idealistic intentions and in many cases was able to reach some of those lofty goals. architect victor gruen, the “father of the mall,” pioneered the idealized image of the mall. in 1952, in a seminal article written for progressive architecture, gruen outlines his vision for the mall: a shopping center must be more than a mere collection of stores and shops. a shopping center must be even more than its name implies – a center for shopping. the regional shopping center must, besides performing its commercial function, fill the vacuum created by the absence of social, cultural, and civic crystallization points in our vast suburban areas.”17 gruen would go on to design the first enclosed shopping mall in the united states, the southdale mall of edina, minnesota, which opened in 1956. over the course of the next twenty years, his firm designed over fifty more shopping malls across the united states, intending to provide “a new outlet for that primary human instinct to mingle with other humans – to have social meetings, to relax aresty rutgers undergraduate research journal, volume i, issue iv together, to enjoy art, music, activities, the theater, films, good food, and entertainment in the company of others.”18 unlike the downtown area of cities, which had performance venues, shopping options galore, restaurants, as well as other social and cultural pleasures, the suburbs at this early juncture generally only had houses and highways. to gruen, the mall was a panacea to the problems of the american suburbs, which he felt were lacking in culture and, what we now might call a “third place.” 3 the mall as a third place? sociologist ray oldenburg popularized the idea of a “third place” in the late 1980s in his influential book, the great good place. oldenburg describes the term “third place” as the “generic designation for a great variety of public places that host the regular, voluntary, informal, and happily anticipated gatherings of individuals beyond the realms of home and work,” and writes that there is a dearth of them in “newer american communities” like the suburbs or the recently-drained inner cities.10 oldenburg rejects the notion that the mall could serve as a third place in the introduction, pointing out that gruen felt the same way: “[gruen] came to reject the designation ‘father of malling’ because his plan was stripped down to commercialism only. he had envisioned a true community center.”11 the great good place contains many such indictments of the mall such as: “merchandizing, not socializing, [which] marks the character of the mall” and that the mall is “a drifting amalgam of nonpersons; there are no ‘characters’ there,” with a specific eye towards the 80s teen obsession with the institution.12 yet, the fond memories of so many gen-xers, of long days spent by the mall fountain or in the food courts, contradicts his conclusions about the shallowness of the mall experience: the shopping mall was the center of the social world for most american suburbanites who came of age in the last part of the 20th century and even the first bit of the 21st. despite the positive glow of hindsight, this sort of handwringing about shopping malls is not unique to oldenburg, nor unique to the 20th century. the early aughts saw the coining of the term “mall rats” to describe these teens that frequented the shopping mall, although the trend of the mall as a teen social space goes back to the 1980s, if not earlier. the term “mall rat” is derisive and implies bad behavior on the part of the teenagers. for example, the youth cultures in america encyclopedia, published in 2016, defines mall rats as “shopping center patrons, usually teenagers who frequent the mall not simply for shopping, but as a space for social interaction and illicit behavior.”15 these socalled “mall rats” should not be demonized but rather grudgingly admired for their ability to carve out a third place for themselves in a world that was chipping away at their access to such spaces due to trends such as lack of public funding for parks, community centers or transportation, a rising drinking age, and moral panic over crime or stranger danger. were mall rats majoring in “consumership and passivity,” as oldenburg phrases it, or were they simply victims to the escalation of capitalism and government austerity? where some may see these mallrats as vapid, troublemaking teen consumers, i see them as young people creating a community of their own, repurposing the commercial space for their own social needs. the linked phenomena of stranger danger, satanic panic, and worries about the illicit or unsavory activities of unsupervised teens in time wore away at the perception of the mall as a safe space for adolescents. authority figures such as mall operators, law enforcement, and parents began to push for stricter regulation of teenagers in shopping malls, which changed the nature of teen movement in the mall space.7 teens were not the only age group that used the mall as a place for socialization, however, as evidenced by the lasting popularity of mall walking groups among the elderly. “mall walking” is a term used to describe the phenomenon of people, typically senior citizens (although it also can include mothers of young children), who use the mall as a place to walk in a temperature-controlled environment. many times, mall walkers will go in groups, both formal and informal. mall walking can provide a sense of camaraderie, as it tends to be a frequent and regular activity with the same people participating every week.2 like the mall rats, mall walkers may aresty rutgers undergraduate research journal, volume i, issue iv shop, but the mall’s function is more social than financial for them. 4 community creation in the consumer space there has been some pushback on the mall’s portrayal as no more than a shrine to consumerism, though i would argue certainly not enough pushback. in his 1996 book main street revisited, richard v. francaviglia, a scholar of religion, positions the mall as a more evolved, almost utopic version of main street.13 main street was the heart of the american small town, with the main purpose of providing retail options, though it also was a place to foster a sense of community. francaviglia argues that on top of the trappings of main street, the mall has the additional benefits of temperature control, a “flourishing social life,” and evolution into a cultural icon of sorts.13 francaviglia was ahead of his time in this opinion; even though the social community of the mall is currently in a state of decline, nostalgia for this particular social world has never been higher. this is especially notable because there was a massive lack of the appreciation for the social culture of the mall back in the mid 90s when he was writing. despite the mall’s prominence as a social space, it was mainly recognized only for its usage as a site of consumption. others have pointed out that the maligned phenomena of mall walkers and mall rats can be seen as a social good, creating thriving subcultures within the world of the mall.6 in his 1986 book, the new religious image of urban america: shopping mall as ceremonial center, religious scholar ira g. zepp, jr. even argues that the mall is a sacred space, providing meaning, community, and ritual for many americans. in an even bolder claim made ten years later, american historian kenneth t. jackson, despite his misgivings about the resources that they devour, claimed shopping malls are “the common denominator of our national life, the best symbols of our abundance,” and that they are “at the core of a worldwide transformation of distribution and consumption. they represent, along with music, computers, suburbs, and skyscrapers, one of america’s major contributions to twentieth-century culture and life.”5 notable is the fact that these claims were made in the mid-90s when malls were still seen as thriving. francaviglia, zepp, jackson, and to a certain extent cohen were some of the few scholars who grasped the full social significance of the shopping mall. many scholars did not grasp the full meaning of the shopping mall until its current state of decline provided clarity. in this case, absence has truly made the heart grow fonder. 5 decline of the mall shopping malls were not built to last, literally or figuratively. shoddy building materials, rushed plans, online shopping’s meteoric rise, and oversaturation have all combined to the mall’s current state of decline. the suburban shopping mall is becoming a thing of the past; dozens of malls have closed in the past 10 years with even more predicted to follow. though countless business newspapers have breathlessly reported on the “retail apocalypse” and the “death of retail” that these mall closures symbolize, the relationship between a declining retail and the death of the mall is not as clean as that framing would have you believe. for example, sociologist zofia bednarowska argues in her 2018 paper entitled “the consumption space paradox: over-retailed areas next to dead malls” that “over-retailed areas, full of stores of retail chains, such as walmart, macy’s etc. that are usually concentrated in american suburbs, also host dead malls – shopping malls that used to be vital and prosperous, now are just left unused and abandoned.”19 notable as well is the atmospheric rise of online shopping with amazon and its competitors being branded as the malls of the internet. somewhat ironic is the fact that the internet, while causing the death of the community found in many malls, has simultaneously given rise to a virtual third space of sorts: online forums, youtube channels, and blogs dedicated to dead malls. though the rise of online shopping was a major factor in the decline of the shopping mall, leading to the destruction of the communities found there, the internet also serves as a place to build communities to honor the memory aresty rutgers undergraduate research journal, volume i, issue iv of the malls of the past. this trend is likely to continue, given that simon malls, the nation’s largest mall operator, and amazon have been in talks to convert old sears and jcpenney’s stores into distribution hubs.1 there is an irony in the fact that the same behemoth that is destroying the financial viability of the mall could also be the entity that preserves the physical space at least a bit longer. even though these transformations preserve the buildings, they do not maintain the malls’ social or sensory significance. there is something to be said for the physical experience of shopping at the mall, with its abundance of sensory input, which in turn contributes to the strong sense of nostalgia associated with the act. in the mall, one could smell the fried snacks from the food court, hear the sounds of muzak and gossip, and feel the coolness of central air – all experiences lost on the internet. in this increasingly digital age, where one can just click “buy” and get the item the next day, people long for the frictionalized shopping experience and surprising sense of community that the mall provides, not just those that lived through the height of the mall but also the young people with a rosy-eyed view of this era.16 the adolescents of the 2020s, dealing with the isolation of the pandemic and the struggles of socializing on line, see the social culture and sensory experience of the shopping mall as something worth envying. in this way, mall nostalgia is not just felt by those who lived through its heyday but also those who never got the chance to experience that historical moment. 6 conclusion the american shopping mall is an institution that has gone through a lot of changes, both in terms of its reputation and of its function, during its fascinating history. in the 1980s and 1990s in particular, the mall served an important social role for suburbanites. this social role was perhaps most important for suburban teenagers, who were at first encouraged and later demonized as “mall rats” for spending time there. mall walkers, while not as heavily critiqued as their adolescent counterparts, also use the mall as a place to socialize. the mall’s initial purpose, of being a cultural and social panacea to the suburbs, never aligned with its reputation as a shrine to consumerism. despite this disconnect, the american shopping mall should be honored for its role as a place for atomized suburbanites to connect with their community. it is time to reassess the mall’s legacy, and perhaps even revive and repurpose the crumbling behemoths that define the suburban skyline. i think malls could capitalize on the nostalgia for the sensory experience and community aspects of that era. the future of malls may very well be in providing a nostalgic experience or spaces for community events like birthday parties or exercise classes. though nothing can return the shopping mall to its former state of glory (economic decline and online shopping have made sure of that), malls can still serve a social function in american society. 7 references [1] esther fung and sebastian herrera, “wsj news exclusive | amazon and mall operator look at turning sears, j.c. penney stores into fulfillment centers,” wall street journal, august 9, 2020, sec. real estate. [2] georgia perry, “mall walkers: the suburban exercisers keeping america wholesome,” the atlantic, september 21, 2015. [3] harriet h. duncan, shirley s. travis, and william j. mcauley. “an emergent theoretical model for interventions encouraging physical activity (mall walking) among older adults.” journal of applied gerontology 14, no. 1 (march 1995): 64–77. [4] juliet b. schor, “in defense of consumer critique: revisiting the consumption debates of the twentieth century,” doi: 10.1177/0002716206299145. [5] kenneth t. jackson, “all the world’s a mall: reflections on the social and economic consequences of the american shopping center.” the american historical review 101, no. 4 (1996): 1111–21. https://doi.org/10.2307/2169636. [6] kyle riismandel, “arcade addicts and mallrats: producing and policing public space in 1980s america,” environment, space, place, (vol. 5, issue 2, 2013), 66. https://doi.org/10.1177/073346489501400105. [7] kyle riismandel, neighborhood of fear: the suburban crisis in american culture, 1975-2001. baltimore: johns hopkins university press, 2020. [8] lizabeth cohen, a consumer’s republic: the politics of mass consumption in postwar america, (new york: vintage books, 2004), 14. [9] thorstein veblen. the theory of the leisure class. (new york, n.y., u.s.a.) penguin books, 1994. https://doi.org/10.2307/2169636 https://doi.org/10.1177/073346489501400105 aresty rutgers undergraduate research journal, volume i, issue iv [10] ray oldenburg, the great good place: cafes, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community, (new york: marlowe & company, 1989), 17. [11] ray oldenburg, the great good place, xiv. [12] ray oldenburg, the great good place, 118-122. [13] richard v francaviglia, main street revisited, (university of iowa press, 1996), 164. [14] sharon zukin and jennifer smith maguire, “consumers and consumption,” annual review of sociology 30, no. 1 (august 1, 2004): 173–97, doi: 10.1146/annurev.soc.30.012703.110553 [15] simon j. bronner and cindy dell clark, youth cultures in america (abc_clio, llc, 2016), 439. [16] steven kurutz, “an ode to shopping malls,” the new york times, july 26, 2017, sec. fashion. [17] victor gruen and lawrence p. smith, “shopping centers: the new building type,” progressive architecture, no. 6 (jun. 1952): 67. [18] victor gruen and lawrence p. smith, “shopping centers,” 68. [19] zofia bednarowska, “the consumption space paradox: over-retailed areas next to dead malls,” acta universitatis lodziensis. folia oeconomica 5 (338), 23. https://doi.org/10.18778/0208-6018.338.02. eliza peterson is a recent graduate of the rutgers university department of american studies. she is also a research assistant at the education and employment research center. she recently finished writing her senior honors thesis, “‘meet me behind the mall’: reassessing the social and cultural value of the american shopping mall,” which received the american studies best thesis award, as well as a henry rutgers scholar award. her thesis explored the american cultural image of the mall and its legacy through the lens of music, movies, tv, and the internet, focusing on themes of nostalgia. she is currently pursuing a career in education. https://doi.org/10.18778/0208-6018.338.02 aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. topic modeling and analysis: comparing the most common topics in 19th-century novels written by female writers lauren rha, sean silver (faculty advisor) ✵ abstract women authors from the 19th-century have had a profound impact on the literary world due to their critical approach to and inclusion of various social phenomena within their work, such as women's rights, sexuality, and human psychology. this paper seeks to contribute to the discussion by quantifying thematic similarities in eight select novels by various female authors of the 19th-century. these novels were chosen due to their contribution to literature and their popularity, common use in college courses around the world, and the prominence of the female authors. this study included utilizing a programming environment known as r studio to perform a topic model. performing a topic model allowed for the discernment of ten main themes or topics that can be generally seen across all eight selected novels, and by extension, 19th-century literature by female authors. the research found initial evidence to support the general understanding of said literature as an endeavor of the themes of social critique and individual consciousness; however, the results were not absolute in conclusion because of the limited size of the corpus. a larger corpus of documents (novels) is necessary to reach further conclusions. 1 introduction in the 19th-century, iconic texts by female authors revolutionized prose fiction and imaginary work written in narrative form through their criticism of social expectations for women, commentary on the class system and expressions of sexuality, and unique focus on characters’ psychological states and moral developments. this paper seeks to contribute to the discussion by quantifying and analyzing these thematic similarities in eight select novels by the following female authors of the 19th-century: jane austen, charlotte brontë, emily brontë, anne brontë, george eliot, louisa may alcott, and elizabeth gaskell. performing such a topic model on literary studies is not yet widely done, although it has proven to successfully extract themes and topics from wordfrequency data. in doing so, perhaps computationally generated themes can be formed and present a new perspective on common and well-studied themes in the 19th-century literature of various female authors. this research uses an approach known as “distant reading” for its analysis. “distant reading” demonstrates the value of reading large numbers of text together and relies on computer-assisted modeling to analyze a larger amount of texts than can be read by any one individual.[15] “distant reading” and the specific practice of topic modeling was performed within the r studio programming environment. the r studio programming environment is a development environment for r, a programming language used for statistical computing and graphics. the novels analyzed by the author included jane austen’s emma, charlotte brontë’s jane eyre, emily brontë’s wuthering heights, anne brontë’s the tenant of wildfell hall, george eliot’s middlemarch, louisa may alcott’s little women, elizabeth gaskell’s north and south, and george eliot’s the mill on the floss. these eight novels were chosen for their popularity, common usage within college classes, contribution to the discipline of literature, and the prominence of their female authors. this research explains how the novels were prepared for topic modeling and produces a specific topic model visualization, an “interactive heatmap,” to compare the frequency of topics found in each novel. aresty rutgers undergraduate research journal, volume i, issue iii 2 methodology installation of packages & preparing the corpus into a document term matrix the first step, distant reading, involved the preparation of a corpus of documents into a document term matrix. a document term matrix represents the words inside a text as a table (or a matrix) of numbers. the document term matrix format was used to implement topic modeling. in doing so, the novels could be read as data instead of narratives. the text file used included the following novels in one document: jane austen’s emma, charlotte brontë’s jane eyre, emily brontë’s wuthering heights, anne brontë’s the tenant of wildfell hall, george eliot’s middlemarch, louisa may alcott’s little women, elizabeth gaskell’s north and south, and george eliot’s the mill on the floss. all of these texts had similar base content material; they all had a female character as their main protagonist. they are all also well-studied novels throughout the world, as many consider these texts “classics.” these texts were accessed through the digital archive project gutenberg. first, the document was reviewed and all the extraneous metadata was removed. the extraneous metadata included introduction pages, acknowledgments, publishing details, title pages, and the table of contents. the only text within the document was from the actual story of each novel. next, each chapter in the document was labeled in a single continuous document. for example, emma, which has 55 chapters in total, was numbered from 1-55. the following novel was jane eyre, which started at 56 to continue the document. this was done in order to make the coding process and analysis by r studio easier. by combining all of the novels into one text document, there would only need to be one file scanned and coded into the program rather than eight individual ones. each chapter was arranged into a contingency table that organized the most frequent words found to the least. then, a relative frequency table was generated by dividing the total count of every word (token) in the chapter text by the total word count in a chapter. a function, stopword(), which was coded by a professor in the english department at rutgers university – new brunswick, professor sean silver, was used to remove words such as “the,” “a,” and “and,” as well as character names. this was done because otherwise, the most frequently appearing words would be these words, which do not represent any theme that can be applied across all eight texts. the text was then transformed into a document term matrix, ready to be prepared and used through topic modeling. topic modeling: a description topic modeling refers to a coding technique that is able to process and analyze large-scale amounts of data (corpus) into key themes or topics. when analyzing novels through a technological lens, there are often two problems: the ambiguity of definition of certain words (i.e. a word like “fan” holds two meanings: a cooling apparatus and an enthusiast for a particular subject), as well as processing large data sets. the former is understood situationally; the human mind is able to identify, through context and critical thinking, the proper definition of such words. however, the computer often struggles with – to call on the aforementioned example – separating the two types of “fans.” in this case, a method known as “key word in context” is used. this method allows the programmer to see the word within the context of the text, so as to better discern what the proper definition for the word is. conversely, unlike the computer, it is very difficult for humans to read and interpret incredibly large amounts of data.[3] as a result, it was imperative that the technique known as topic modeling was used to address and offer solutions for both of these concerns. within topic modeling, the identified largescale document is “read” by the computer and separated into key themes or topics. these topics are determined by the likelihood that certain words will appear together within the novel; each topic consists of words that have similar meanings or often emerge in context of each other. this is known as “collocation” and is essentially equivalent to concepts or themes. after the identification of major themes, the estimated proportions of the topics within the documents are identified. in this research, a visualization was then generated: an interactive heatmap. the in aresty rutgers undergraduate research journal, volume i, issue iii teractive heatmap is a graphical representation of the data where each of the individual values that are contained within the matrix is represented as different colored squares. when hovering over each colored square, the row label, column label, and calculated value of the data appears. in this way, topic modeling hovering allows for the organization and summarization of “electronic archives at a scale that would be impossible by human annotation.”[3] the number of topics/themes one can generate (“k”) is arbitrary and up to the discretion of the programmer. if the “k” value chosen was too small, the topics generated would be too broad to classify as “main themes.” however, if the “k” value chosen was too large, there would be too many overlapping topics. figure a: interactive heatmap of topics most frequently found in 19th-century novels by female authors generated interactive heatmap of topics most frequently found in 19th-century novels. figure b: examples of the interactive square. each interactive square displayed the name of the novel, the topic, and a value that represented the sum of all of the probability values generated in the topic modeling sequence. aresty rutgers undergraduate research journal, volume i, issue iii figure c: the 10 generated topics in trial 5. figure d: the 10 generated topics in trial 6 for comparison. aresty rutgers undergraduate research journal, volume i, issue iii topic modeling: implementation latent dirichlet allocation (lda), the simplest topic model, was employed within this research, as the “intuition behind lda is that documents exhibit multiple topics [themes].”[3] in lda, “all the documents [in this case, the novels] in the collection share the same set of topics, but each document exhibits those topics with different proportion.”[11] i chose to have lda generate ten topics/themes (𝑘𝑘 = 10) and to return the top 25 terms from each topic and the estimated proportions of topics in the documents. because the corpus of documents was divided into chapters and not novels, i had to create a separate code to differentiate certain chunks of chapters as certain novels. for example, i coded for chapters 1-55 to be for emma, chapters 56-93 to be jane eyre, and so forth. using this data, i then plotted the generated values that show the estimated proportions of topics in each novel. there were ten total trials of the interactive heatmap and topic modeling performed. 3 results after performing ten trials of the interactive heat map and topic modeling, although there was a slight variation in the order of words and changes in a few number of terms each trial, results displayed that the generated topics and terms were generally the same. trial 5 was randomly chosen. any trial would have sufficed because there was stability between the various trials and the generated themes. in general, there were not many topics that were found prominently amongst all of the novels – only topic 6 was found averagely within all of the novels. instead, there were concentrations of certain topics in certain novels. this can be seen in topic 3’s prevalence in emma and topic 1 and topic 4’s prevalence in middlemarch, and so forth. experimentation on this topic has led to the conclusion that the most popular word in a topic changes about forty percent of the time between essentially identical topic models. the most significant and comprehensive themes amongst the novels included topic 4 and topic 6. topic 4 touched upon much of women’s worth within the social constructions of the 19th-century (“husband,” “life, “marriage,” “self”), while topic 6 was focused more on the mind and body (“eyes,” “face,” “look,” “heart,” “hand,” “felt,” "voice”). the rest of the topics – although they were found in some capacity amongst all the novels, were generally only concentrated within one or two novels and did not show a real, significant, or comprehensive theme. instead, it can be understood that the remaining themes mainly consisted of a general melting pot of the most commonly appearing words within each respective novel (perhaps due to the plot, characterization of certain characters, and/or an individual writers’ writing style as opposed to chosen themes). 4 discussion although it is difficult to come to concrete conclusions about shared themes/topics because the corpus was small in comparison to analyzing thousands of novels written by female authors in this century, i still found some support through the generation of topics based on word frequency for our understanding of 19th-century literature by female authors as an endeavor of social critique and individual consciousness of women during the 19th-century. using the technique of topic modeling (interactive heatmap), i assessed eight texts by seven unique anglophone female writers. multiple trials were performed in order to distinguish and observe which themes appeared in highest frequency, as well as to assess whether there was a notable difference in the generated themes for each respective trial. although there was slight variation in the order of words on the spreadsheet and a few number of terms changed each trial, the results repeatedly showed the keywords to be virtually the same (figure c compared to figure d). for example, if one trial had the word “family” within a topic about agreeableness and neighborly interaction (topic 1 in figure c), another trial may not have the specific word of “family,” but still had keywords such as “quite,” “father,” “woman,” and “little.” this is due to the nature of topic modeling. within topic modeling, a label is automatically assigned to each topic based on its top term. then, the code changes the topic model in various ways to see if the decided label turns up again.[16] 19th-century literature is often characterized aresty rutgers undergraduate research journal, volume i, issue iii by its understanding of the psychological state within stressful and restrictive social environments or when confronted by moral dilemmas.[13] 19thcentury literature dealt with ideas of moral decline and development and addressed the different modes and extent of control within certain psychological, social, cultural, and moral frameworks. it “delve[d] [into] the internal life of the characters… through a focus on the thoughts and motivations of the characters rather than their occupations and external settings alone.” [13] topic 6 is indicative of this; it included words such as “felt,” “mind,” “love,” “eyes,” “face,” “stood,” and “looked.” as seen from the interactive heatmap, this topic is found averagely in all of the texts. topic 6’s words have an emphasis on mind and body. words such as “felt,” “mind,” and “love” represent the psychological state and emotion, while words such as “eyes,” “face,” “stood,” and “looked” are grounded to the physical body; this connection between mind and body can be interpreted as representations of self-control. within sally shuttleworth’s study, charlotte bronte and victorian psychology, she meditates on this idea of self-control, writing: “rigorous control and regulation of the machinery of mind and body would offer a passport to autonomous selfhood and economic liberty.”[14] however, she also importantly notes that there are “two conflicting models of psychology found in victorian economic discourse: the individual is figured both as an autonomous unit, gifted with powers of self-control, and also as a powerless material organism, caught within the operations of a wider field of force.”[14] in considering this, topic 6, – rich with words such as “felt” and “look,” – strengthens the general understanding and characterization of 19th-century literature and its preoccupation with the consciousness and social standing in society. within these novels, “males hold positions of political, institutional, and sometimes of economic power denied to females,” which can be seen in topic 4, with words such as “husband,” “feeling,” “self,” “marriage,” “world,” “living,” and “experience.”[14] this topic touches on the aspect of 19thcentury society that defined much of a woman’s legal and social worth. although this connection is in line with our understanding of victorian society, it was significant that topic 4 on women’s worth based on social construct was not as prominent as topic 6 on the mind and body in all of the texts. this showed the authors’ intent to put an emphasis on the psychology and inner world of these female characters, rather than their marriage, in defining them and giving them a voice as individuals. the prominence of topic 6 demonstrated that “females hold a kind of psychological and moral power that is exemplified in their status as paradigmatic protagonists.”[12] in addition, within a quantitative study conducted by john a. johnson, joseph carroll, jonathan gottschall, and daniel krueger, the data concluded that within 19th-century victorian literature, female protagonists scored the “highest of agreeableness,” among other characteristics. agreeableness was defined as a “pleasant, friendly disposition and tendency to cooperate and compromise, versus a tendency to be self-centered and inconsiderate”.[12] this tendency toward altruism is “more strongly associated with maturity, [and]… as vehicles for improving society.”[12] although this could be understood as a great emphasis on the improvement of society as a whole, it also demonstrates the female characters’ need to be overwhelmingly “pleasant” within society. it was unsurprising that topic 3 of my research included words that evoke some sort of community or degree of agreeableness amongst neighbors, such as “quite,” “little,” “time,” “sure,” and “dear.” as i was interested in understanding the proper context of these words, i performed a keyword in context. words like “quite” were used to express full and absolute agreeableness on a subject; examples include: “to feel quite sure,” “something quite fresh,” and “you do quite right.” similarly, the keyword “sure” evoked ideas of obliging mannerisms; examples include: “to be sure,” “i am sure of having their opinions with me,” and “i am sure you are a great deal too kind.” the heavy concentration of this “agreeable ness” theme within the first novel, emma, is quite appropriate of the text, as it is a novel about a young female and her social life, experiences, and relationships within the small town of highbury. although the novel is an exploration of emma’s relationship aresty rutgers undergraduate research journal, volume i, issue iii with her mind and body (ideas of maturity and development), she comes to develop her understanding of self-control and her emotional state through her interactions with others. as mentioned before, i still found certain words and, thus, topics that were, understandably, more concentrated within certain novels. as just discussed, topic 3 was very concentrated within emma. although this was the most prominent case, in general, there seemed to be topics that were strongly associated with one particular novel over the others. 5 conclusion this research aimed to add to the current discussion surrounding 19th-century literature by female authors by quantifying and analyzing various thematic elements within eight select novels. these themes included social expectations for women, commentary on class and expressions of sexuality, and a focus on the protagonist’s psychological state or moral development. this research is significant because it allowed a new method of analysis of 19thcentury literature by female authors. the method of topic modeling utilized within the r studio coding environment is effective in generating themes based on word-frequency data, but it must be noted that there are limitations to this method of analysis as it completely rejects the traditional format of reading novels. within novels, there is a plot and specific storyline. the novel is organized and divided in a way that best supports the development of this plotline. however, topic modeling does not discern nor care for the sequence of the story and is instead only interested in quantifying the text – or “data” – and extracting and putting certain groups of words into “themes” or “topics” so as to quantitatively display the author and novel’s thematic intent. this research would have benefited from a larger corpus of works in order to come to more conclusive, specific, and generally applicable themes to be applied to novels during this time period. in ad dition, although words such as “you,” “i,” and “it” were included within the stopword() function, they still appeared in the results. they were joined by a strange symbol: <e2>, which i was regrettably unable to remove. this was most likely due to oversight within the coding script. ideally, the analyses would be redone with a fully cleaned corpus because the appearance of the <e2> symbol limits the ability to draw conclusions and undermines the interpretability of the topic wordlists. although the results of this analysis on the famous novels of 19th-century female writers were telling, they were not absolute in their conclusion and lacked a large enough corpus to properly determine the main themes amongst a large number of victorian novels. if repeated, this research should take on a larger number of novels. there was some evidence of the societal restrictions imposed on women during this time period (topic 4), yet conversely, there was also some evidence of the authors’ intent to dispel these societal, political, and legal factors in defining the female individual by defining them through their mind, consciousness, and body (topic 6). evidence showed the importance of psychology, the body, and the inner worlds of their female protagonists in relaying the stories of women and giving them the opportunity to voice their thoughts and feelings through literature∎ 6 acknowledgements thank you to professor sean silver of the english department at rutgers university—new brunswick for his assistance and encouragement of this research. aresty rutgers undergraduate research journal, volume i, issue iii 7 references [1] alcott, louisa may, little women. (melbourne; london; baltimore: penguin books, 1953) https://www.gutenberg.org/ebooks/514 [2] austen, jane, emma. (london; new york: penguin books, 2003) https://www.gutenberg.org/ebooks/158 [3] blei, david. “introduction to probabilistic topic models.” (2011) https://www.eecis.udel.edu/~shatkay/course/papers/uintrototopicmodelsblei2011-5.pdf [4] brontë, anne, the tenant of wildfell hall. (oxford: blackwell, 1931) https://www.gutenberg.org/ebooks/969 [5] brontë, charlotte, jane eyre. (new york: penguin books, 1985) https://www.gutenberg.org/ebooks/1260 [6] brontë, emily, wuthering heights. (london; new york: penguin books, 2003) https://www.gutenberg.org/ebooks/768 [7] burrma, rachel. “the fictionality of topic modeling: machine reading anthony trollope's barsetshire series.” big data and society 2, no 2 (2015) https://works.swarthmore.edu/fac-english-lit/286 [8] eliot, george, middlemarch. (london: vermont: j. m. dent; charles e. tuttle, 1997) https://www.gutenberg.org/ebooks/145 [9] eliot, george, the mill on the floss. (edinburgh; london: w. blackwood and sons, 1860) https://www.gutenberg.org/ebooks/6688 [10] gaskell, elizabeth cleghorn, north and south. (harmondsworth, penguin, 1970) https://www.gutenberg.org/ebooks/4276 [11] jockers, m. l., and mimno, d. “significant themes in 19thcentury literature.” poetics 41 (2013): 750–769. [12] johnson, j. a., et al. “portrayal of personality in victorian novels reflects modern research findings but amplifies the significance of agreeableness.” journal of research in personality (2010) http://personal.psu.edu/faculty/j/5/j5j/papers/victorianpersonality.pdf [13] sen, debashish. psychological realism in 19th century fiction: studies in turgenev, tolstoy, eliot and brontë. (united kingdom: cambridge scholars publishing, 2020). [14] shuttleworth, sally. charlotte brontë and victorian psychology. (new york: cambridge university press, 1996). [15] underwood, ted. "a genealogy of distant reading.” digital humanities quarterly, 11, no. 2, (2017) http://www.digitalhumanities.org/dhq/vol/11/2/000317/000317.html [16] yang, y., et al. “the stability and usability of statistical topic models.” acm transactions on interactive intelligent systems, 6, no.2, article 14 (2016) https://dl.acm.org/doi/abs/10.1145/2954002?download=true lauren rha is a rising first-year graduate student at the rutgers graduate school of education. she hopes to become an esl teacher for immigrants and refugees coming into the united states. her interest in and study of 19th-century literature motivated her to research common themes across eight famous victorian novels. the research presented was done as a final project within professor sean silver's english 315 class during her second semester as a junior, where she learned the fundamentals of basic coding for literary analysis within the r studio environment. https://www.gutenberg.org/ebooks/514 https://www.gutenberg.org/ebooks/158 https://www.eecis.udel.edu/%7eshatkay/course/papers/uintrototopicmodelsblei2011-5.pdf https://www.eecis.udel.edu/%7eshatkay/course/papers/uintrototopicmodelsblei2011-5.pdf https://www.gutenberg.org/ebooks/969 https://www.gutenberg.org/ebooks/1260 https://www.gutenberg.org/ebooks/768 https://works.swarthmore.edu/fac-english-lit/286 https://www.gutenberg.org/ebooks/145 https://www.gutenberg.org/ebooks/6688 https://www.gutenberg.org/ebooks/4276 http://personal.psu.edu/faculty/j/5/j5j/papers/victorianpersonality.pdf http://www.digitalhumanities.org/dhq/vol/11/2/000317/000317.html https://dl.acm.org/doi/abs/10.1145/2954002?download=true aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. agency in the writing center: examining the importance of student autonomy in higher education ryan rodriguez abstract in college-level english courses, students often struggle to achieve satisfactory results in their writing. to remedy this, they seek help at campus writing centers, where a tutor helps them improve their writing skills and their academic performance. yet, students experience tension between the classroom and the writing center that universities should seek to minimize. in my research, i discovered how different learning methods may either foster or suppress student autonomy. further, i found that current methods—such as the course-embedded model for mitigating the tension between the writing center and the classroom—fail to empower the student. using rutgers university and its style of minimalist tutoring as a benchmark, i discuss the topics of autonomy and agency, student-led negotiation with authority, lack of academic motivation, and how we can bridge the pedagogical gap between the writing center and the classroom. introduction common on university campuses is the writing center, an institution for students in search of help with writing-based coursework. assigned to tutors, students can intimately collaborate with a peer rather than their classroom professor. the writing center’s aim is two-fold: first, to offer immediate support to students struggling in their english classes specifically, and second, to improve students’ general compositional skills. distinct from the classroom, traditional writing centers, such as those provided at rutgers university, are supplemental resources used solely at the discretion of the student. a student may thrive in a writing course and still seek tutoring, or may fail paper after paper but still decline assistance. the writing center is an open invitation, advertising free help to all students in writing courses who need it and, more importantly, who actively want it. what then becomes the subject of debate is how colleges can effectively implement writing centers. such discourse heavily revolves around how differing approaches to tutoring impact students’ improvement. however, the discussion insufficiently examines how the classroom and writing center environments noticeably differ, and to what extent the former might actually impede the latter. a writing center tutor’s instructional style differs greatly from a student’s https://creativecommons.org/licenses/by-nc-sa/4.0/ https://creativecommons.org/licenses/by-nc-sa/4.0/ aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 2 primary professor. further, whereas a writing center is an optional resource, the classroom is a non-negotiable, mandatory space. these distinct environments demonstrate how students operate differently in the presence and absence of authority. given these issues, how can we make these individualized, mentor-based writing centers function effectively alongside their classroom counterparts? in other words, how can universities minimize the tension students experience between the two environments? to explore this, i examined both existing literature and my own experiences as a tutor at the plangere writing center at rutgers university. i found that writing centers promote student autonomy in a way the classroom does not: the renewed sense of self-sufficiency the student receives generates significant motivation to succeed academically. and, while there are methods that bring both environments closer together, such as the course-embedded model, maintaining the distinct identity of each environment is crucial in giving students autonomy. a collaborative, peer-led learning process is one students will rarely encounter within traditional classroom settings. so, if the hierarchies present in the classroom are extended into the writing center, its benefits are eliminated. in this essay, i will indicate the distinctions between these two environments to help better understand how each functions in relation to the student. furthermore, i will offer steps educators can take to bridge the gap between them. before moving forward, important clarifications are necessary. first, the following arguments pertain to the accepted standard of minimalist tutoring that has spread throughout higher education since jeff brook’s (1991) initial introduction of the concept. this style rests on the student’s ability to learn autonomously, as well as the tutor’s ability to promote this autonomy. second, my criticism of the classroom is not meant to be perceived as harsh; my intent is merely to highlight present flaws that inhibit student autonomy. negotiating autonomy and authority autonomy describes the extent to which a student experiences self-determined independence separate from outside influence. troisi (2015) states that “people feel autonomous when they perceive their actions emanate from themselves rather than outside sources,” and that “[f]eeling controlled by others rather than by the self undermines a person’s sense of autonomy and her/his intrinsic motivation to perform tasks” (84). passivity, on the other hand, occurs when students cannot explore avenues of personal interest. in such instances, instruction is severely limited in scope; it does not consider how students may want to learn but instead what they must learn. unsurprisingly, this approach does not inspire motivation and interest in course material. van lier (2008) offers a method of assessing students’ relative passivity and autonomy. the proposed 1-6 scale ranges from the least ideal, passive, to most ideal, committed. “committed” students “voluntarily enter into a debate with one another and create a collaborative agency event” (3). a “committed” student engages in the classroom and is fully invested in discussion. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 3 of course, this is not limited to discussion-based classroom settings; it can be expressed in any environment where learners can interact with the instructor and the material being learned— whether it be answering questions, collaborating with peers on assignments, communicating with an instructor during office hours, and so on. most importantly, voluntary contributions demonstrate the student’s direct engagement and genuine interest in their education. writing centers are designed to cultivate such voluntary participation. thus, rutgers and countless other higher education institutions implement a “minimalist” approach to tutoring. in a guidebook for rutgers writing center tutors, the role of the writing center is described as such: "making the student do all the work" has become the mantra of what is called "minimalist tutoring," as first described by jeff brooks and as practiced at the rutgers writing centers (and at practically every writing center across the country)… in the minimalist model, the writing tutor is a mentor, coach, or task master who guides students through the process of revision and helps keep them focused on the project at hand by breaking it down into smaller and more manageable tasks. (goeller, kalteissen 2008, 7) this method necessitates student autonomy: if the student does not want to improve, the method falls apart. yet this generally does not happen. whether influenced by extrinsic motivations (e.g. good grades) or intrinsic motivations (e.g. personal development), a student who enrolls in tutoring already demonstrates a desire to improve. thus, in my research, i found that students were willing to intimately engage with their assignments. when they committed to finding solutions for themselves, in direct collaboration with their tutor, they exhibited increased autonomy. the writing center allows the student to speak their mind and to question what they think is right or wrong. however, i fear that students only experience such motivation during tutoring sessions, since the traditional classroom is less conducive to student autonomy. it works within the hierarchical system a university imposes, and there is little negotiation available. thus, college-level curricula allow little to no opportunity to restructure the pedagogy, especially in larger classes that move through material quickly. teachers often instruct what to do rather than giving students a range of choices; rogoff et al. (2003) characterize it as “assembly-line instruction” with no community among students or active participation in the course material. without a sense of choice, student passivity sets in. such a setting does not permit the individualistic approach some students require. ultimately, students suffer as a result: freeman (2014) revealed the stunning disparity in academic success between learning in the traditional classroom versus an active learning environment, finding similar results across various disciplines (see figs. 1 & 2). aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 4 these writing center and traditional classroom also differ in that the classroom has a much stronger authority figure than the writing centers. healy notes, “the advantage tutors have over teachers in this enterprise is that even if students try to invest tutors with authority, tutors can resist that role, while teachers, as long as they give grades, have a harder time shedding their image as authority figures” (21). additionally, kail and trimbur (1987) indicate that “peer tutoring based on collaborative learning taps into the network of mutual aid already based in student culture,” including study groups, shared homework problems among friends, and so on (9). the level of comfort students experience due to tutor-student rapport affects student motivation. in managing classrooms, professors generally cannot be a personable guiding hand for each student, whereas tutors give the student a space where they are not talked down to or have to satisfy curricular expectations. levels of comfort should be and are achieved in both settings, of course, but tutors find themselves in a favorable position: lacking true authoritative leverage, they exist on an even playing field with their students. this dynamic ultimately lets students be truer to themselves, searching for the approval of a role model instead of an authority figure. the primary differences, then, between the writing center and classroom environment are the extent to which each encourages student autonomy, and the nature of each environment's authority. where the instructor in a classroom is the singular, ultimate authority figure, the tutor is a peer with a guiding role. where the classroom is a group of young adults under the professor’s supervision, the writing center is a personal conference with a mutual interest in improvement. in the classroom setting, healy (1993) writes that students “see themselves as acted upon rather than as agents of their own destiny, as subject to the authority of others rather than subjects of their own academic journey” (17). conversely, healy states that “writing centers, as ‘semi-autonomous’ spaces, can serve as one of those forms of social organization through which students learn to negotiate issues of authority and learn to take more responsibility for their own learning” (18). while authority defines the classroom, cooperation defines the writing center. where one cultivates student passivity, one fosters student autonomy. students react differently depending on the environment they inhabit, and the writing center stimulates independent student thought to a degree the classroom does not. writing centers allow students to determine the path of their writing, not under the supervision of an instructor, but with the guidance of a peer. active participation in the writing process offers students the opportunity to act independently. removed from passive instruction in the collective setting, they can fully engage in a one-on-one conversation that restructures the typical educational hierarchical structure. thereby, they achieve autonomy. it would be neglectful not to mention that the classroom is not entirely—as i have characterized it until this point—an environment disagreeable to effective education, nor is it a place where autonomy is perpetually nonexistent. as i stated previously, the writing center is a supplementary resource to the classroom whereas the classroom is vital to the learning process. baringer and mccroskey (2000) reported that the immediacy experienced in a traditional classroom heightens appreciation for learning (178). alexander astin framed student aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 5 involvement as the “physical and psychological efforts students put forth toward their academic experiences,” ultimately defining the classroom as the major factor in a student’s educational experience. with this in mind, vincent tinto noted that “students only formally experience education in the actual classroom” (sidelinger 2010, 88). there is no doubt then that the classroom affords a beneficial learning environment. additionally, it develops an array of social skills. what becomes important is how to optimize its implementation, using resources like campus writing centers to improve the student’s educational experience. unfortunately, students often experience tension between these two environments. i found that, when working with such authority figures as professors, students often were more concerned with pleasing their instructors than improving their writing. rutgers sophomore a.p. related his personal account of tutoring. “instead of taking the risk of arguing for a less favorable but more interesting side,” he said, “i went into tutoring sessions asking the question, ‘do you think my professor will like this [phrase] or [that one] more?’ this, in turn, changed the sessions from how to improve my writing so that i am able to do it on my own to both my tutor and i collaborating in ways to impress my professor.” rutgers sophomore k.d. similarly described tension between her professor and her tutor, feeling “hindered by the way [she needs] to balance the two” audiences. clearly, the relationship between these environments must be more effectively mediated. applying student autonomy: columbia college addressing the tension students experience between the writing center and the classroom begins with understanding that they operate differently despite their symbiotic relationship. bridging the gap between the two entails improving the student’s role in this cooperative relationship. and forging associations between them is not difficult by any means; it only requires the effort necessary to address the needs of students who visit the writing center. in on location: theory and practice in classroom-based writing tutoring, ottery and colleagues (2004) conducted a case study where tutors, teachers, and writing center consultants in chicago’s columbia college collaborated with student learners in the writing process. writing center consultants found that “[w]orking and meeting with instructors and the program coordinators offered tutors a chance to affect procedure and pedagogy, as they were able to provide [teachers] with information about how the students were reacting to the class” (ottery et al. 2004, 63). this study shows how students become more or less forthcoming depending upon the power dynamic of their learning environment. the tutors could then communicate with their students’ teachers, who gained insight they would not otherwise have access to. instructors could use this new information to change how they address student’s needs. conversely, writing centers at rutgers, as is the case at many other universities, remain completely separate from the classroom. they may exist within the rutgers english department’s writing program, and professors may advertise them as an optional resource; however, characterizing them as a direct extension of the classroom is inaccurate, since they do aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 6 not directly consult with writing-based courses. tutors do not communicate with students’ professors, thus denying students the benefits demonstrated by the columbia college study. one example, from rutgers english professor michael duffy, recounts a time where this disconnect ultimately hurt a student’s grade: i had a situation with a student a few years ago who was allegedly "misled" by their tutor to the extent the tutor allegedly gave the student bad revision advice on their final 201 paper. the tutor's alleged advice made the student's claims more general and the student earned a grade they were not happy with. i should also say that the advice contradicted my comments on the student's draft. [however, t]he student's claim that "my tutor thought it was a good idea" isn't a good excuse because, like i said, the student must take ownership of their work—it's the student who earns the grade. of course, a student is responsible for the work they submit. the inherent problem here, however, is the contested loyalty students experience. a student meets with a rutgers writing center tutor for a weekly, 80-minute session. consequently, the rapport established between a tutor and a student builds, making it likely that students feel a closer attachment with their peer tutors than their professors. in the account above, a student bypassed explicit instruction from their professor and opted to heed the advice of their peer. while the fault still lies with the student, students should not be forced to choose which party to listen to. adopting a columbia college-like model eliminates contradictions between the professor and the tutor. ottery, therefore, highlights the advantages of improved classroom-writing center communication. additionally, ottery’s example of accessibility not only provides an actual pathway linking the writing center and the classroom, but also gives the student the individualistic attention they need in order to succeed academically. current discussion much of the current discourse on writing center pedagogy examines the pros and cons of different tutoring approaches. these approaches include traditional, independent writing centers as well as other integrated models, such as classroom-based tutoring or the courseembedded model, which will be discussed in the next section. however, discourse fails to acknowledge how significantly classrooms directly impact tutors and the writing center. in other words, the conversation skews toward approaches to tutoring itself instead of questioning prominent external influences that impact the student. for instance, corbett (2013) asks, “what can tutors do when they find themselves more closely attached to writing courses or curricula, to teacher expectations, [or] to student concerns?” (82) instead of asking about the student’s external influences. acknowledging these problems, rather than focusing exclusively on tutoring methods, gives a wider scope to the discourse—it considers the whole of the student rather than devoting all attention to the work alone. similarly, rutgers writing centers work to improve student writing skills without acknowledging why they are successful in the first place. but i believe this is an area in which rutgers can enhance its approach—one that is already aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 7 effective in its application of autonomy. malavika amenon, a tutor at plangere writing center, believes that although a lot of tutoring is “common sense,” how tutors are taught to help students is slightly flawed. she argues that tutor training “should give us a trial session to better prepare us,” and that the internship “should be more hands-on in focusing how we can be effective tutors instead of learning how to write essays that reach teachers’ expectations.” as a tutor myself, i agree with her input. granted, rutgers’ writing centers do functions exceptionally well. but although rutgers’s minimalist approach is effective in promoting autonomy, tutors and students alike need more guidance on how the writing center fits into the grander scope of learning. up to this point, i have argued that ignoring this significant facet to the writing-based tutoring model inhibits the effectiveness of our student aid. without addressing this detail, tutors cannot fully understand the dynamics at play when a student feels tied between two different figures of authority, two different environments, and two different modes of thinking. thus, i hope to elucidate conflicts between a pair of environments whose contradictory approaches place students at a disadvantage. the course-embedded model significant efforts have already been made to bring the classroom and writing center closer together. the proposed and widely accepted “course-embedded model” enjoys much praise (carpenter, whiddon, and dvorak. 2014; squibb, mikkelsen 2016; pagnac et al. 2014; henry, bruland, and sano-franchini 2011; titus et al. 2014). but while its goal is to link classroom instruction and tutoring more intimately, the model ultimately caters to the classroom and relegates the writing center to a subsidiary role. building on speigelman and grobman’s on location (2004)—a foundational work for bridging classroom and writing center pedagogy—carpenter and colleagues (2014) highlight the importance of layering writing center instruction within the classroom curriculum. they express displeasure with the traditional idea that the writing center and classroom should remain separate, stating, “course-embedded programs encourage teachers and tutors to work at the intersections of writing center and writing classroom spaces, creating unique opportunities and challenges for both parties” (3). yet in doing so, the model establishes the student’s curriculum as the main focus of the learning process. thus, universities seize the writing center’s role and appropriate it for the classroom. by doing so, this model supplants the autonomous space for students with aspects of the classroom that promote passivity. kail and trimbur (1987) write that “[t]he curriculum-based model operates through official channels,” and rather than giving them full responsibility, it “makes peer tutors an extension of the writing program” (6). building upon this, healy (1993)—equating peer tutors in a course-embedded model to tas—also argues why students deserve tutoring removed from the constraints of the classroom: aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 8 [s]tudents [...] deserve direct support that is unmediated by the instructor, and that support can be better provided by a tutor than a ta—by someone, in other words, who is not affiliated with the course. by standing apart from the classroom, tutors provide a means of interrogating academic hierarchy. they provide an audience whose relationship to a student’s writing is not governed by the same kind of “oughtness” as is the instructor’s. (20) beholden to the course curriculum, the course-embedded writing center model takes away choice from the student. students are forced to engage with tutoring rather than personally deciding if it is best for them. students cannot possibly achieve lier’s “committed” status if their choice is taken from them. furthermore, the dynamic between students and tutors greatly changes, since the tutor is more invested in advancing classroom standards than establishing a rapport with the student. the tutor becomes synonymous with the classroom, and tutors’ goals are torn between satisfying the needs of the instructor and the student. such blurring of boundaries effectively erases the writing center altogether in favor of the classroom. thus, squibb’s (2016) study concluded that the learning acquired from the course-embedded model instruction “did not translate to higher student achievement as represented by course grades and grade point average” (164). yet the course-embedded model still receives praise. this discrepancy can be attributed to “chaos [that] sometimes clouds writing center administrators’ ability to accurately account for the factors that contribute to the successes and failures of course-embedded writing support” (webster, hansen 2014, 52). such chaos is caused by resistance to the implementation of the course-embedded model. webster and hansen rightfully point out that tensions exist between tutors and faculty when renegotiating their authoritative hierarchy, indicating that instructors may resist yielding administrative duties to tutors in this new dynamic. amid current praise for this newer method of writing center tutoring, we lose sight of thoughtful critiques regarding such an ambitious model’s application. as andrea lunsford (2000) reminds us, “[w]e shouldn’t fool ourselves that creating new models of authority […] is a goal we can hope to reach in any sort of straightforward way” (71). the model also necessitates training tutors to exist within the curriculum of the writing course. when universities implement a course-embedded model, they must dedicate resources to ensure tutors meet certain standards. tutors may be expected to “attend class and sometimes engage in teaching activities by conducting workshops, forming peer-response groups, and providing individual conferences during class time” (carpenter, whiddon, and dvorak 2014, 4). while these expectations improve student learning, they also place an undue burden on tutors. such a burden, on top of students’ preexisting course load, discourages potential tutors from enlisting in the writing center’s ranks. rutgers university’s current system of tutoring makes it virtually impossible for such a model to function; it is simply not feasible to require such an extreme investment from tutors. ultimately, large universities like rutgers cannot enlist student workers without interfering with their personal schedules, and it is one crucial reason aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 9 separation of classroom and writing center are required. difficult as it is to admit, it is not always about the methodology in theory but also its practicality. conclusion the main takeaways from this discourse should be that the classroom and writing center function very differently, and that autonomy and independence are important for student success. to attain this goal, we must strike a balance between these two separate entities. the solution lies between two extremes: total separation and the course-embedded model. we should aim to open channels of communication between the writing center and the classroom, without letting the classroom supersede the supplementary writing center. to this end, tutors and teachers alike need to cultivate a dedicated commitment to student agency—practicing encouragement instead of coercion. adopting columbia college’s method, or a variation on this approach, is an appropriate measure that links the classroom and the writing center in a firmer relationship. importantly, primary teacher/secondary tutor correspondence forms a bridge while keeping them distinct. this is the direction we should take and one we hope to see more readily employed throughout (but not limited to) higher education. students feel frustrated obeying the education system. they struggle on the path toward success, and they often know they need additional aid to achieve their goals. so, it is up to educators to create stimulating environments that inspire students’ drive to learn and respect their autonomy. fostering this environment in the classroom is a challenge; but once it is achieved, the opportunity to effectively teach students presents itself with stunning clarity. references baringer, doreen k., and james c. mccroskey. "immediacy in the classroom: student immediacy." communication education 49, no. 2 (2000): 178-186. brame, c. "active learning." vanderbilt university center for teaching (2016). brooks, jeff. “minimalist tutoring: making students do all the work.” writing lab newsletter 15.6 (1991): 1-4. print. carpenter, russell, scott whiddon, and kevin dvorak. "guest editor introduction: revisiting and revising course-embedded tutoring facilitated by writing centers." praxis: a writing center journal (2014). corbett, steven j. “negotiating pedagogical authority: the rhetoric of writing center tutoring styles and methods.” rhetoric review 32 (1): (2013) 81–98. freeman, scott, sarah l. eddy, miles mcdonough, michelle k. smith, nnadozie okoroafor, hannah jordt, and mary pat wenderoth. "active learning increases student performance in science, engineering, and mathematics." proceedings of the national academy of sciences 111, no. 23 (2014): 8410-8415. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 10 goeller, m., and k. kalteissen. "the task: a guide for tutors in the rutgers writing center." (2008). hall, emily, and bradley hughes. "preparing faculty, professionalizing fellows: keys to success with undergraduate writing fellows in wac." the wac journal 22 (2011): 21-40. healy, dave. "a defense of dualism: the writing center and the classroom." writing center journal 14, no. 1 (1993): 16-29. henry, jim, holly huff bruland, and jennifer sano-franchini. "course-embedded mentoring for first-year students: melding academic subject support with role modeling, psychosocial support, and goal setting." international journal for the scholarship of teaching and learning 5, no. 2 (2011): n2. kail, harvey, and john trimbur. "the politics of peer tutoring." wpa: writing program administration 11 (1987): 5-12. lunsford, andrea a. "refiguring classroom authority." the ethics of writing instruction: issues in theory and practice 4 (2000): 65. ottery, jim, jean petrolle, derek john boczkowski, and steve mogge. "writing and reading community learning: collaborative learning among writing center consultants, students, and teachers." in on location: theory and practice in classroom-based writing tutoring, edited by spigelman, candace and grobman, laurie, 60-71. university press of colorado, (2005). pagnac, susan, shelley bradfield, cyndi boertje, elizabeth mcmahon, and gregory teets. "an embedded model: first-year success in writing and research." praxis: a writing center journal (2014). rogoff, barbara, ruth paradise, rebeca mejía arauz, maricela correa-chávez, and cathy angelillo. "firsthand learning through intent participation." annual review of psychology 54 (2003). sidelinger, robert j. "college student involvement: an examination of student characteristics and perceived instructor communication behaviors in the classroom." communication studies 61, no. 1 (2010): 87-103. squibb, sara davidson, and susan mikkelsen. "assessing the value of course-embedded information literacy on student learning and achievement." college & research libraries 77, no. 2 (2016): 164-183. titus, megan l., jenny l. scudder, josephine r. boyle, and alison sudol. "dialoging a successful pedagogy for embedded tutors." praxis: a writing center journal (2014). troisi, jordan d. "student management teams increase college students’ feelings of autonomy in the classroom." college teaching 63, no. 2 (2015): 83-89. van lier, leo. "agency in the classroom." sociocultural theory and the teaching of second languages 163 (2008): 186. webster, kelly, and jake hansen. "vast potential, uneven results: unraveling the factors that influence course-embedded tutoring success." praxis: a writing center journal (2014). aresty rutgers undergraduate research journal, volume i, issue ii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. sliding mode-based traction control for an autonomous martian rover julian bowne, dr. annalisa scacchioli (faculty advisor) ✵ abstract a traction control system was developed for an autonomous martian rover using a sliding mode controller. the main inspiration for this project was nasa’s mars rover, curiosity, which suffered severe wheel damage due to the lack of an effective traction control system. a control system was sought out to effectively prevent wheel damage, slippage, and soil failure for a martian rover. it was initially hypothesized that a sliding mode controller would be most effective to control the vehicle’s traction. a simulink model was created with a deformable soil-rigid tire mathematical model in order to simulate the traction control system. the sliding mode controller was tested to be more robust and stable compared to a proportional-integral-derivative (pid) controller for the rover. the results elaborate the possible applications for this project, which spans across commercial and military rovers, rescue robots, and planetary rovers in the private and global space industry. 1 introduction traction control has been a growing concern for the robotics and automotive industry. for example, traction control can be applied to hybrid cars, rescue robots, and planetary rovers [2,8,12]. the main inspiration for this research was the severe damage of the wheels of the mars rover, curiosity, that resulted from an incomplete understanding of the martian terrain and a lack of a suitable traction control system[1]. nasa later produced an algorithm to reduce further damage to the wheels and to continue its exploration of mars, but this method was not in the scope of difficulty for this project[14]. instead, a similar approach to kazuya yoshida's article about slip-based traction control was followed[17]. the goal of this research is to create a simulation to test a model-based traction control system for a martian rover, like curiosity. this project consists of two main components: mathematical modeling of the vehicle-terrain interaction and the controller, which generates a voltage applied to the wheel’s electric motor. this voltage forces the slip ratio to approach a desired value. in other words, the controller accelerates the vehicle efficiently in the deformable martian terrain. in yoshida's research, a slip-based model was developed using the terramechanics, which is the study of soil properties, of a rigid wheel on loosedeformable soil[17]. the terrain model used in this article was the deformable soil-rigid tire model, which was then adopted into the current project[15]. this figure 1: one of curiosity’s damaged wheels. description: the holes present on the middle tire, on the right side of the image, was caused by sharp rocks found in gale crater. due to the lack of a traction control system, the damage to the tire grew rapidly. aresty rutgers undergraduate research journal, volume i, issue ii model consists of several equations of terramechanics; specifically the interaction of wheeled vehicles on various surfaces. the equations involved model soil that is deformed by the vehicle’s wheels, as well as the resistive and driving force and torque generated by the soil. these relations allow for a more accurate simulation for the soil and vehicle interaction. this model was also chosen based on the computing restrictions of the computer used. more advanced options like finite element models require computers with higher level of performance. another key component is the controller. in yoshida's research, a proportional-integral-derivative controller (pid) was used to control the slippage[17]. after thorough investigation, a faster and more robust method was sought out. a faster method specifically meant a quicker settling time (time to reach 2% of the desired slip ratio value), and more robust method meant a controller that functions better with more uncertainties. the sliding mode controller was used for its robustness and effectiveness[5]. the more advanced controller is a nonlinear control method that uses a discontinuous control signal. it is a type of variable structure control, meaning it alters the dynamics of a nonlinear system by applying a switching control. mathworks’ softwares matlab and simulink were integral tools used for this research[9]. matlab is a programming tool with its own language that expresses matrix math more directly. simulink is a program for modeling, simulating, and analyzing dynamical systems. its main interface is a block diagramming tool that allows the user to model and simulate a system made up of various interconnected equations. they were ideal for this project because simulink and matlab are frequently used for professional vehicle based-control system design in the field. simulink is user-friendly for mathematical modeling because of its block-based coding. this means that instead of writing lines of code, the control system can be designed with simple “blocks” that represent mathematical operations and tasks. matlab was useful for inputting initial conditions and extracting results for plots. both of these softwares were used together to simulate and extract results. 2 modeling a negative feedback system occurs when the output of a system is fed back into itself in order to influence its output. these system configurations are required for most control methods. in order to model a traction control system with block diagrams in simulink, three basic components of a negative feedback loop are needed: 1. the plant is the mathematical model of the dynamic equations of the problem. 2. the actuator is the controlled motor. 3. the controller is inputted with the error of the feedback loop (the desired slip ratio minus the plant’s modeled slip ratio) and outputs the voltage applied to the actuator. to begin understanding the problem of slip-based traction control on deformable soil, a basic understanding of a vehicle control system was established through the exploration of a simple anti-lock braking system (abs). two examples were used as a starting point: the first being an example from mathworks' website and the second being a rudimentary simulink block diagram[4,10]. the second example incorporates the quarter car model for the vehicle's motion equations. equation set 1: vehicle model and slip ratio. aresty rutgers undergraduate research journal, volume i, issue ii the first set of equations modeled were the vehicle's motion equations. the quarter car model was used, which isolates each individual wheel to carry the load of a quarter of the car's weight. this allows the problem to be simplified to controlling a single wheel’s slip. the derived equations are newton's second law for a quarter mass vehicle and newton's second law for circular motion for the wheel. the input variables in the quarter car equations are the driving force, resistive force, motor torque, and soil torque (reaction torque generated by the soil). the acceleration taken by the two equations are integrated to calculate the velocity of the vehicle and the angular velocity of the wheel. these values along with the wheel’s radius is used to calculate the slip ratio. in order to develop a relation for the slip ratio and the forces and torques generated by the soil onto the wheel, a soil-vehicle mathematical model must be utilized. a simple model that was initially used for testing the controller was the pacejka’s magic tire formula[11]. it is a trigonometric function (made up of sine and arctangent) with generally four terrain specific coefficients. the more complex model that was used for the final results was the deformable soil-rigid tire model. it is composed of three integrals to calculate the two forces and one torque. the integrated functions are the components of shear and normal stress produced by the soil. the martian soil is represented by certain soil coefficients in the equations to calculate the shear and normal stress as functions of slip. figure 2: vehicle model. description: this diagram represents a simplified model, which focuses on a single wheel’s dynamics. 𝑀𝑀 is the mass of the car, 𝑟𝑟 is the radius of the wheel, 𝑣𝑣𝑥𝑥 is the vehicle’s velocity, 𝜔𝜔 is the angular velocity of the wheel, 𝐹𝐹𝑑𝑑 is the driving force, and 𝑇𝑇𝑚𝑚 is the motor torque. table 1: vehicle model and slip ratio variables. aresty rutgers undergraduate research journal, volume i, issue ii figure 3: deformable soil-rigid tire model. description: the diagram represents the tire interaction with de-formable soil. the main focus of the diagram is the shear and normal stresses represented by 𝜏𝜏 and 𝜎𝜎, respectively. the stress is experienced by the tire, and two resultant forces are generated, the driving force 𝐹𝐹𝑑𝑑 and the resistive force 𝐹𝐹𝑅𝑅. equation set 2: terramechanics equations. table 2: terramechanics variables. aresty rutgers undergraduate research journal, volume i, issue ii parameters used in this project for the equations mentioned earlier came from two sources. these include yoshida’s article for soil properties and nasa’s curiosity specifications for the rover’s dimensions and mass[13,17]. below is a table of the parameters: in order to utilize the deformable soil-rigid tire model, the integrals to calculate the forces and torque must be solved. several methods were considered for finding the solutions. first, an analytical method was investigated. however, using matlab's symbolic toolbox (a solver that can be used for integrals) only resulted in a non-simplified output. since the equations only can be solved numerically, taylor series approximations were used. adding higher order derivatives to the approximations makes it more accurate, which is shown in the graph below. the 5th order approximations were utilized in simulink to complete the mathematical model. this approximation was sufficient since higher slip ratio values (closer to one) do not occur in the simulation. the slip ratio was determined to be the state variable, which means it is a variable that describes the state of the system. the friction or driving force, which is maximized for the ideal performance of the simulated vehicle, is a function of this slip ratio. the voltage is the control variable which is then inputted into a dc motor to generate a corresponding motor torque. a model of a dc motor was required for the actuator of the simulation. after investigation, the armature circuit was a good starting point to model an ideal motor[3]. the simple armature circuit consists of a voltage source, resistor, inductor, and rotor with a fixed magnetic field. for the dc motor, newton’s second law and kirchhoff’s voltage law were derived. also, proportional relations were determined for the torque and electromotive force. by utilizing these equations, an input of voltage produces a corresponding motor torque. table 3: parameters for the overall system. aresty rutgers undergraduate research journal, volume i, issue ii list of relevant terms negative feedback loop – a system where the output is fed back into itself in order to influence its future output. plant – the mathematical model that has an input and output (without a feedback loop) actuator – the component of a system which is responsible for moving or controlling a device or machine. controller – the component of a system that uses the system’s state to control its output. quarter car model – a model that simplifies a vehicle into an isolated wheel and a quarter of the total mass. newton’s second law – the acceleration is directly proportional to the magnitude of the net force, and inversely proportional to the mass. slip ratio – the value expressed in a ratio that represents the slipping behavior of a vehicle’s wheel. driving force – the force, also called the friction force, is the force that causes the vehicle to accelerate forward. pacejka’s magic tire formula – a simple formula that uses the slip ratio as an input and outputs the driving force. different plugins of coefficients can model different terrains. deformable soil-rigid tire model – a set of equations that model a complex deformable soil like the martian terrain. taylor series approximation – a way of approximating complex functions, using values of its derivative. higher order expansions (more derivatives) allow for a more accurate approximation. shear stress – the force applied to an object in the direction parallel to the cross-sectional area. normal stress – the force applied to an object in the direction normal to the cross-sectional area. armature circuit – an electrical circuit that involves a rotating part or in the case of a dc motor, an input current that interacts with a fixed magnetic field to produce a torque. kirchhoff's voltage law – a form of conservation of energy for a closed electrical circuit. the law states that the sum of the voltage around a circuit must equal zero. figure 4: taylor series approximation for driving force. description: this graph shows the increase in the accuracy of the taylor series approximation as higher order derivatives are added. the approximations are compared to the numerical curve, which is colored blue. figure 5: dc motor armature circuit. description: this diagram shows the electrical circuit on the left, and the rotor and electric field on the right. the circuit contains a voltage source, a resistor, an inductor, and a rotor. aresty rutgers undergraduate research journal, volume i, issue ii 3 the controller the purpose of the controller is to output a motor voltage signal to force the slip ratio to approach the desired value. the pid controller was first experimented with and later used to compare final results. it consists of a proportional gain, an integral gain, and a derivative gain. these multiplicative coefficients allow the user to tune the controller for different dynamical systems. the sliding mode controller is a more advanced method than the pid. the controller, essentially, has a phase where the state variables are forced to approach a function called the sliding surface. once on the surface, the sliding phase begins by trying to bring this sliding function to zero (as well as the error of the desired and modeled slip ratios). figure 7: simulink model. description: the three main components (left to right) are the controller, labeled smc for sliding mode controller, the actuator, which is the dc motor, and the plant, which is the deformable soil-rigid tire model. the smc is inputted with three tuning coefficients labeled k, lambda, and psi as well as the error and its rate of change. it then outputs a voltage applied to the actuator. the actuator then outputs the motor torque, which is inputted into the plant. based on the vehicle-terrain model, the plant outputs the slip ratio which is fed back into the system to calculate the error. this completes the negative feedback loop that simulates a traction control system for a martian rover. figure 6: sliding mode behavior. description: first, the sliding mode controller forces the system onto the sliding surface, denoted by the reaching phase. then, it oscillates about this surface until the error is zero, which is called the sliding phase. aresty rutgers undergraduate research journal, volume i, issue ii the mathematics behind the sliding mode is divided into the sliding surface design and the control input design. for the sliding surface design, the function is made up of the error and a certain number of its derivatives. for the control input design, the task is to steer the sliding surface function towards zero (in other words, to reduce the error). the control input for this project is the voltage applied to the electric motor powering the wheel. in the context of the problem, figure 6 variables (x1 and x2) are the wheel’s angular velocity and the vehicle’s velocity, respectively. these variables are the state variables of the system because they are used to calculate the slip ratio. similar to the pid controller, the sliding mode controller has coefficients for tuning. these coefficients are involved in the equations for the sliding surface function and the control input function. after modeling the necessary mechanisms for controlling slip, the models must be integrated into simulink for creating the simulations. to create the negative feedback system, different blocks for mathematical computation are incorporated to create a block diagram version of the equations for the model and controller. once all the coefficients and initial conditions have been inputted, the simulation can commence. list of relevant terms pid controller – proportional-integral-derivative controller. it uses a balance of the three to minimize the error of the control system. sliding mode controller – a nonlinear control method that alters the dynamics of a system with a discontinuous control signal that forces the system to “slide” along a function that minimizes the error. sliding surface – the bounded function in which the sliding mode controller forces the system onto. control input – the manipulated discontinuous signal that forces the system onto the sliding surface. tuning – the way a control method is customized for the appropriate scenario. state variables – variables that describe the mathematical state of the dynamical system. 4 results the results for the preliminary pacejka model simulation showed the working pid controller and different traction control responses for each of the terrains: dry asphalt, wet asphalt, concrete, snow, and ice. the desired slip ratio was different for the terrains. the value was determined by maximizing the driving force using the pacejka’s magic tire formula. when looking at the plots, it is evident that the pid controller overshoots the desired slip ratio. additionally, the oscillations in the response are not minimized. when comparing the different responses, the pid’s response for ice had more oscillations at a higher frequency. this shows the lack of robustness for the pid controller. the rise and settling time for the plots were nearly identical for the five terrains. these preliminary results were useful to see the behavior of the pid for the simulated wheel and vehicle. in addition, these results already show the limitations of the pid controller for a simpler terrain model represented by the overshoot and oscillations. the main results of this project illustrate how a sliding mode-based traction control system works on a simplified martian terrain using a deformable soil-rigid tire model. the sliding mode controller outperformed the pid controller in stabilizing the slip ratio to a value of 0.16. looking at the step responses, the sliding mode controller did not overshoot the desired value, while the pid did. in addition, the response of the sliding mode controller was critically damped, which was ideal. the initial response for the sliding mode controller was also faster, which is desirable for a traction control system. this is evident in the faster settling time, which was reduced by around a factor of 7 (decreased from 0.7 to 0.1 seconds). additionally, the rise time was reduced by around a factor of 3. looking at the tracking error plots, the error for the sliding mode controller response shows a single peak with a very small width. the same peak is evident in the pid response, but it is wider. this means the width of the initial error pulse is significantly reduced. this further supports that the sliding mode controller has better performance for minimizing the error. to finalize, the sliding mode controller is more practical for a physical rover because of its faster and more efficient response. aresty rutgers undergraduate research journal, volume i, issue ii figure 8: preliminary pid with pacejka’s magic tire formula for dry asphalt. description: the plot shows the slip ratio being controlled to achieve the desired value for dry asphalt. figure 9: preliminary pid with pacejka’s magic tire formula for wet asphalt. description: the plot shows the slip ratio being controlled to achieve the desired value for wet asphalt. figure 10: preliminary pid with pacejka’s magic tire formula for dry concrete. description: the plot shows the slip ratio being controlled to achieve the desired value for dry concrete. figure 11: preliminary pid with pacejka’s magic tire formula for ice. description: the plot shows the slip ratio being controlled to achieve the desired value for ice. figure 12: preliminary pid with pacejka’s magic tire formula for snow. description: the plot shows the slip ratio being controlled to achieve the desired value for snow. aresty rutgers undergraduate research journal, volume i, issue ii figure 13: step response using pid. description: the slip ratio was successfully controlled using the pid controller. however, it overshot the desired slip ratio, and it created many oscillations. figure 14: tracking error using pid. description: the error plot shows the delay in the pid controller’s response and the oscillations present. figure 15: step response using sliding mode. description: the plot shows the step response of the sliding mode controller. the controller had a critically damped response with a faster rise and settling time compared to the pid. figure 16: tracking error using sliding mode. description: the plot shows the tracking error of the sliding mode controller. the error decreases rapidly and does not oscillate like the pid’s tracking error. aresty rutgers undergraduate research journal, volume i, issue ii 5 conclusion slip-based traction control is a very useful tool for autonomous rovers. these techniques can be applied for a wide-range of purposes: planetary exploration, rescue missions, and commercial and military endeavors. the simulations were the beginnings of creating an autonomous rover with a traction control system. the results showed the validity of the mathematical modeling and the controller design, which opened a doorway for furthering the complexity of this research. to summarize, the negative feedback-traction control system was broken down into three components: the plant, the actuator, and the controller. the plant was the mathematical model of the vehicle and terrain interaction (mostly the deformable soil-rigid tire model). the actuator was the hypothetical dc motor that powered the wheels (modeled according to a simple armature circuit). the (final) controller was the sliding mode, which worked more efficiently than the pid. all of these combined with the correct initial conditions simulated a slip-based traction control system for a rover on mars. in conclusion, the method utilized led to a better understanding of the physics of deformable soil and the limitations of a model-based traction control system. in addition, the results confirmed the hypothesis that a sliding mode controller would be more effective than a pid controller. this project further confirms the use of a sliding mode controller for a traction control system, which has been tested in recent research[6]. this project can be a trailblazer for sliding mode controller research specifically for a martian rover. the computational results will be used on a physical rover in order to make mars exploration easier and more efficient. for future works, the traction control system will be tested on a physical rover (prototype being assembled in the picture above) and on a deformable test bed. in addition, theoretical work about using a sliding mode controller integrated with data-driven modeling and control will be researched in order to advance this technique. these results will be compared to that of the taylor series approximation∎ 6 acknowledgements i would like to thank my research advisor dr. annalisa scacchioli for her phenomenal guidance. she has always been so supportive of me, and without her this project would never have come to fruition. also, i would like to thank the aresty research program for giving me this amazing opportunity. finally, i would also like to thank arjun singh, vishal rachapudi, and galileo wang for their impeccable assistance, which led to the creation of the prototype rover. 7 references [1] arvidson, r., degrosse, p., grotzinger, j., heverly, m., shechet, j., moreland, s., newby, m., stein, n., steffy, a., zhou, f., zastrow, a., vasavada, a., fraeman, a., & stilly, e. (2017). relating geologic units and mobility system kinematics contributing to curiosity wheel damage at gale crater, mars. journal of terramechanics, 73, 73–93. https://doi.org/10.1016/j.jterra.2017.03.001 [2] gonzalez, r., apostolopoulos, d. & iagnemma, k. improving rover mobility through traction control: simulating rovers on the moon. auton robot 43, 1977–1988 (2019). https://doi.org/10.1007/s10514-019-09846-3 figure 17: rover prototype. description: the prototype is a six-wheeled, rocker-bogie suspension rover. the frame and wheels were 3d printed. the computer onboard the rover is an arduino uno. the green cylinders are lipo batteries to power the rover. the motor and electronic speed controllers have not been added yet. https://doi.org/10.1016/j.jterra.2017.03.001 https://doi.org/10.1007/s10514-019-09846-3 aresty rutgers undergraduate research journal, volume i, issue ii [3] dc motor speed: simulink modeling. control tutorials for matlab and simulink motor speed: simulink modeling. http://ctms.engin.umich.edu/ctms/index.php? example=motorspeed§ion=simulinkmodeling#1 [4] emheisen, a., ahmed, a., alhusein, n., sakeb, a., & abdulhamid, a. (2015). car wheel slip modelling, simulation, and control using quarter car model, international journal of engineering trends and technology (ijett), vol. 28, no. 6, 291–293. [5] example on sliding mode control. example on sliding mode control file exchange matlab central. https://www.mathworks.com/matlabcentral/ fileexchange/52429-example-on-sliding-mode-control [6] han, k., choi, m., lee, b., & choi, s. b. (2018). development of a traction control system using a special type of sliding mode controller for hybrid 4wd vehicles. ieee transactions on vehicular technology, vol. 67, no. 1, pp. 264-274 [7] iagnemma, k., & dubowsky, s. (2004). mobile robots in rough terrain. berlin heidelberg:springer-verlag. [8] li, s., liao, c., chen, s., & wang, l. traction control of hybrid electric vehicle. 2009 ieee vehicle power and propulsion conference, dearborn, mi, 2009, pp. 1535-1540, doi: 10.1109/vppc.2009.5289563. [9] matlab. (2019). version 9.7.0 (r2019b). natick, massachusetts: the mathworks inc. [10] modeling an anti-lock braking system. matlab & simulink. https://www.mathworks.com/help/simulink/slref/modeling-an-antilock-braking-system.html [11] pacejka, h. b. (2002). tire and vehicle dynamics. society of automotive engineers. [12] reina, g., giannoccaro, n.i., & messina, a. (2013). traction control for four-wheel drive robots. [13] summary. nasa, 9 sept. 2019, https://mars.nasa.gov/msl/spacecraft/rover/summary/ [14] toupet, o., biesiadecki, j., rankin, a., steffy, a., meirion-griffith, g., levine, d., schadegg, m., & maimone, m. (2018). traction control design and integration onboard the mars science laboratory curiosity rover. 2018 ieee aerospace conference. [15] wong, j. y. (1978). chapter 2. in theory of ground vehicles. john wiley & sons. [16] yoshida, k., hamano, h. (2002) motion dynamics of a rover with slip-based traction model, proc. 2002 ieee int. conf. on robotics and automation, 3155-3160. [17] yoshida, k., hamano, h., & watanabe, t. (2003). slip-based traction control of a planetary rover. b. siciliano and p. dario (eds.): experimental robotics viii, star 5, pp. 644–653. http://ctms.engin.umich.edu/ctms/index.php?example=motorspeed%c2%a7ion=simulinkmodeling#1 http://ctms.engin.umich.edu/ctms/index.php?example=motorspeed%c2%a7ion=simulinkmodeling#1 https://www.mathworks.com/matlabcentral/fileexchange/52429-example-on-sliding-mode-control https://www.mathworks.com/matlabcentral/fileexchange/52429-example-on-sliding-mode-control https://www.mathworks.com/help/simulink/slref/modeling-an-anti-lock-braking-system.html https://www.mathworks.com/help/simulink/slref/modeling-an-anti-lock-braking-system.html https://mars.nasa.gov/msl/spacecraft/rover/summary/ aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. is it all in our heads? an investigation into american and historical legacies of racism and social frameworks that perpetuate racial inequalities in twenty-first century healthcare systems meghana nampally ✵ abstract racial bias in the healthcare system, originating from the eminent founders of science and medicine, has numerous adverse effects on black populations and continues to have harmful consequences today (byrd and clayton, 2001). from poor clinical decision-making to preventing people of color from entering prominent fields in medicine, racism is ubiquitous in medicine and healthcare (byrd and clayton, 2001). the impact of racial bias on patient care is of great interest with many studies illustrating the detrimental impacts of bias on minority groups, specifically in black communities. however, there is additional research that concludes racial bias does not play a role in patient care or in medicine (dehon et al., 2017). the lack of acknowledgment within academia concerning racial disparities in healthcare and science further oppresses black voices. with my research, i investigate the extent to which various biased social frameworks in healthcare, medicine, and science negatively impact black individuals. i also address major historical events during the creation of the modern-day healthcare system and how these events perpetuate racism today. focusing on the twenty-first century, i demonstrate that systemic historical and social events during this period eternalize racism in the modern-day american healthcare system. key terms: paternalistic racism, competitive racism, race-conscious professionalism, psychologizing racism 1 introduction confronting racial inequalities in healthcare requires examining the historical legacies of academia and of federal policies. neil lewis jr., an assistant professor of communications research in medicine at weill cornell medicine, describes “that you can’t understand, much less change, people’s health behaviors without reckoning with larger social structures and systemic forces” (blackwood, 2021). beginning in the nineteenth century, residential segregation excluded people of color from obtaining adequate healthcare. this enabled and sustained “structural racism in other forms, including… the unjust distribution of high-quality health care” (bailey et al., 2021). exacerbating the effects of residential segregation, academic discrimination against colored communities preserves outdated racist attitudes and ideologies (bailey et al., 2021) while also perpetuating distrust between marginalized communities and healthcare services (dula, 1994). another facet of this bias in academia can be seen in the underrepresentation of minorities in science and healthcare. in dr. michael byrd and dr. linda clayton’s paper “race, medicine, and healthcare in the united states: a historical survey,” they delineate two different forms of racism: paternalistic racism and competitive racism (byrd and clayton, 2001). the former is the view that black people are immature, child-like, and inferior — dispositions that are tolerated unless they deviate from socially acceptable roles (byrd and clayton, 2001). in the latter, aresty rutgers undergraduate research journal, volume i, issue iv emancipated slaves were seen as competition for scarce resources (byrd and clayton, 2001). these act as the foundations for the modern-day concept of race-conscious professionalism, wherein black individuals in prestigious medical and academic fields feel a dual obligation to succeed (powers et al., 2016). the idea of psychologizing racism, as described in “a socioecological psychology of racism: making structures and history more visible,” is the concept where one overly focuses on individual bias while neglecting the systemic and historical racism behind various institutions (trawalter et al., 2020). this recognition of excessively analyzing the role of individual accountability rather than moving to understand global structural forces at work is significant as it elucidates the true depth of racism in perpetuating racial disparities. in this paper, i consider three different historical and social domains that perpetuate racial inequalities in the twenty-first century. the first is the role of academic discrimination in fueling fear and distrust in the healthcare system. here, i also examine the consequences of this bias through the under-treatment of black patients, and through the disproportionate number of physicians treating people of color. i then examine the significance of redlining in racial segregation. redlining is the use of racial demographics to assess which communities would receive investment (bailey et al., 2021). the ramifications of this practice manifest as areas with increased pollutants, which worsen healthcare outcomes in colored communities (li and yuan, 2021) and demonstrate pollution inequity (abraham et al., 2021). finally, the lack of representation in academic and healthcare settings is analyzed using the framework of race-conscious professionalism. academic discrimination, redlining, and underrepresentation in healthcare and academic institutions are interrelated realms that perpetuate and eternalize racial inequalities in the modern-day healthcare system because of their historical and social contexts in the united states. 2 academic discrimination racial inequalities exacerbate distrust between scholars and people of color. a prominent example of this is the tuskegee syphilis study. during this study, black men were infected with syphilis without their consent (trawalter et al., 2020). after this information was disclosed to study participants, decreased healthcare usage was seen among black men, which led to increased mortality rates (trawalter et al., 2020). trawalter et al. illustrate that “the disclosure of the study in 1972 is associated with a decrease in healthcare utilization, presumably due to decreased trust in the medical community, and a commensurate increase in mortality among [black] men” (trawalter et al., 2020). this event is of historical significance when discussing the perpetuation of racial discrepancies in healthcare as there is a history of treating black people as inferior in the academic community. it is important to note that the distrust demonstrated by the tuskegee syphilis study is not an isolated event in academic history. james marion sims, the “father of modern gynecology,” rose to prominence for creating a surgical remedy for the obstetric fistula based on his experimentation on black, enslaved women (cronin, 2020). during his experiments, he failed to use anesthetic ether on the individuals he experimented on, despite having access to such resources (cronin, 2020). however, sims did use anesthesia on his wealthy, white patients (khabele et al., 2021). the use of sims’ work in modern medicine represents the exploitation of slaves and suggests a potential cause for the mistrust black individuals have toward academia (conteh et al., 2022). the results of this distrust between medical professionals and marginalized communities can be seen in the racial proportion of physicians to underprivileged, colored communities. dr. miriam komaromy and her colleagues found that in areas with five times as many black residents, the number of black physicians was commensurate with the number of residents (komaromy et al., 1996). this disproportion in race illustrates that colored communities mainly trust physicians of their own race and ethnicity, as well as black physicians feeling the need to practice in communities with large populations of their own race and ethnicities. this is reinforced by “the fact that the physician's race or ethnic aresty rutgers undergraduate research journal, volume i, issue iv group predicted whether he or she would care for greater-than-average numbers of black or hispanic patients” (komaromy et al., 1996). this further supports the idea that racial disparities are perpetuated by mistrust in healthcare settings. race was a contributing factor to where individuals would practice; therefore, individuals of the same race as their primary care provider feel more comfortable receiving aid from members of their own race. academic discrimination against black individuals additionally stems from promoting biased teachings in medical schools, specifically in the perception of pain. in a 2016 study to determine racial attitudes, medical students held beliefs that black patients feel less severe pain than that of white patients (bailey et al., 2021). this bias leads to overt disparities in treatment regimens as “[black patients] are less likely than white [patients] to receive pain medication and, when they do, they receive less” (trawalter et al., 2012). the disparity in treatment between white and black patients by physicians and students contributes to the distrust that black patients feel as these false beliefs perpetuate the care of “greater-than-average numbers of black or hispanic patients'' by physicians of the same race (komaromy et al., 1996). this undertreatment of black patients because of academic bias intensifies the skepticism and suspicion of medical practices, ultimately leading to disparities in healthcare as seen through the disproportionate number of black physicians practicing in communities with large populations of similar ethnicities. 3 redlining in residential segregation redlining makes use of racial compositions to assess investment opportunities for communities (bailey et al., 2021). it involves the conscious discrimination against black people from obtaining financial resources that would aid them in acquiring adequate housing as well as other necessities (bailey et al., 2021). redlining impacts the proximity to which individuals receive adequate education, nutrition, recreation, and medical care services as “neighborhoods influence the collective resources” these individuals receive (li and yuan, 2021). segregated neighborhoods face greater barriers to recruiting and retaining physicians, which limits individual access to healthcare services (white et al., 2012). these neighborhoods have limited resources (e.g. diagnostic imaging services) which also contribute to healthcare disparities (white et al., 2012). due to disparities in resources, “women whose residential neighborhood[s are] characterized by a lower quality-built environment are also at increased risk of adverse perinatal outcomes such as preterm birth” (anthopolos et al., 2014). further augmenting the impact of redlining, historically disadvantaged neighborhoods have “a higher risk for covid-19 infection in zctas with present-day economic and racial privilege” (li and yuan, 2021). here, li and yuan (2021) illustrate a greater risk for covid-19 infection in present-day redlined areas. this mirrors the impact of poor-quality environments due to previous residential segregation, as this leads to an increased risk of “adverse perinatal outcomes” (anthopolos et al., 2014). therefore, modern-day redlined zones illustrate the perpetuation of racial inequalities in healthcare outcomes through limiting the number of resources marginalized communities can gain access to due to the progression of “racially segregated communities [becoming] economically segregated, resulting in the large-scale disinvestment often characterizing majority non-white neighborhoods” (anthopolos et al., 2014). this ultimately leads to overt consequences such as “preterm birth through poor-quality built environment [and] poor-quality housing stock” (anthopolos et al., 2014). clearly, anthopolos, li, and yuan illustrate that standard of care is impacted by geographic factors. redlining has influenced racial inequalities in healthcare outcomes through increased exposure to pollutants and other toxins, which further illustrates the lack of investment in black communities. this disparity is demonstrated as “better holc neighborhood grades are associated with lower levels of airborne carcinogens and higher levels of tree-canopy coverage (which mitigates air polluaresty rutgers undergraduate research journal, volume i, issue iv tants and heat)” (bailey et al., 2021), whereas predominantly black communities face “pollutant exposure through proximity to neighboring industrial plants or landfills, water leakage, mold, lead paint, pest infestation, and poor ventilation” (abraham et al., 2021). this demonstrates pollution inequity as white individuals create most of the fine-particulate pollution due to their overconsumption of goods; however, black and latinx minorities face the consequences and inhale this pollution (abraham et al., 2021). li and yuan further illustrate pollution inequity as the government and society’s “devoid of investment” in black communities leading to “the institutionalized segregation of capital (e.g., loans and investments) from black people [which] shaped the socio-spatial arrangement of goods and services [e.g. medical care] in the usa” (li and yuan, 2021). the fact that white individuals do not face the same consequences as their black counterparts demonstrates the disadvantage and inequality that these marginalized communities endure, ultimately leading to more adverse health outcomes such as asthma and low birth rates (li and yuan, 2021). li, yuan, and bailey et al. illustrate that the racial inequalities between black and white individuals because of racial segregation keep people of color in disadvantaged and disinvested neighborhoods. these past policies continue to perpetuate racial disparities in healthcare since white individuals do not experience this increased risk of illness. clearly, redlining and racial segregation promote racial discrepancies in healthcare. 4 underrepresentation in healthcare and academia the absence of minority groups in healthcare positions and academic settings illustrates the depth of racial inequalities in society. this lack of representation can be seen in the recruitment and retention of black faculty, as a study conducted in 2010 demonstrated that “among faculty members who had been hired in 2000, blacks were less likely to have been retained than any other demographic group” (ansell and mcdonald, 2015). this exclusion of black professors and faculty from academia is also perpetuated through “poor education and school quality; lack of role models; financial cost of education and training; and persistent bias, stereotyping, and racism” (powers et al., 2016). the paucity of black representation in academic and healthcare settings leads to the existence of racial inequalities in healthcare in the form of extrinsic factors such as “persistent bias, stereotyping and racism” (powers et al., 2016). these factors lead to “only 2.9% of all faculty members at u.s. medical schools [being] black” (ansell and mcdonald, 2015). this contributes to the disproportionate ratio of black to white physicians by creating environments in which black medical students lack black role models, resulting in fewer people of color in these fields. the effects of underrepresentation in academia and healthcare can be illustrated through the idea of “race-conscious professionalism” where african americans understand the implications of their professional success in race politics and marginalized communities (powers et al., 2016). this leads to black physicians experiencing a dual obligation to reach professional excellence in order to protect their communities. the two-fold responsibilities that black physicians experience are not limited to the present day; many of the first formally trained physicians used their scientific credibility and community leadership to build hospitals to care for black communities, while also bolstering the african-american professional class (powers et al., 2016). race-conscious professionalism gives insight into the intrinsic obstacles that black physicians face as “most have experienced or witnessed, firsthand, inequalities in the access to, and quality of, health care” (powers et al., 2016). since many black physicians have experienced the disparities that their communities are facing, they feel an obligation to aid their communities, which leads to a disproportionate number of “black and hispanic physicians locat[ing] their practices [to] areas with higher proportions of residents from underserved minority groups (komaromy et al., 1996). the lack of black representation in academia and healthcare perpetuates racial inequality as it leads to increased pressures placed on minority physicians to practice in marginalized communities aresty rutgers undergraduate research journal, volume i, issue iv who have also “witnessed inequalities in the access to, and quality of, health care” (powers et al., 2016), resulting in “black and hispanic physicians consistently car[ing] for disproportionately high numbers of [black and hispanic] patients” (komaromy et al., 1996). this burden is a direct result of the lack of inclusivity in science and medicine for black individuals, which leads to incommensurate physician demographics. confining colored physicians to practice in underprivileged areas to protect and advocate for their own racial and ethnic groups preserves racial inequality as this responsibility is placed solely on colored minorities while their white counterparts are liberated from this accountability. 5 differentiating systemic bias from individualistic bias while i have argued that racial inequalities in healthcare must be viewed from a systemic lens through observing past historical and social abuses, there are also those who argue that an individualistic lens is more suitable. sabin et al. illustrate prioritizing the individualistic perspective as “physicians [holding] strong implicit associations for black patients as being ‘less cooperative’ and demonstrating that this implicit bias was related to quality of care” (sabin et al., 2009). they support viewing racial inequalities as an individual’s responsibility because this has direct consequences in clinical decisionmaking. this can include treatment plans for patients, as with the use of thrombolysis for coronary symptoms. for example, in green et al’s study to measure implicit bias in physicians about race, physicians that favored white patients more than black patients were more likely to treat their white patients with thrombolysis for coronary symptoms (green et al., 2007, as cited in sabin et al., 2009). therefore, physicians who maintain strong implicit biases contribute to inappropriate, adverse courses of treatment toward black patients, resulting in disparities that can explicitly be seen in the fact that “relative to [white] americans, [black] americans experience higher rates of diseases, disability, and premature death” (trawalter., et al 2012). although examining the impact of physician implicit bias on clinical decisions is important in evaluating racial disparities, solely relying on individual accountability negates the impact of greater systemic and structural forces. the significance of evaluating systemic elements when discussing racial inequalities can be seen in “the systematic disinvestment… within segregated black neighborhoods [which] has resulted in under-resourced facilities with fewer clinicians, which makes it more difficult to recruit experienced and well-credentialed primary care providers and specialists and thereby affects access and utilization” (bailey et al., 2021). therefore, when examining the influence of systemic factors such as the disinvestment in black neighborhoods, it transcends the quality of care. because access to resources is limited in black communities, this leads to fewer physicians and difficulty in obtaining a higher standard of healthcare. this showcases the greater depth and effect behind the nature of care that black neighborhoods receive. viewing racial disparities in terms of barriers such as redlining and race-conscious professionalism is more effective in understanding racial inequalities because it provides reasoning for the pervasive disadvantages that black people face, independent of individual actions. regardless of physician bias, black patients “consistently have much higher rates of premature, preventable death and poorer health throughout their lives” (bassett, 2015). this can be attributed to greater forces that are deeply entrenched in institutions and policies rather than interpersonal interactions, as this enables a holistic framework for addressing the obstacles and adverse outcomes that are ubiquitous in black communities. in addition, global factors are significant because the “ongoing exclusion of and discrimination against people of african descent throughout their life course, along with the legacy of bad past policies, [continues] to shape patterns of disease distribution and mortality” (bassett, 2015). this further reinforces that the “higher rates of diseases, disability and premature death” (trawalter et al., 2012) in black demographics are not limited to implicit bias. this only provides a partial picture of the factors that perpetuate racial inequalities as aresty rutgers undergraduate research journal, volume i, issue iv these consequences continue to exist past individual interactions. 6 conclusion perpetuating racial inequalities are the result of direct and indirect social and historical factors pervasive in academia and healthcare systems. the academic prejudice against black communities promotes distrust in medical services, leading to the undertreatment of black patients and the inordinate number of black physicians practicing in these racialized neighborhoods. this academic prejudice is further supported by the underrepresentation of minorities in academia and the responsibility placed upon black scholars to excel in their fields. moreover, structural mechanisms that continue to preserve racial disparities surpass academia and science and can be observed in residential segregation, which leads to unequal access to healthcare services and negative health outcomes in black patients. these systemic forces depict the extensive nature of racial bias in medicine and society in the united states as these are not limited to individual interactions. the decisions made by individuals are based on historical and cultural bias. the ubiquity of the impact of structural racism is supported through the act of psychologizing racism, where the excessive analysis of one’s own prejudice, discrimination, and stereotypes towards people of color invalidates the experiences of black minorities (trawalter et al., 2020). this results in minimizing the true effect of institutional bias and injustice. it is important to validate the role of various systemic elements when examining the barriers placed upon black communities rather than focusing on individual biases. recognizing the social and historical contexts behind modern-day healthcare practices gives us insight into racial inequalities and disparities in academia and medicine. moving forward, reform requires the recognition and reconciliation with the past abuses against black individuals while also continuing to educate on historical and social policies. this entails medical schools educating on past historical injustices against black communities and the consequences of these abuses in modern-day medicine. additional initiatives to combat underrepresentation in academia should be proposed and enforced to dismantle systemic and structural racism most effectively in healthcare and beyond. this requires implementing programs that consider the social and historical contexts unique to the black experience. understanding the frameworks that perpetuate racial disparities today allows for a greater appreciation of black experiences as it elucidates long-standing obstacles that validate distrust in academia and medicine, prevent access to healthcare, and demonstrate underrepresentation in academia. enduring change can only stem from accepting that systemic racism is not rigid and absolute but can be remedied through actively educating and dismantling biased policies and institutions∎ 7 references [1] abraham, p., williams, e., bishay, a. e., farah, i., tamayo-murillo, d., & newton, i. g. (2021). the roots of structural racism in the united states and their manifestations during the covid-19 pandemic. academic radiology, 28(7), 893–902. https://doi.org/10.1016/j.acra.2021.03.025 [2] ansell, & mcdonald, e. k. (2015). bias, black lives, and academic medicine. the new england journal of medicine, 372(12), 1087–1089. https://doi.org/10.1056/nejmp1500832 [3] anthopolos, r., kaufman, j. s., messer, l. c., & miranda, m. l. (2014). racial residential segregation and preterm birth: built environment as a mediator. epidemiology (cambridge, mass.), 25(3), 397–405. https://doi.org/10.1097/ede.0000000000000079 [4] bailey, z. d., feldman, j. m., & bassett, m. t. (2021). how structural racism works — racist policies as a root cause of u.s. racial health inequities. the new england journal of medicine, 384(8), 768–773. https://doi.org/10.1056/nejmms2025396 [5] blackwood, k. b. (2021, april 5). faculty examine racism ‘embedded’ in us health care. cornell chronicle. retrieved from https://news.cornell.edu/stories/2021/04/faculty-examine-racism-embedded-us-health-care [6] bassett. (2015). blacklivesmatter — a challenge to the medical and public health communities. the new england journal of medicine, 372(12), 1085–1087. https://doi.org/10.1056/nejmp1500529 [7] byrd, & clayton, l. a. (2001). race, medicine, and health care in the united states: a historical survey. https://doi.org/10.1016/j.acra.2021.03.025 https://doi.org/10.1056/nejmp1500832 https://doi.org/10.1097/ede.0000000000000079 https://doi.org/10.1056/nejmms2025396 https://news.cornell.edu/stories/2021/04/faculty-examine-racism-embedded-us-health-care https://news.cornell.edu/stories/2021/04/faculty-examine-racism-embedded-us-health-care https://doi.org/10.1056/nejmp1500529 aresty rutgers undergraduate research journal, volume i, issue iv journal of the national medical association, 93(3), 11s–34s. [8] conteh, gagliardi, j., mcgahee, s., molina, r., clark, c. t., & clare, c. a. (2022). medical mistrust in perinatal mental health. harvard review of psychiatry, 30(4), 238–. https://doi.org/10.1097/hrp.0000000000000345 [9] cronin. (2020). anarcha, betsey, lucy, and the women whose names were not recorded: the legacy of j marion sims. anaesthesia and intensive care, 48(3_suppl), 6–13. https://doi.org/10.1177/0310057x20966606 [10] dehon, weiss, n., jones, j., faulconer, w., hinton, e., sterling, s., & choo, e. k. (2017). a systematic review of the impact of physician implicit racial bias on clinical decision making. academic emergency medicine, 24(8), 895–904. https://doi.org/10.1111/acem.13214 [11] dula, a. (1994). african american suspicion of the healthcare system is justified: what do we do about it? cambridge quarterly of healthcare ethics, 3(3), 347–357. https://doi.org/10.1017/s0963180100005168 [12] khabele, holcomb, k., connors, n. k., & bradley, l. (2021). a perspective on james marion sims, md, and antiblack racism in obstetrics and gynecology. journal of minimally invasive gynecology, 28(2), 153– 155. https://doi.org/10.1016/j.jmig.2020.10.027 [13] komaromy, grumbach, k., drake, m., vranizan, k., lurie, n., keane, d., & bindman, a. b. (1996). the role of black and hispanic physicians in providing health care for underserved populations. the new england journal of medicine, 334(20), 1305–1310. https://doi.org/10.1056/nejm199605163342006 [14] li, & yuan, f. (2021). historical redlining and resident exposure to covid-19: a study of new york city. race and social problems, 1–16. https://doi.org/10.1007/s12552-021-09338-z [15] powers, white, a. a., oriol, n. e., & jain, s. h. (2016). race-conscious professionalism and african american representation in academic medicine. academic medicine, 91(7), 913–915. https://doi.org/10.1097/acm.0000000000001074 [16] sabin, nosek, b., greenwald, a. g., & rivara, f. p. (2009). physicians’ implicit and explicit attitudes about race by md race, ethnicity, and gender. journal of health care for the poor and underserved, 20(3), 896–913. https://doi.org/10.1353/hpu.0.0185 [17] trawalter, s., bart-plange, d.-j., & hoffman, k. m. (2020). a socioecological psychology of racism: making structures and history more visible. current opinion in psychology, 32, 47–51. https://doi.org/10.1016/j.copsyc.2019.06.029 [18] trawalter, hoffman, k. m., & waytz, a. (2012). racial bias in perceptions of others’ pain. plos one, 7(11), e48546–e48546. https://doi.org/10.1371/journal.pone.0048546 [19] white, haas, j. s., & williams, d. r. (2012). elucidating the role of place in health care disparities: the example of racial/ethnic residential segregation. health services research, 47(3pt2), 1278–1299. https://doi.org/10.1111/j.1475-6773.2012.01410.x meghana nampally is a fourth-year undergraduate student at rutgers university-new brunswick. she is majoring in cell biology and neuroscience with a minor in psychology. she plans to pursue a ph.d./md program after rutgers and hopes to practice in underrepresented areas. meghana’s research was inspired by the racial disparities that are often overlooked in the medical field. as a future healthcare professional, she wanted to educate herself and others on the inequalities that people of color face in medicine and in academia. her research began as a final project for her english 201 class; however, she wanted to continue educating herself on these inequities, leading to her aresty undergraduate research journal submission. https://doi.org/10.1097/hrp.0000000000000345 https://doi.org/10.1177/0310057x20966606 https://doi.org/10.1111/acem.13214 https://doi.org/10.1017/s0963180100005168 https://doi.org/10.1016/j.jmig.2020.10.027 https://doi.org/10.1056/nejm199605163342006 https://doi.org/10.1007/s12552-021-09338-z https://doi.org/10.1097/acm.0000000000001074 https://doi.org/10.1353/hpu.0.0185 https://doi.org/10.1016/j.copsyc.2019.06.029 https://doi.org/10.1371/journal.pone.0048546 https://doi.org/10.1111/j.1475-6773.2012.01410.x microsoft word [formatted] 6 changing verbal label assignments selects the memory system for responses in an immediate visual recognition task aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. changing verbal label assignments selects the memory system for responses in an immediate visual recognition task alexa becker, mengxue kang, arnold glass (faculty advisor) ✵ abstract the dual system hypothesis posits the existence of two neural systems for memory and learning in the mammalian brain: the habit system and the improvisational system. this study sought to determine whether both systems are involved in a visual recognition task originally outlined in sternberg (1966) and whether each system could be selectively engaged on the basis of response assignment.[13] seventeen undergraduate students participated in an immediate visual recognition task where they responded whether or not a test consonant was present in a previous study sequence of one to six consonants by pressing one key for same or another key for different. when the different response was assigned to the spatially right “j” key, reaction time for targets and lures was a function of the study sequence size, indicating that the study sequence was serially scanned and compared with the test item by the habit system. however, when the same response was assigned to the spatially right “j” key, reaction time was not a function of study sequence size, indicating that the test item was not compared with the study sequence and responses were instead determined by perceived recency/novelty of the test item by the improvisational system. differences in reaction time depending on response assignment suggest the selection of one memory system over the other based on verbal labels assigned to response keys in different spatial locations. verbal label refers to the label of same or different assigned to the response keys in the experiment instructions. results expand upon sternberg (1966)—which used the same visual recognition task design as this study but did not account for response assignment, obscuring evidence of contributions from both memory systems—and provide more evidence for the dual-system hypothesis by demonstrating the involvement of both memory systems in immediate visual recognition.[13] 1 introduction sternberg’s (1966) seminal study on memory scanning found that, for a study sequence of one to six digits presented one at a time, visual recognition response time (rt) for a test digit was an increasing function of the length of the study sequence for both targets (a test digit that was present in the study sequence) and lures (a test digit that was not present in the study sequence).[13] these results suggested the existence of a visual recognition system that computes by (exhaustively) serially scanning a representation of a study set for a match with a test item. less than a decade after sternberg’s study, evidence was found for the existence of an additional visual recognition system that computes a test item’s perceived recency (referring to a test item that one recognizes as just having been seen in a previous study set) or novelty (referring to a new test item that one does not recognize as just having been seen in a previous study set) without scanning through a representation of a corresponding study set. whereas sternberg’s results suggested the serial scanning of a study set in memory for a match with a test item, atkinson & juola’s (1973 & 1974) suggested a purely perceptual judgment of the recency/novelty of a test item without comparison to a study set.[1,2] sternberg and atkinson & juola both aresty rutgers undergraduate research journal, volume i, issue iii found evidence for two different systems of visual recognition; however, researchers at the time had no explanation for why there would be more than one. evidence of the contributions of two distinct neural systems to animal navigation provided the first evidence of two distinct neural systems for recognition. this led to the proposal of the dual-system hypothesis, a possible resolution to the mystery of sternberg and atkinson & juola’s findings.[9,11] the dual-system hypothesis posits that two integrated, yet distinct systems of learning and memory exist in the mammalian brain: the improvisational system and the habit system. the improvisational system includes the hippocampus and surrounding medial temporal regions, the inferior temporal cortex, and the occipital cortex. it is responsible for constructing visual representations of the world and making responses to novel targets. the habit system includes the basal ganglia, the supplementary, premotor, and ventrolateral areas of the frontal cortex, the parietal cortex, and occipital cortex. it is responsible for encoding and serially generating sequences of actions.[5] see figure 1 for an fmri scan depicting portions of these systems. a variety of behavioral findings—such as spatial navigation, habit formation, and goal-directed actions—have been found to have a clear neural basis in the context of this dual-system model.[15] the improvisational system makes recognition judgments based on a test item’s perceived recency or novelty.[14] the most likely explanation for how the improvisational system computes recency/novelty is on the basis of the neural response to a perceptual stimulus. a repeated visual target activates exactly the same neurons, and the response of those neurons decreases as a function of the repetition due to habituation. so, if a test item was just shown in a previous study set, the improvisational system will detect that habituated (hence reduced) neural pattern and compute recency (a match with one of the just-seen study items). if it does not detect a habituated neural pattern upon presentation of the test item, it will compute novelty (a new item; no match). well before the establishment of the dualsystem hypothesis, groves and thompson (1970) first demonstrated that this was a plausible basis for the detection of recency/novelty.[6] contrarily, the habit system makes recognition judgments by retrieving a representation of the study items from memory and serially scanning the set of study items figure 1: clusters in the left/right hippocampus (a) and left/right anterior parahippocampal gyrus (b), which are part of the improvisational system. (c) shows clusters in the left/right caudate, which is part of the habit system. a better illustration of the habit and improvisational systems does not yet exist unfortunately, but these scans from sinha and glass (2017) show the crucial parts of each system. aresty rutgers undergraduate research journal, volume i, issue iii for a match with a test item.[3] sinha & glass (2017) found evidence for the involvement of both of these memory systems in a same/different matching task where participants judged whether two four-consonant strings were the same or different (e.g., study string “bdcf” vs. test string “bcfd”).[12] participants were presented with a study string followed by a test string to which they had to respond with either same or different. rt for different responses was an increasing function of the left-to-right location of the first difference between study and test strings, indicating that the habit system made different responses by serially comparing the consonants in the study and test strings from left to right. it follows from this line of reasoning that same responses ought to be the slowest since the participant presumably must compare all four test string items to all four study string items to determine that they are the same. however, rt for same responses was faster than different responses, indicating that the improvisational system was making same responses based on the perceived recency of the test string as a whole chunk without serially comparing each item in the study and test strings. fmri results corroborated this interpretation and showed that different responses were associated with activity in both the caudate and hippocampus while same responses were associated with just hippocampal activity. based on a connectivity analysis of this brain activity, sinha & glass (2017) posited that when the test string is presented, the improvisational system begins a holistic perceptual comparison between the study and test strings while the habit system simultaneously initiates a serial left-to-right comparison of the individual study and test string consonants.[12] during this holistic comparison, mental maps of the test and study strings are compared; if the strings match across all four positions, a strong match signal is produced by the hippocampus and a same response is made. when the study and test strings are identical, the match signal produced by the hippocampus is powerful enough to inhibit the serial comparison process initiated by the caudate and produce a same response. when the study and test strings differ, the match signal is not strong enough to inhibit the serial comparison process, so the habit system takes over and the serial comparison continues until a mismatch is found and a different response is then made. the functional roles of the improvisational and habit systems in a visual recognition task where the study and test items were both four-consonant strings were evident from the results of sinha & glass (2017).[12] the two systems appear to work in parallel and one of them responds depending on whether the test string is the same as or different from the study string. however, there was still the question of how the systems would function when the test item was a single consonant. kang, norman, zhou, becker, & glass (2021) replicated the same experimental design as sinha & glass (2017) with a fourconsonant study string, except the test item was a single consonant that either was (eliciting a same response) or was not (eliciting a different response) present in the study string. they found that when the different response was assigned to the spatially right response key and the same response was assigned to the spatially left response key, rt for same responses to targets was an increasing function of the test consonant’s position in the study string, indicating that the habit system made responses by serially scanning the study string until a match (if there was one) with the test consonant was found. however, when the response assignment was switched—the different response was assigned to the spatially left response key and the same response was assigned to the spatially right response key—rt for same responses was not significantly different among the four possible positions in the study string. this indicated that the improvisational system made responses based on the perceived recency/novelty of the test consonant without scanning the study string for a match.[7] kang et al. (2021) also conducted an iteration of this experiment where participants used two fingers on the same hand to make responses with keys in different spatial locations. they found the same result, confirming that the assignment of a verbal label to a key in a relative right or left spatial position determines which memory system is selected for the task. this refutes an alternative explanation that the assignment of verbal label same or aresty rutgers undergraduate research journal, volume i, issue iii [a] the participants in this experiment were students from two different rutgers courses. “course a” completed the experiment first, running 144 regular trials per session. after observing significant results from “course a”, the researchers increased the regular trials for “course b” to 360 per session to see if the effect observed in the results of “course a” remained. of the sample of 17, 10 participants were in “course a” and 7 participants were in “course b”. groups 1 and 2 mentioned above—which are the groups the researchers are comparing—are each composed of students from both of these courses. the final results detailed later are a combination of the results of both courses, but these combined results do not significantly differ from the initial results of just “course a” (data not shown). different to a specific hand causes the hemisphere associated with that hand to control which memory system is selected for the task.[7] therefore, when the study item is a four-consonant string and the test item is a single consonant, one system is selected to make responses in the task over the other based on the assignment of verbal labels to response keys in different spatial locations. sternberg (1966) did not take into account the effect of response assignment on rt; since it is not mentioned anywhere in his paper, he either only tested one response assignment or varied the response assignment but did not discriminate between rts from trials of different response assignments in his analyses. the purpose of this experiment was to investigate the effect of response assignment on rt by designating it as a separate independent variable in a replication of sternberg (1966). the researchers aim to determine whether response assignment influences the control of the improvisational versus habit system in sternberg’s immediate visual recognition task. based on the results of kang et al. (2021) where both memory systems were differentially involved in the task depending on the response assignment, the researchers hypothesize that the finding of contributions from both memory systems in sternberg (1966) may have been obscured originally by his disregard for response assignment’s effect on rt. thus, it could be the case that both memory systems are indeed involved in this task and that one system is selected to make responses over the other based on response assignment. 2 methodology participants were seventeen students distributed across two psychology courses at rutgers university-new brunswick in new jersey. at the time that this study was run, the researchers did not have access to a larger online subject pool; however, one researcher had remote access to a pool of students enrolled in his psychology courses and the study here was relevant to their course material, so they were used as subjects. participants consented to taking part in experiments as a feature of the course they were enrolled in and received course credit for their participation. eleven were female and six were male, aged 18-23. sixteen were right-handed and one was left-handed based on self-reports of handedness. stimuli were capitalized consonants randomly selected from a pool of all consonants (excluding y). in a previous replication of sternberg (1966), glass (1993) used three different kinds of stimuli in sternberg’s original design: digits, consonants from the first half of the alphabet, and consonants from the second half of the alphabet.[4] there was no difference in results based on the stimuli. study sequences were one to six consonants in length. each consonant was presented one at a time for 1.2 seconds each with no inter-stimulus intervals, consistent with the design and timing in sternberg (1966). in the “right key = different” response assignment condition, participants responded by pressing the spatially left “f” key if the test consonant was present in the study sequence (a same response) and the spatially right “j” key if it was not (a different response). in the “right key = same” condition, participants had the opposite response assignment. each participant completed one session with the “right key = different” response assignment and one session with the “right key = same” response assignment. participants were randomly assigned to one of two groups: group one (n = 9) completed the “right key = different” session first and the “right key = same” session second. group two (n = 8) completed the sessions in the opposite order.[a] all participants completed twenty practice trials at the beginning of each session to understand how the trials would run. rts from practice trials were not included in the larger analysis. there were an equal number of regular trials for each possible aresty rutgers undergraduate research journal, volume i, issue iii study sequence size (one to six). within the group of trials for a particular study sequence size, half of the subsequent test consonants were targets (test items presented in the study sequence) and half were lures (test items not presented in the study sequence). of the trials where the test consonant was a target, there were an equal number of trials for each possible position of the test consonant in the study sequence. the experiment was administered remotely; participants were given python code files containing the experiment code and instructions on how to run it through psychopy on their personal computers.[10] before the trials began, instructions were presented (see figures 2a & 2b). though there were no explicit instructions regarding which hands or fingers to use to respond, the researchers assume participants used their left hand to press the “f” key and their right hand to press the “j” key as that is the normal typing position. however, the researchers are more concerned with the spatial location of the response keys rather than which hands the participants used to respond; as explained earlier, it is the assignment of verbal labels to response keys in different spatial locations that determines the memory system (and thus visual recognition method) used to respond, not the assignment of verbal labels to different hands.[7] for each trial, a study sequence of one to six capitalized consonants was shown—one consonant at a time in the center of the screen—and then a fixation asterisk was presented followed by a test consonant. the purpose of the fixation asterisk was to focus the participant’s gaze to a particular point (the center of the screen) where the test consonant was about to be displayed so that rt did not include the time it takes for a participant to direct their gaze towards the test consonant, only the time it took for them to respond. the fixation asterisk was black for two seconds and then red for one second to indicate that the test consonant was about to appear. participants responded as fast as possible whether the test consonant had been present in the previous sequence. the test consonant disappeared immediately after a response was made and feedback on accuracy was presented, as was done in sternberg (1966).[13] feedback was one of four possible sentences (depending on the response) displayed for three seconds that said “correct, the test letter was/was not in the sequence” or “incorrect, the test letter was/was not in the sequence”. fixation asterisks appeared in between successive trials as well. 3 results a within-subjects anova was selected to assess the effect of response assignment condition (“right key = different” or “right key = same”), judgment (same response or different response), and study sequence size (1-6 consonants) on rt. the researchers selected a within-subjects anova because each subject completed one session for each response assignment, and they wanted to measure the difference in rt for those conditions for each subject. all variables were within-subjects variables. figure 2a: first instructions screen displayed upon opening the experiment. figure 2b: second instructions screen. aresty rutgers undergraduate research journal, volume i, issue iii the degrees of freedom, f-statistic, p-value, and effect size for all variables and interactions are contained in table 1. the effects of response assignment and judgment were not significant. the effect of study sequence size was significant, 𝑝 = 0.002. the interaction of response assignment and judgment was significant, 𝑝 = 0.034. the interaction of response assignment and study sequence size was significant, 𝑝 = 0.013. there was an effect of study sequence size on rt in the “right key = different” condition for targets (𝑝 < 0.001) and lures (𝑝 < 0.001), but there was no effect of study sequence size on rt in the “right key = same” conditions for targets or lures. this result is shown in figure 3, which includes the average rt for each condition/study sequence size across all subjects. none of the other interactions— between judgment and study sequence size and between condition, judgment, and study sequence size—were significant. variable/interaction degrees of freedom f-statistic p-value effect size (observed power) response assignment (ra) 1 0.647 0.433 0.118 judgment 1 2.605 0.126 0.329 study sequence size 5 4.285 0.002** 0.952 ra x judgment 1 5.353 0.034* 0.585 ra x study sequence size 5 3.086 0.013* 0.850 ra x study sequence size (right key = different – targets) 5 7.383 < 0.001** 0.999 ra x study sequence size (right key = different – lures) 5 14.189 < 0.001** 1.000 ra x study sequence size (right key = same – targets) 5 0.419 0.834 0.154 ra x study sequence size (right key = same – lures) 5 0.962 0.446 0.327 judgment x study sequence size 5 0.852 0.536 0.281 ra x judgment x study sequence size 5 0.402 0.846 0.150 table 1: ∗ indicates 𝑝 < 0.05. ∗∗ indicates 𝑝 < 0.01. aresty rutgers undergraduate research journal, volume i, issue iii figure 2: rt is only a function of study sequence size in the "right key = different" condition, statistically verified by the withinsubjects anova. rts (y-axis) are averages for the entire set of participants. error bars represent the standard error of the sampling distribution of each condition/study set size, specified in the table below (table 2). table 2: values for the standard error bars in figure 3. aresty rutgers undergraduate research journal, volume i, issue iii 4 discussion the purpose of this experiment was to investigate the effect of response assignment on rt in a replica of sternberg (1966). the researchers hypothesized that sternberg’s disregard for response assignment’s effect on rt may have obscured evidence for contributions from two memory systems— hence two methods of visual recognition—in his task. the finding of significant differences in rt between the “right key = different” and “right key = same” conditions suggested that serial scanning of the study sequence by the habit system only occurred in the “right key = different” condition. it was only in this condition that rt was a function of study sequence size, indicating serial comparison of the test item to each item in the study sequence; longer study sequences require more time to serially scan, and thus result in longer rts. the fact that rt was not a function of study sequence size in the “right key = same” condition suggests that the improvisational system made judgments of the perceived recency/novelty of the test consonant (based on detection of a habituated/non-habituated neural pattern, respectively) without comparing it to the study sequence. this expands upon sternberg’s original findings and provides more evidence for the dualsystem hypothesis. our results suggest that both the habit system and the improvisational system are indeed involved in this task, and an experimenter can control the memory system selected to make responses in the task by assigning different verbal labels to response keys in different spatial locations (in this case, different keys on a keyboard). the researchers do not have a definite answer as to why the habit system is selected for the task when the different verbal label is assigned to the spatially right response key and the same verbal label is assigned to the spatially left response key (right key = different condition) while the improvisational system is selected for the task when the response assignment is the opposite (right key = same condition). even though kang et al. (2021) confirmed that the assignment of verbal labels to response keys in different spatial locations determines which memory system gets selected for the task—not the assignment of verbal labels to different hands— there is still no clear answer as to why the two response assignment conditions result in the selection of different memory systems. for the same task, why should merely switching the verbal labels assigned to the response keys cause the participant to use a different memory system? the researchers suggest that the verbal label assigned to the spatially right key could be the deciding factor. language is localized in the left hemisphere, which processes information from the right side of space. therefore—since the verbal label assigned to the spatially right response key is initially processed by the languagedominant left hemisphere while the verbal label assigned to the spatially left response key is initially processed by the non-language-dominant right hemisphere—the language-dominant left hemisphere may interpret the task and select the appropriate memory system depending on verbal label assigned to the spatially right response key. perhaps when the label of different is assigned to the spatially right response key (right key = different condition), the experiment is interpreted as a difference-detection task that implies serial comparison, and so the habit system is selected to make responses. conversely, when the label of same is assigned to the spatially right response key (right key = same condition), the experiment is interpreted as a recency/novelty detection task that implies a purely perceptual judgement, and so the improvisational system is selected to make responses. however, this is merely a suggestion; the verbal label assigned to the spatially left key could instead be the determining factor, or it could be the spatial position of both keys. to determine if it is indeed the verbal label processed by the language-dominant hemisphere that controls which system is selected for the task, future studies ought to measure participant language dominance directly by using brain imaging techniques to see if the location of one’s language center has any effect on their responses. if individuals with atypical right hemisphere language localization use the opposite neural system as their typical counterparts to respond when the response assignment is the same, this would provide evidence for the notion that it is the verbal label processed by the language aresty rutgers undergraduate research journal, volume i, issue iii dominant hemisphere that controls the memory system selected for the task. there is an increased incidence of atypical right hemisphere language localization in strongly left-handed individuals.[8] however, it only occurs in about 25% of those individuals, and there was no evidence that the one left-handed participant in this study was using the opposite neural system as the other participants to respond when the response assignment was the same. they were also not an outlier regarding rt. a limitation of our experiment is that due to the move to remote instruction as a consequence of the covid-19 pandemic, participants completed this experiment in their own homes at their leisure, so environmental confounds could not be controlled as they normally would be in a laboratory setting. however, this experiment could still better control for those confounds while remaining remote, perhaps by running the sessions over video call to monitor participants. the sample size was also small, and the sizes of each group were slightly uneven (group 1, n = 9; group 2, n = 8). participant fatigue could have also been a confounding factor as breaks did not appear during any of the sessions. further replications of this experiment that include fmri recordings ought to be done to see if activation in the caudate/hippocampus like that observed in sinha & glass (2017) and kang et al. (2021) is found during this task∎ 5 acknowledgements this research was conducted under dr. arnold glass in his lab. data analyses referenced in the results section were facilitated by our lab’s graduate student, mengxue kang. 6 references [1] atkinson, r. c., & juola, j. f. (1973). factors influencing speed and accuracy of word recognition. fourth international symposium on attention and performance, (583–611). [2] atkinson, r. c., & juola, j. f. (1974). search and decision processes in recognition memory. contemporary developments in mathematical psychology: vol. 1. learning, memory & thinking. [3] checkosky, s. f., & baboorian, n. (1972). memory search for cvc and ccc trigrams. journal of experimental psychology, 96(1), 158–163. [4] glass, a.l. (1993). the role of generation in recognition. scandinavian journal of psychology, 34(3), 255-267 [5] glass, a. l. (2019). within the framework of the dual-system model, voluntary action is central to cognition. attention, perception, & psychophysics, 81(7), 2192 – 2216. [6] groves, p.m., & thompson, r.f. (1970). habituation: a dualprocess theory. psychological review, 77(5), 419-450. [7] kang, m., norman, m., zhou, w., becker, a., & glass, a. l. (2021). spatial location of response key selects the neural system for visual recognition (submitted, under review) [8] knecht, s., dräger, b., deppe, m., bobe, l., lohmann, h., flöel, a., ringelstein, e. b., henningsen, h. (2000). handedness and hemispheric language dominance in healthy humans. brain, 123(12), 2512-2518 [9] packard, m. g., & mcgaugh, j. l. (1996). inactivation of hippocampus or caudate nucleus with lidocaine differentially affects expression of place and response learning. neuropsychology of learning and memory, 65, 65–72. [10] peirce, j. w. (2007). psychopy psychophysics software in python. journal of neuroscience methods, 162, 8–13. [11] poldrack, r. a., & packard, m. g. (2003). competition among multiple memory systems: converging evidence from animal and human studies. neuropsychologia, 41, 245–251. [12] sinha, n., & glass, a. l. (2017). dissociating medial temporal and striatal memory systems with a same/different matching task: evidence for two neural systems in human recognition. the journal of general psychology. [13] sternberg, s. (1966). high-speed scanning in human memory. science, 153(3736), 652–654. [14] suzuki, w. a., & naya, y. (2014). the perirhinal cortex. annual review of neuroscience, 37, 39–53. [15] yin, h. h., & knowlton, b. j. (2006). the role of the basal ganglia in habit formation. nature reviews neuroscience, 7, 464– 476. my name is alexa becker. i am a rising senior majoring in psychology and cognitive science at rutgers-new brunswick. my hobbies include playing the bass, reading, and powerlifting. my research interests include visual recognition and, more broadly, the neural systems involved in memory and learning as well as aging and memory. i have been working in dr. arnold glass’ research lab for two years. dr. glass has been a professor of psychology at rutgers for 46 years. the experiment outlined in this paper was done virtually (as a result of the covid-19 pandemic). i composed the write-up, designed this experiment, provided participants with materials needed to complete the experiment remotely, and collected results data. our lab’s former graduate student, mengxue kang, assisted with designing the experiment and facilitated the statistical analyses. she has since successfully defended her dissertation and now works as a data scientist for cvs. aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. cost-effective remote operated vehicle zachary smolder, jingang yi (faculty advisor) ✵ abstract remote operated vehicles (rovs) are robotic submersibles controlled typically by a person at the surface of a water body. rovs can be applied to surveillance, environmental, and data recording jobs or tasks. the vehicle design may be modified to remove or add additional capabilities depending upon the specific purpose of the rov. in this paper, we explore using remote operated vehicles as a cheap and affordable water exploration platform. rov’s high cost is a prohibitive barrier to entry, preventing widespread adoption of rov for personal, research, and conservation uses. to address this problem, our paper explores a cost effective rov with video capturing and directional control capabilities. using state-of-the-art robotic technologies, a cost-effective competitive rov is designed and constructed. this rov was tested to a depth of 7 meters and has the potential to reach depths of up to 30 meters per its design. 1 introduction the field of robotics has always been known for high costs, especially when remotely operated vehicles (rovs) are taken into context.[4] to reduce these costs, corners are typically cut. this can include implementing a rov with a low resolution camera as well as building rovs out of flimsy materials. current rovs on the market include blue robotics, bluerov2, and geneinno’s tether titan. these rovs, similar to others, are expensive, typically costing more than one thousand dollars. furthermore, much of the market is saturated at this high price point but with limited capabilities in the rovs. most of these rovs are not able to be programmed and can only be controlled with the manufacturer's program or application. the idea of creating cost-effective robotics without compromising missions is one that we seek to explore throughout this research. our target audience is the rov community at large: specifically, our research is aimed at those who want an open-source, versatile, and quality rovs sold at a cost-effective price. in addition to creating software to control the rov, we seek to give the community the ability to add any water instruments and tools to the rov, allowing the robot to be marketable towards research or conservation organizations. the goal of this project is to design and fabricate a costeffective rov system that has similar functionality and performance to its higher-priced counterparts. the process of designing and building the rov as well as later design changes are detailed below. 2 methodology in order to facilitate the fabrication of the rov, several criteria were specified to come up with the initial layout of the robot. some of these initial ideas can be seen in the sketches in figure 1. the rov would have to have the following components: 1. hull and outer structure; 2. waterproof enclosure for electronics; 3. thrusters; and 4. electronics (including electronic speed controller, ethernet connection, batteries, camera, central computer, and general signal cables) after listing the initial specifications for the rov, we explored prior work surrounding rov designs. at first, the rov design just had simple sketches to brainstorm potential candidates for what it could look like. the thought process used to create these simple designs revolved around past rovs already created; these designs would be altered based upon technological specifications required for our purposes. aresty rutgers undergraduate research journal, volume i, issue iii one similarity between the sketches and past designs was how the four thrusts were mounted on the side. this allowed for easy directional control and several industry rovs use this design for this reason. a change that was made to the industry-inspired design was the easily accessible electronics compartment in the center of the rov. in these sketches, the original idea was to have a hatch that opened to allow for easy access on the bottom. figure 1 shows a rough sketch that includes preliminary thoughts based upon initial research. the vehicle was sketched as shown so that one could see side, front, and top-down views. numerous issues were present in some of these initial ideas. some of these issues included the rov’s inability to fit in a waterproof enclosure. the hatch idea was scrapped since waterproofing a seal around the hatch is challenging. additionally, these initial sketches had no clear method on how to secure the motors. at this time, the rov was conceptualized to have five motors: the rov would have four motors for depth and turning control and one motor for horizontal propulsion. in figure 2, there are several conceptualizations of the future designs that pertain to various parts created after extensive thought and research. the final design was one that incorporated several different ideas that we learned after much thought and deliberation. this thought process included the fact that we reduced the motor count to four since the fifth motor was ruled nonessential. the rationale behind this decision was that it would just get in the way, making the overall rov more expensive and heavier with no navigational benefit. this conclusion was met after realizing that the existing four thrusters could be arranged in a way to meet x y z navigation. furthermore, a more cylindrical design was decided upon which revolved around the waterproof enclosure center. the research shows that creating a unique shape with materials other than cast acrylic would be more expensive than solely building from a cast acrylic cylinder. figure 1: initial basic sketches of the rov design which includes a 5-thruster format. figure 2: basic sketches of part designs including motor housings and the hull that would go over an acrylic cylinder. aresty rutgers undergraduate research journal, volume i, issue iii based on this research, the simpler the design was, the more efficient the rov would be. this observation was a design philosophy that supports the use of both the fewest thrusters and making the hull as easy as possible to manufacture in the name of cost. this final design incorporates, in addition to previously mentioned design concepts, an acrylic enclosure in the center of the rov with a holding bracket for the electronics. furthermore, the four thrusters are held in a two-part forward-backwards thrust-based system, in tandem with a rise-and-dive based system. two of the motors are held vertically for rise and dive maneuvers. this is done alongside two motors held horizontally for forwards and backwards control. this final design stands to be efficient due to features like its cast acrylic center, as this reduces build time and stands to still be extremely strong. furthermore, the motor configuration also allows for only four motors to be necessary while not compromising stability or navigational ability. figure 3 and figure 4 show cad models of the final design based upon these constraints of motor configuration and a cast acrylic center. i. hull design / waterproof enclosure the inner acrylic enclosure was the priority for manufacturing. this was because all other parts of the rov would revolve around this central design element. in the end, cast acrylic was chosen to be used for the physical cylinder and half sphere cap. these two parts are the objects shown in figure 5, and in the case of this specific rov design, they are responsible for holding the electronics. the cast acrylic was chosen due to its strength and ability to be welded together to make a final product. weld-on 4 was used to properly seal the cylinder to the cap. a metal cap, manufactured by an outside party due to time constraints, was attached to the opposite side of the waterproof enclosure. figures 5-7 show the design, fabrication, and welding process of the water proof enclosure. the metal cap on the end opposite the dome of the waterproof enclosure was manufactured to be removable and enable cable wires to pass through to outside components of the rov. several ideas were considered to allow a waterproof fitting around the cables to prevent water leakage. originally, we planned to use standard weatherproof wire covers with hot glue. this was a failure figure 3: (isometric view) final cad design of the rov which uses a 4-motor design format. figure 4: (front view) this view shows the inside of the clear acrylic cylinder which will house the electronics. aresty rutgers undergraduate research journal, volume i, issue iii figure 5: (top left) this view showcases the cad model of the acrylic cylinder and dome when assembled. figure 6: (top center) the process of welding the acrylic dome to the acrylic cylinder required clamping and the use of weld-on 4. figure 7: (top right) the two halves of the acrylic are now welded together and ready for a hull to be attached. figure 8: (bottom left) initially weatherproof wire covers were used and they were found to not work adequately for the pressure cables would experience at desired testing depths. figure 9: (bottom center) the now potted cables need to dry for 48 hours in order to form a hardened seal around the cables and in the glands. figure 10: (bottom right) the dry glands were inserted into the endcap which separates the water from the rest of the electronics in the electronics enclosure. aresty rutgers undergraduate research journal, volume i, issue iii upon testing, which can be seen in figure 8. the hot glue leaked water after multiple tests and the weatherproof wire covers did not form a decent seal with the cables with which they interfaced. we instead decided to use the conventional approach of “potting”, as shown in figure 9. “potting” is when one sits cables in manufactured glands and fills the glands with epoxy to completely waterproof them. this allows for a permanent, waterproof bond around all the cables. figure 10 shows these glands put into the end cap. during the beginning phases of deciding upon materials and the fabrication processes of the hull, cost-effectiveness was kept in mind. this was to continue the overall pursuit of a cost-effective rov product. 3d printing was used to create the outer shell of the rov. due to its cheap yet sturdy nature, polylactic acid plastic (pla) made a great candidate for materials. figure 11 shows a 3d-printed hull over the acrylic electronics enclosure. pla was used to 3d print the main hull and motor housings. then, epoxy resin was used to make the pla stronger and create an even plane on which the water can glide, as shown in figure 12. this figure shows parts of the hull after having a fresh coat of polyester resin applied as well as the paint rollers used to apply this resin. after this process, the resin was sanded and polished evenly. a body filler coat was applied to fill in wherever the resin was not properly leveled. finally, black spray paint was used to cover the finished product. figure 13 shows all parts of the hull to be assembled. ii. propulsion design propulsion was a serious challenge both in terms of cost and power. the rov required a lot of motors based upon the design. the power constraints were also a serious issue. the decision to use 64mm 5-rotor 4500 kv brushless motors was made. the rationale behind this choice was that these motors had to be brushless to function underwater. these motors provided enough thrust with balanced power consumption to be wise choices (each motor is using power efficiently). we designed and fabricated a 3d-printed motor housing that allowed the motors to function perfectly in their desired loca tions. this housing can be seen in the middle of being 3d printed in figure 14. iii. electronics design several design decisions had to be made during the electronics and electronics tray fabrication process. the electronics tray had to be created so that it would fit in the electronics enclosure and put the computer, batteries, and electronics speed controllers (escs) in reasonable and safe places repetitively. in the end, a 3d-printed form factor or ergonomic design proved to be successful. this tray had holes to allow for zip-tying to keep cables from being loose and threading places to put wire through. figure 15 shows how the tray looked with electronics included on it. for the electronics, a raspberry pi 4 was used as the brain of the whole rov. this computer oversees running code, sending signals to parts of the robot, and, thus, controlling the robot’s movement and the use of electronic components. the raspberry pi 4 is a cheap yet powerful computer that worked well with the robot’s limited internal space while still allowing for uncompromising computing performance. in addition to these strengths, the raspberry pi 4 allows for users to connect any additional instruments to the rov easily by just connecting them to the pi and writing code. two 5200 mah lipo batteries were used to power the four motors. finally, one 1000 mah lithium-ion battery was used for the raspberry pi 4. motor bullet connectors were soldered in-house. batteries, motor leads, and the rov computer were wired. this was then finally put on the tray and into the enclosure. iv. coding the script used for the control of the rov was coded on python due to the power and simplicity of the language. the rov was controlled by connecting a raspberry pi 4 in the rov to a raspberry pi 4 on land by an ethernet cable. after a connection was established, software was used to see the desktop environment of the rov computer. once in this environment, the control script was run on the rov’s computer. this script functioned by sending pwm signals to the motors by branching if-else command aresty rutgers undergraduate research journal, volume i, issue iii figure 11: (top left) the hull was covered in car filler to properly smooth out any imperfections from the polyester resin process, hence its gray color. this hull is soon ready to be attached to the acrylic cylinder. figure 12: (top center) the parts of the hull were given polyester resin and left to dry for days to create a strengthened and smooth surface. figure 13: (top right) all the parts can be seen ready to put together with metric screws. figure 14: (middle left) a motor housing is in the middle of being 3d-printed. figure 15: (middle center) the electronics tray is fully loaded up with batteries, computer, camera, and electronic speed controllers. figure 16: (middle right) the electronics tray has been put together inside of the electronics enclosure. figure 17: (bottom left) the rov is fully assembled and ready for testing and implementation. figure 18: (bottom center) the rov’s buoyancy is very positive but the robot exemplifies adequate horizontal control. figure 19: (bottom right) the rov is being connected to initially with a separate control computer in order to begin operation. aresty rutgers undergraduate research journal, volume i, issue iii structures. once a forward, left, right, or dive command was sent via the control computer, the rov would respond by sending pwm signals to each of its respective motors. this code proved to be effective and can be seen in the appendix. 3 water test results once the fabrication was complete, testing took place. four testing iterations were conducted at two locations. all of these tests had sunny weather conditions with low water motion. during the first of these tests, buoyancy was measured on the rov. the observations centered on whether or not the rov was positively or negatively buoyant and by how much was the rov buoyant. the rov was found during this first test to be 500 grams positively buoyant. this issue was resolved through the attachment of a diver belt to the rov with 500 grams on the belt. during the second test, horizontal thrust and navigation was tested at a shallow location. now that the buoyancy was fixed, the rov stayed at a specific depth underwater. the robot during this test was able to perform fantastic left, right, and forward movement operations. this test was deemed successful because the rov was able to triumphantly navigate underwater and maintain a waterproof electronics enclosure while doing so. the rov was designed to function for about a half hour and at depths of 30 meters; however, the rov was never fully tested at depths of 30 meters. it was successfully tested at depths of about 7 meters due to the unavailability of a deeper body of water within time constraints. with increased depth comes increased pressure, which is why we feel it is necessary to physically test at 30 meters before deeming the rov fully capable in its depth-traversal abilities. during the third and fourth tests, the full capabilities of the rov were tested. depth as well as horizontal maneuverability was observed and analyzed. the rov during both these tests was able to exemplify left, right, forward, rise and dive capabilities at depths of 7 meters. at this depth, there were no traces of water found in the robot's electronics enclosure, and the camera used for vision was still in complete working order. other rovs on the market have been rated to go depths of 7 meters or more, making the robot unsuccessfully validated to beat the depth rating of other rovs for a competitive price. however, we still deemed these tests a success because the robot was able to strongly navigate and operate for about 100 minutes at a time, beating our original 30-minute estimate. the reasoning behind these results was the robot's ability to control the speed of its motors instead of running full throttle for the full thirty minutes. this operation time is comparable to several similarly priced rovs like the thorrobotics’ tenchrover, which costs $1,380 and can operate for 120 minutes during a single session. during the tests, several aspects of the product were both challenged and validated. in regard to the buoyancy issue, we addressed this by attaching a diver belt to the rov each testing time, this was an effective solution to a serious problem. it did, however, have room to improve as the belt can shift during operation. in the future, weights should be implemented into the hull physically, such as in extra pockets in the hull for weights to be inserted or making the hull thicker. this will make the buoyancy not shiftable during operation and result in a cleaner design. alternatively, if we were to keep the original hull design, we could make a form factor weight to secure in the electronics enclosure. this could accomplish similar results to redesigning the hull. in addition to the buoyancy issue, propulsion could have been improved to be more powerful. this extra power could help the rov navigate better in turbulent waters. communication could have been made better in the sense of cost. currently, the rov requires a second computer for control, but this could be solved through the use of a smartphone instead. one validating aspect of the tests was the ability of the rov to stay completely waterproof during operation. all electronics were left dry after posttest inspection in the waterproof enclosure. this inspection included visually and electronically testing each of the electronic components inside of the rov after each test. additionally, the rov was able to showcase operating times of up to 100 minutes which was similar to slightly higher priced rovs. furthermore, the rov was able to navigate at distances aresty rutgers undergraduate research journal, volume i, issue iii of 10 meters horizontally from shore. the biggest limiting factor in the rov’s ability to navigate at large distances is its ethernet cable length, which, for more money, could be adjusted for farther expeditions. the rov was recorded being able to display uninterrupted video feed while underwater with its camera. navigation at the speed of 1 knot or .5 m/s was recorded. due to the fact that this speed was recorded all in favorable environmental conditions, it should be noted that rov performance in rough waves would result in lower speed measurements. altogether, the robot was able to exemplify competitive battery life, navigational control, and waterproof capabilities, deeming it to be an outstanding rov. 4 discussion & conclusion looking back at the work over the past several months, we felt that the project has done a lot in terms of cost-effective robotics solutions. on average, rovs cost about between $1000 and $4000. this robot costing $900 means that it is cheaper than even the lowest priced commercial solution. this low cost allows for a profitable yet cost-effective sale price. this project has been successful in creating a cheap and capable rov. due to the rov’s opensource design, adding additional instruments would be as easy as just wiring these tools to the rov’s internal raspberry pi 4 and writing code to use these tools. several ideas can be focused on as examples of creating cost-effective rovs. design methods, such as that of designing around a cylindershaped piece of cast acrylic, proved fruitful in being cost effective while not sacrificing waterproof capabilities. additionally, the idea of 3d printing a hull structure around the cast acrylic proved to be good in allowing for a cheap and versatile platform to which components can attach. also, 3d printing was found to be a promising new technology that, when treated, was an inexpensive and strong waterproof material for parts attached to the hull. simplicity was the main guide for design decisions and was found to be a helpful philosophy when keeping costs low while keeping uncompromised capability. altogether, the project validates that it is possible to cut costs while still delivering a competitive rov. the result of the efforts of this project culminated into a robot that was able to have battery life comparable to other rovs with functional directional control. this was done at a fraction of the cost compared to many other rovs and, therefore, has implications to the rov community. we have shown that rovs can be produced and sold with decent functionality at competitive costs. the project was successful in meeting our original criteria. the rov was able to properly navigate waters with decent battery life and functional watertight compartments all at a cost lower than other rovs in the market. the project has several limitations that could be improved and changed in the future. this includes propulsion, communication, and buoyancy, which all have room for improvement in the current design. propulsion could be modified so that the thrusters are held more effectively and are more powerful. communication could have been improved on the robot, as the rov could have been made to be controlled with a phone, and, assuming the user already has a phone, would further reduce costs. furthermore, if the buoyancy was manufactured into the hull instead of being corrected with a dive belt, the rov would operate more efficiently since there would be no chance of shifting weight. all these places for improvement show several branches for further research and work to be done to explore these ideas∎ simplicity was the main guide for design decisions and was found to be a helpful philosophy when keeping costs low while keeping uncompromised capability. aresty rutgers undergraduate research journal, volume i, issue iii 5 acknowledgements we would like to express appreciation for the project sponsor blue robotics inc. for giving a discount on the water enclosure cap. furthermore, we would like to thank the faculty mentor professor jingang yi for editing the report and for guidance on the project. 6 references [1] aguirre-castro, o. a., inzunza-gonzález, e., garcía-guerrero, e. e., tlelo-cuautle, e., lópez-bonilla, o. r., olguíntiznado, j. e., & cárdenas-valdez, j. r. (2019). design and construction of an rov for underwater exploration. sensors (basel, switzerland), 19(24), 5387–. [2] ahmed, y., yaakob, o., & sun, b. (2014). design of a new low cost rov vehicle. jurnal teknologi. 69. [3] aniwaa. (2021) “the 11 best underwater drones in 2021.” aniwaa.com, retrieved 12 jan. 2021 www.aniwaa.com/buyers-guide/drones/best-underwater-drones/ [4] cherdo, l. (2020) “best underwater drones 2020: the 13 best rovs this year.” aniwaa [5] martos, g., abreu, a., gonzalez, s., & tremante, a. (2013). remotely operated underwater vehicle (rov) 100% report (unpublished b.s. thesis). retrieved from https://mme.fiu.edu/wp-content/uploads/2013/12/f13-sr-t-1.pdf [6] rahimuddin et al. (2018) j. phys: conf. ser. 962 012017 zachary smolder is a rutgers school of engineering sophomore. he is majoring in mechanical engineering and plans to pursue a career in entrepreneurship and film. he is very passionate about robotics and their multivariable applications. another interest of zachary’s is the great issue of microplastics in the world’s oceans which he hopes to help out in during his career. throughout quarantine, zachary decided that due to flexibility and freedom in his schedule he would do independent research during the fall 2020 semester. under the guidance of professor chigang yi, zachary was able to complete his project while at home and working remotely. zachary is glad for the opportunity to be able to do research during covid and looks forward to continuing his research endeavors while on campus. http://www.aniwaa.com/buyers-guide/drones/best-underwater-drones/ https://mme.fiu.edu/wp-content/uploads/2013/12/f13-sr-t-1.pdf aresty rutgers undergraduate research journal, volume i, issue iii 7 appendix below is the total cost and expenses for the project. it is important to note that the rov only cost $930.65, seeing as many rovs in the modern day cost thousands of dollars.[3] this expense chart sums up the true innovation and benefit of the design. note that in order to calculate the cost of labor, a rate of $12.95 was used for a total labor time of 12 hours which would be enough in tandem with 3d printings and other tools to assemble the robot. figure 20: detailed expense and cost report showing the total price of the rov’s construction. aresty rutgers undergraduate research journal, volume i, issue iii here is a comparison chart further showing the great value our rov has over more expensive competition. while the rovs below have better specifications than our rov in some ways, their rovs are not open source and cost significantly more than ours for minimal additional capabilities. the only exception to this is blue robotics’ bluerov2, which has open-source abilities but at a price four times higher than that of our rov. notilo plus ibubble (cost is $4,499) ● not open source ● battery life of 60 minutes ● diving depth of 60 meters geneinno t1 (cost is $2,499) ● not open source ● battery life of 240 minutes ● diving depth of 150 meters aquarobotman nemo (cost is $1,799) ● not open source ● battery life of 180 minutes ● diving depth of 100 meters bluerov2 (cost is $4,268) ● is open source ● diving depth of 100 meters aresty rutgers undergraduate research journal, volume i, issue iii figure 21: schematic showing the wiring of the rov by component. this was used to have a better overview of how to wire the rov during construction. aresty rutgers undergraduate research journal, volume i, issue iii python control script import os import time os.system ("sudo pigpiod") time.sleep(1) import pigpio import picamera import pygame camera = picamera.picamera() esc1 = 27 esc2 = 17 esc3 = 5 esc4 = 6 pi=pigpio.pi() pi.set_servo_pulsewidth(esc1,0) pi.set_servo_pulsewidth(esc2,0) pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) max_value = 1600 min_value = 500 print ("initiating launch sequence") def calibrationsequence(): pi.set_servo_pulsewidth(esc1,0) pi.set_servo_pulsewidth(esc2,0) pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) print("press enter to confirm calibration") inp = input() if inp == '': pi.set_servo_pulsewidth(esc1,max_value) pi.set_servo_pulsewidth(esc2,max_value) pi.set_servo_pulsewidth(esc3,max_value) pi.set_servo_pulsewidth(esc4,max_value) print("press enter again") inp = input() if inp == '': pi.set_servo_pulsewidth(esc1,min_value) pi.set_servo_pulsewidth(esc2,min_value) pi.set_servo_pulsewidth(esc3,min_value) pi.set_servo_pulsewidth(esc4,min_value) print ("one moment") time.sleep(10) pi.set_servo_pulsewidth(esc1,0) pi.set_servo_pulsewidth(esc2,0) pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) time.sleep(2) print("arming") pi.set_servo_pulsewidth(esc1,min_value) pi.set_servo_pulsewidth(esc2,min_value) pi.set_servo_pulsewidth(esc3,min_value) pi.set_servo_pulsewidth(esc4,min_value) controlsequence() def controlsequence(): print ("thrusters engaging") time.sleep(1) speed = 500 print ("+ for increase, for decrease") while true: pi.set_servo_pulsewidth(esc1,speed) pi.set_servo_pulsewidth(esc2,speed) pi.set_servo_pulsewidth(esc3,speed) pi.set_servo_pulsewidth(esc4,speed) inp = input() if inp == "-": speed -= 100 elif inp == "+": speed += 100 elif inp == "0": stop() else: print ("invalid control") def forward(): pi.set_servo_pulsewidth(esc1,max_value) pi.set_servo_pulsewidth(esc2,max_value) def left(): pi.set_servo_pulsewidth(esc1,max_value) pi.set_servo_pulsewidth(esc2,0) pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) def right(): pi.set_servo_pulsewidth(esc1,0) pi.set_servo_pulsewidth(esc2,max_value) aresty rutgers undergraduate research journal, volume i, issue iii pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) def dive(): pi.set_servo_pulsewidth(esc1,800) pi.set_servo_pulsewidth(esc2,800) pi.set_servo_pulsewidth(esc3,max_value) pi.set_servo_pulsewidth(esc4,max_value) def launch(): print ("beginning underwater exploration") print ("control with wasd") camera.start_preview(fullscreen=false, window=(100,20,640,480)) pi.set_servo_pulsewidth(esc1,0) pi.set_servo_pulsewidth(esc2,0) pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) while true: inp = input() if inp == "w": print ("forward") forward() elif inp == "a": print ("left") left() elif inp == "d": print ("right") right() elif inp == ("s"): print ("dive") dive() else: print ("finish launch") stop() def stop(): pi.set_servo_pulsewidth(esc1,0) pi.set_servo_pulsewidth(esc2,0) pi.set_servo_pulsewidth(esc3,0) pi.set_servo_pulsewidth(esc4,0) pi.stop() def startup(): print ("launch sequence...press 1 to start, press 2 to begin system control, and press 0 to abort") inp = input() if inp == "1": calibrationsequence() elif inp == "2": launch() elif inp == "0": stop() else: print("abort") startup() aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. workforce compensation policies for lead teachers in state-funded preschool zaynab khan, allison friedman-krauss (facult advisor) ✵ abstract each state in the us has different compensation parity policies for their early childhood education programs. currently, public preschool teachers often have similar qualifications to k-3 teachers but earn significantly lower salaries. compensation parity policies ensure that equivalent work and qualifications are compensated with equivalent pay and benefits. using data collected by the national institute of early education research (nieer), i organized and analyzed policy data from all state-funded preschool programs in the u.s., with a focus on state compensation parity policies for lead preschool teachers. ultimately, my purpose was to understand state-funded preschool compensation parity for lead teachers in order to identify areas of improvement for the workforce within early education programs. i initially hypothesized that pre-k programs that required pay parity did not cost states more per child than pre-k programs that did not require pay parity. literature from nieer and other early education research institutions has shown that quality early education programs are critical in a child’s formative years and suggests that a more satisfied workforce yields more positive outcomes for children. parity policies in state-funded pre-k programs are not highly correlated to spending per child or program quality. moreover, parity policies improve workforce well-being and should still be incorporated into state-funded pre-k. results show that there are only six programs across four states that require full salary parity for lead preschool teachers and no states require benefit parity for lead preschool teachers in both public and private settings. no clear pattern has emerged between compensation parity policies and state preschool spending or program quality. 1 introduction i. what is compensation parity? many preschool teachers earn substantially less than their counterparts in k-3 despite having comparable education and training.[3] a lack of parity in both salaries and benefits between preschool and k-3 teachers has important implications for both the teachers themselves as well as for the children they serve. in this study, i apply the center for the study of child care employment’s (cscce) definition of workforce compensation parity to analyze the workforce compensation parity policies in state-funded preschools in the u.s.[10] according to cscce, pay parity occurs when preschool teachers and k-3 teachers with equal qualifications and work experience receive equal pay (table 1). components of compensation salary parity benefit parity fu ll pa r it y equivalent starting salary with k-3 lead teachers with the same qualifications equivalent salary schedule with k-3 lead teachers with the same qualifications salary schedule is prorated for differences in length of workday or year requires retirement benefits requires paid time off requires healthcare benefits requires other benefits pa r ti a l pa r it y not equivalent or no salary schedule some benefits are equivalent, but others are not table 1: compensation parity & related forms of compensation improvement note. definitions of parity as defined above. source: adapted from whitebook, marcy, and caitlin mclean. “in pursuit of pre-k parity.” center for the study of child care employment, apr. 2017. aresty rutgers undergraduate research journal, volume i, issue iii table 1 outlines how cscce defines compensation parity. this definition divides parity into multiple categories such as full parity, partial parity, sub parity, and alternative forms of compensation. parity is equivalent compensation which includes: 1) starting salary, 2) salary schedule, 3) benefits, and 4) payment for professional responsibilities for employees with similar qualifications. full parity is when teachers receive equivalent treatment across all four parity categories. partial parity is the equivalent treatment across some of the parity categories. for example, starting salary could be the same, while salary schedule, benefits, and/or payment for professional responsibilities may differ.[1] sub-parity is similar to partial parity but does not entail equivalent treatment across all four parity categories. alternative forms of compensation include strategies to improve compensation but do not fall into one the three other categories.[1] only full and partial parity will be further discussed in this paper. early childhood education in the u.s. consists of a variety of programs such as the federal head start program, childcare, state-funded preschool, or private school. this paper will specifically discuss state-funded preschool. state-funded preschool is defined as a program that is directed, funded, and controlled by the state, serves mainly 3 and 4 year-olds, and focuses on early childhood education.[3] furthermore, state-funded preschool is a distinct program from state-subsidized childcare, programs that do not need to include early learning components. some states such as new jersey have multiple preschool programs which are all funded and instituted by the state. each state is responsible for the licensing and administration of these programs. state-funded preschool programs may be provided in both public (e.g., public schools) and nonpublic settings (e.g., childcare centers) when states have mixed-delivery systems. this paper will specifically focus on parity for lead teachers in state-funded pre-k, in contrast to assistant teachers or other staff employed at statefunded preschool programs. lead teachers are often required to have the similar qualifications to k-3 teachers (such as a bachelor’s degree). assistant teachers often do not need the same level of education as lead teachers and are typically not present in k-3 programs. qualifications requirements for lead teachers can be compared to those of k-3 teachers as they require similar education levels. ii. why is compensation parity important? prior research suggests that teacher compensation affects both teachers as well as their students.[9] for example, improved well-being of the preschool workforce is associated with higher quality preschool programs. one recent review found that teacher stress was associated with poor teacherchild relationships and ineffective teaching of social and emotional skills.[9] teachers’ financial instability was identified as a significant source of stress and thus associated with less positive child outcomes. low emotional wellness of teachers is associated with negative effects on students such as misbehavior, indicating the relationship between teacher wellness and student outcomes.[7] furthermore, low wages are correlated with high turnover rates.[9] consequently, this may negatively affect program quality for children. parity policies in general are associated with higher wages for lead teachers and can improve teacher well-being. compensation parity is critical because it is associated with higher wages for preschool teachers, putting them on a pay scale like their k-3 colleagues. while both k-3 and many preschool teachers are required to have the same educational backgrounds (e.g., at least a ba), preschool teachers typically earn significantly less than k-3 teachers. teachers in the federal head start program with at least a ba still make, on average, $23,000 less than public k-3 teachers.[2] in a study conducted by the research institute child trends, researchers examined alabama’s first class pre-k program after the state increased funding for salary parity in 2016.[5] teachers began receiving pay similar to public k-3 teachers. in this qualitative study, some preschool directors reported aresty rutgers undergraduate research journal, volume i, issue iii that they could lead more effectively since parity resulted in easier recruitment of high-quality teachers. furthermore, the researchers also noticed an increase in teamwork after the policy took effect. from the preschools the researchers surveyed, all directors reported high staff retention rates of teachers. more lead teachers opted to remain employed within alabama’s first class pre-k program than before, and directors and teachers noted that there was an increased amount of interest from public school teachers to find employment in alabama’s first class pre-k program. however, this did not apply to auxiliary staff, who were not affected by the new parity policy. teachers also reported better economic and personal well-being after salary parity went into effect.[5] while this study was conducted in 2016, the parity policies remain in place in alabama’s first class pre-k.[8] taken together, this research by child trends and the paper early care and education teacher well-being: associations with children's experience, outcomes, and workplace conditions: a research-to-policy brief suggests that compensation parity is an essential ingredient for improving preschool program quality. as a result, it can potentially improve outcomes for children in their formative years and teachers’ well-being. further, this research highlights that it is important for states to prioritize compensation parity policies in their state-funded preschool programs and that it is necessary to understand the landscape of compensation parity policies in state-funded preschool. iii. what are the main questions relating to the state of compensation policies in statefunded preschool? in response to the existing literature that indicates the importance of the compensation parity policies, i examined the existing structure of parity policies within state-funded pre-k. in this exploratory paper, i addressed the following questions: 1) what do we know about preschool parity policies? 2) which state-funded preschool programs institute partial or full parity for lead teachers? 3) how do parity policies differ across public and nonpublic settings within state-funded preschool programs? 4) how are parity policies related to teacher salaries for public programs? 5) to what extent do parity policies relate to state preschool spending and program costs? 6) to what extent do parity policies relate to the quality of state-funded preschool programs? ultimately, this research focuses on pay parity policies within the different state-funded preschool programs across the united states and their relationship to cost and compensation in pre-k. this research was conducted to understand if programs with pay parity policies spent more per child and if programs with pay parity policies had a higher wage for teachers. 2 methodology the national institute for early education research (nieer) has tracked state-funded preschool enrollment, spending, and policies in order to support quality programs across states since 2002. recently, nieer has taken an in-depth look into compensation parity policies, including the extent to which they support the preschool teaching workforce. data used in this paper pertain to the 20172018 school year. nieer surveys all 50 states, the district of columbia, and u.s. territories about their current policies regarding their state-funded preschools. all states that have state-funded preschool programs have provided requested data. this includes 61 preschool programs across 44 states and d.c.; the programs serve over 1 million children. this information is compiled into the annual state of preschool yearbook.[3] to explore lead preschool teacher compensation parity policies, i used data about state-funded preschool policies, such as each state’s required teacher qualifications, average salary of lead preschool teachers, and average salary of lead k-3 teachers, for programs in both public and nonpublic settings. that is, although all programs receive state funding, many states use a mixed-delivery system where some preschool programs operate in public aresty rutgers undergraduate research journal, volume i, issue iii [a] consistent with u.s. government statistical reporting practices, the district of columbia will be referred to as a “state” throughout this report. hence, we report 45 “states” providing state-funded preschools. schools and others operate in nonpublic settings such as private childcare centers or head start programs. while data were collected on the 44 states, d.c., and territories (only guam) that have a state-funded preschool program, only the states and d.c. were included in the analyses for this study.[a] quality as discussed throughout the paper refers to the nieer’s quality standards, which evaluate programs according to 10 benchmarks: early learning and development standards, curriculum support, teacher and assistant teacher educational background, training, and professional development, class sizes, staff-student ratios, and continuous quality improvement system. i conducted a thorough literature review of nieer’s previous publications on compensation parity as well as other credible publications to understand compensation parity policies and their importance. to analyze the current state of compensation parity for lead teachers in state-funded preschool programs, i analyzed raw survey data collected by nieer for the 2017-2018 state of preschool yearbook. i also conducted statistical analyses, including correlations and regression analyses, on nieer’s data to explore associations between compensation parity policies, state preschool spending, and preschool quality. for the regression analysis between parity policies and state preschool spending, the length of the day was included as an independent variable in the regression. the dependent variable, adjusted spending per child in each state’s pre-k program, was regressed against the following binary variables: if they program offered part day, full day, or extended day service and whether the program offered partial salary parity, full salary parity, and benefit parity within their programs. a regression analysis was also conducted between parity and quality benchmarks. given that prior literature finds an association between teacher compensation and preschool quality, i also explored the association between state preschool programs’ parity policies and nieer’s 10 benchmarks for preschool quality.[6] examples of nieer’s quality standard benchmarks include supports for curriculum implementation, qualifications of lead (at least a b.a.) and assistant teachers (at least a child development associate (cda)), class size (maximum 20 children) and staff to teacher ratios (1:10 or better), as well as other non-workforce related specifications (such as required health screenings for children). all the programs are graded on quality on a scale from 1-10. the programs are evaluated across the 10 policy standards and then given an average total score. 3 results & discussion i. what do we know about preschool parity policies? of the 44 states and district of columbia that offered state funded preschool, some states required similar qualifications for lead preschool teachers and public-school k-3 teachers. according to nieer’s 2018 workforce special report, 28 states required that all lead teachers in state-funded preschool have at least a bachelor's degree (b.a.). 25 states required that lead teachers have an additional state-specified certification. however, of those states that require both a b.a. and a certification, only four —hawaii, new jersey, oklahoma, and rhode island —require starting salary parity.[3] ii. which state-funded preschool programs institute partial or full parity for lead teachers? i explored state preschool parity policies to determine which states have full and partial parity policies in place for lead preschool teachers. nieer collected data on state policies regarding both salary parity and benefit parity in state-funded preschool. as per the data from the 2017–2018 state of preschool yearbook, there were no states that required both salary and benefit parity for all lead teachers in state-funded preschool programs, which includes teachers in programs that offer both nonpublic and public settings. fifteen state-funded preschool programs have policies in place that meet the requirement for at least one component of salary parity, such as equivalent starting salaries but not an equivalent sal aresty rutgers undergraduate research journal, volume i, issue iii ary schedule. however, there are only six statefunded preschool programs that require full salary parity, which requires all three components (equivalent starting salary, equivalent salary schedule, and proration of salary based on work hours, see table 1) for lead teachers in both public and nonpublic programs. these six programs are in four states: california, minnesota, new jersey, and oregon. only six programs meet at least one component of benefit parity, while no state program meets all three categories of benefit parity (see table 1) for teachers in all settings. iii. how do parity policies vary across public and non-public settings? in many state-funded preschool programs, parity policies differ for preschool teachers employed by public and non-public schools. parity policies are more likely to be in place for preschool teachers employed by public rather than non-public schools. according to the nieer workforce special report, “in states requiring the same degree requirements for all teachers, almost 70% (of those with data) reported wage disparities where preschool public school teachers earned up to $21,136 more than private preschool teachers in the same program. disparities are even larger when comparing state-funded preschool teachers where a bachelor’s degree is required to public school k–3 teachers. preschool pay gaps of $20,000 to $30,000 per year are common.”[3] moreover, there are 19 state-funded preschool programs that require full salary parity only for teachers employed by public programs. oregon, california, new mexico, texas, kentucky, minnesota, west virginia, north carolina, south carolina, new jersey, delaware, rhode island, vermont, and maine have some form of salary parity. only three states (nevada, d.c., and georgia) require full benefit parity for lead preschool teachers employed by figure 1: map of salary and benefit parity policies for lead teachers in state-funded preschool programs employed by public schools compared to k-3 lead teachers note: map outlining parity by state public programs. aresty rutgers undergraduate research journal, volume i, issue iii [b] while the state of california considers tk a kindergarten program, it meets nieer’s definition of a state-funded preschool program[3]. despite tk being viewed as kindergarten by ca, there is still a large salary gap between preschool teachers in tk and elementary school teachers. public schools. of these, only the district of columbia and nevada require both salary and benefit parity for state-funded preschool teachers employed by public schools. iv. how are parity policies related to teacher salaries for public programs? one result of the lack of parity policies in state-funded preschool programs is that preschool lead teachers in public programs with similar qualifications as public k-3 teachers often receive significantly less pay for similar work. according to the 2017–2018 yearbook data, among the states that reported preschool teacher salary, state-funded preschool teachers employed by public schools earned $7,456 less on average than public elementary school teachers. this gap is even larger when comparing lead state-funded preschool teachers employed by nonpublic settings with k-3 teachers; on average, this gap is $17,729.[3] i also explored the extent to which salary parity policies in state-funded preschool were related to lead preschool teacher salaries. however, only 30 states provided data on average salaries for lead preschool teachers. when examining the states that require full salary parity for all programs, there still exists a significant wage gap between preschool teachers and k-3 teachers. for example, in california’s transitional kindergarten (tk), a preschool program that serves 4-year-olds, the wage gap is $18,126 despite tk being considered an early year of kindergarten by california.[b] similarly, in oregon’s preschool promise program, the wage gap is $15,413. both programs institute salary parity, yet there still exists a large discrepancy in wages.[3] the inconsistency of parity policies and teacher salaries across the different programs can be partly attributed to the split between public and nonpublic program designations. for example, new mexico institutes full salary parity for preschool teachers in public schools but not those teaching in private settings (though still part of the state-funded program). the wage gap between state-funded, public school lead preschool and k-3 teachers in new mexico is only $2,745. conversely, in arkansas, where there are no salary parity policies, the wage gap between public lead preschool teachers and public k-3 teachers is $8,824. even after adjusting for cost of living, the wage gap in arkansas is significantly higher than the wage gap in new mexico.[3] v. to what extent do parity policies relate to state spending and program costs? a regression analysis showed a statistically insignificant correlation between salary parity policies and state spending per child for preschool. as an example, nevada and the district of columbia are the only two states that have parity policies for both salary and benefits for lead preschool teachers employed in public settings (see table 2). yet the state spending per child is significantly different in these two states. the district of columbia spends $17,545 per child, which is the most compared to any other state program. the state of nevada spends $4,025 per child (which is only slightly lower than the median of $4,769 per child). when adjusted for cost of living, nevada spends $4,124 per child and the district of columbia spends $15,008. nevada still institutes full compensation parity for public programs despite spending significantly less than the only other state that institutes full parity policies for both salary and benefits. however, it is worth noting that nevada’s programs are part-day and only serve students for 2.5 hours a day while the district of columbia’s program serves students for 6.5 hours a day (as indicated in table 2). table 3 indicates that it is somewhat surprising that there is only a weak but nonsignificant correlation between spending per child and parity policies. this is likely due to other variables such as differing operating times (as in the example of district of columbia and nevada). that is, some programs are full-day while others are half-day, which affect costs. programs also vary in the total number of hours and/or days per year, which also affects costs. shorter programs will cost significantly less than programs that serve students for longer periods of time. results of the regression analysis indicated only a weak and nonsignificant correlation between pay parity policies and the spending per child of each aresty rutgers undergraduate research journal, volume i, issue iii [c] full day programs serve students for at least 5.5 hours a day. extended day programs serve students for more than 6.5 hours a day. part day programs serve students for at least 3 hours a day but not more than 5.5 hours. table 2: state preschool spending per child, minimum operating schedule, and parity policies[c] note. this table represents data available on the operating schedule, spending per child, partial and full salary parity, and full benefit parity. aresty rutgers undergraduate research journal, volume i, issue iii b (se) p part-day program (n =35) -375.9859 1165.289 p > .05 full-day program (n=10) 949.3526 1428.907 p > .05 extended-day program (n=7) 857.1525 1640.042 p > .05 partial salary parity (n=29) -113.4493 1137.561 p > .05 full salary parity (n=19) 2041.8102 1223.313 p > .05 full benefit parity (n=3) 1266.2790 1996.638 p > .05 b (se) p part-day program (n =35) 1.5262705 0.7695571 p < .05 full-day program (n=10) 2.8829284 0.9436502 p < .01 extended-day program (n=7) 2.3051779 1.0830838 p < .05 partial salary parity (n=29) 1.4399083 0.7512457 p > .05 full salary parity (n=19) -0.9485807 0.8078760 p > .05 full benefit parity (n=3) -1.7629463 1.3185800 p > .05 table 3: regression analysis predicting spending per child from teacher salary parity table 4: regression analysis predicting quality standard benchmarks from teacher salary parity aresty rutgers undergraduate research journal, volume i, issue iii state program (𝑝𝑝 > .05). there are several limitations regarding finding a relationship between parity policies and state spending. many states have varied operating schedules. while the analysis accounted for operating schedule, it did not account for how many calendar days children are served. furthermore, class size, which varies amongst programs and is often a driver for program costs, was not controlled for. vi. how does parity relate to the quality of programs? results of the regression analysis from table 4 indicated only a weak and nonsignificant correlation between pay parity policies and the number of quality standard benchmarks met by each state preschool program (𝑝𝑝 > .05). in 2018, the programs that met every minimum quality benchmark were in michigan, alabama, and rhode island. nieer does not include parity as one of their ten quality benchmarks and does not guarantee quality. none of these states institute full parity for their lead preschool teachers. however, both rhode island and alabama required partial salary parity in their programs. the results suggest that there is no direct link between policies related to quality and parity policies.[3] however, it is possible that preschool program quality is affected by consequences of parity such as high turnover rates. similar to state spending, there are limitations regarding the relationship between quality and parity policies. while the literature suggested a relationship between parity and teacher well-being (and consequently quality), there are several factors not included in the analysis that could explain the weak correlation. the quality standards benchmarks represent policies related to quality but do not represent actual program quality. further, the benchmarks include markers of quality that do not relate to the lead teachers at all such as a policy requiring that children receive health screenings. 4 summary & conclusion in conclusion, compensation parity policies are policies that require preschool teachers to receive equal compensation, including both salary and benefits, on par with similarly qualified k-3 teachers. in general, parity policies should be associated with higher wages. however, results of this analysis indicate that few states institute parity policies in statefunded preschool. teachers employed by public schools are more likely to be supported by parity policies in comparison to preschool teachers employed by nonpublic settings. moreover, there seems to be little correlation between state preschool spending per child and parity policies, but some evidence that a smaller wage gap between public preschool and public k-3 was associated with preschool programs having parity policies. in addition, qualitative studies indicate that parity policies improve teacher well-being which can improve program quality of state-funded programs. however, i found no direct link between program quality and enforced parity policies. this study is a first step in understanding state-funded preschool compensation parity for lead teachers. results indicate that most statefunded preschool programs have a long way to go to ensure that preschool teachers receive compensation packages on par with their k-3 counterparts. future research could explore barriers that contribute to these discrepancies. governments can create labor policies, such as parity and starting salaries, for their own workforce and therefore for teachers in public settings. however, it likely would be seen as government overreach for states to regulate workresults indicate that most statefunded preschool programs have a long way to go to ensure that results indicate that most state-funded preschool programs have a long way to go to ensure that preschool teachers receive compensation packages on par with their k-3 counterparts. aresty rutgers undergraduate research journal, volume i, issue iii force policies for privately-run (non-public) programs even if they receive state funding.[1] furthermore, many states will delegate the implementation of parity policies to municipalities. some local governments will require parity policies in their districts while others will not, resulting in different parity policies and varying wages across a state. indeed, for many preschool programs that did not have compensation parity policies, these policies were determined at the local level by school districts. additionally, as more state-funded preschool programs implement parity policies, future research could explore how changes in parity policies over time affect teacher salaries and turnover as well as preschool quality, spending, and child outcomes. current budgeting deficits due to covid-19 pandemic may result in the regression of parity policies, though so far this is not evident.[4] while backtracking on parity policies is not a desirable outcome, it may serve as an opportunity to research the effects of compensation parity on teacher well-being, program quality, and child outcomes. ensuring that the preschool workforce is well-compensated will be crucial to ensure that preschool programs have positive effects on young children’s development, as well as for teachers’ own well-being∎ 5 acknowledgements i would like to thank dr. allison friedmankrauss as well as the national institute of early childhood education research for their support in this project. 6 references [1] barnett, w. s. & kasmin, r. (2017). teacher compensation parity policies and state-funded pre-k programs. new brunswick, nj: the national institute for early education research and berkeley, ca: center for the study of child care employment, university of california, berkeley [2] barnett, w. s., friedman-krauss, a., weisenfeld, g., horowitz, m., kasmin, r., & squires, j. h. (2017, may). the state of preschool 2016. http://nieer.org/wp-content/uploads/2017/09/full_state_of_preschool_2016_9.15.17_compressed.pdf [3] friedman-krauss, a., barnett, s., garver, k., hodges, k., weisenfeld, g., & dicrecchio, n. (2019, april). the state of preschool 2018. http://nieer.org/wp-content/uploads/2019/08/yb2018_full-reportr3wappendices.pdf [4] friedman-krauss, a., barnett, s., garver, k., hodges, k., weisenfeld, g., & gardiner, b. (2021, august). the state of preschool 2020. https://nieer.org/wp-content/uploads/2021/08/yb2020_full_report_080521.pdf [5] gebhart, t., carlson, j., harris, p., & epstein, d. (2020 june). workforce perceptions and experiences with the alabama early care and education salary parity policy. childtrends.org [6] jennings, p. a. (2014). early childhood teachers’ well-being, mindfulness, and self-compassion in relation to classroom quality and attitudes towards challenging students. [7] jennings, p., & greenberg, m. (2009). the prosocial classroom: teacher social and emotional competence in relation to student and classroom outcomes. review of educational research, 79(1), 491-525. retrieved may 18, 2021, from http://www.jstor.org/stable/40071173 [8] office of the governor of alabama. (2021, may 18). governor ivey announces new first class pre-k classrooms in first round of funding. office of the alabama governor. https://governor.alabama.gov/newsroom/2021/05/governor-iveyannounces-new-first-class-pre-k-classrooms-in-first-round-offunding/ [9] smith, s., & lawrence, s. (2019, march). early care and education teacher well-being: associations with children's experience, outcomes, and workplace conditions: a research-to-policy brief (rep.). (eric document reproduction service no. ed597580) [10] whitebook, m. &mclean, c. (2017). in pursuit of pre-k parity: a proposed framework for understanding and advancing policy and practice. berkeley, ca: center for the study of child care employment, university of california, berkeley and new brunswick, nj: the national institute for early education research. zaynab khan is a graduate from the class of 2021. she majored in mathematics and minored in political science and economics. she was an honors college student as well as a member of the douglass residential college community. as part of the aresty research assistant program, she worked with the national institute of early education research at the rutgers graduate school of education to conduct research on early education policy, specifically on workforce compensation for early education teachers. she hopes to eventually pursue higher education in economics and conduct further policy related research. http://nieer.org/wp-content/uploads/2017/09/full_state_of_preschool_2016_9.15.17_compressed.pdf http://nieer.org/wp-content/uploads/2017/09/full_state_of_preschool_2016_9.15.17_compressed.pdf http://nieer.org/wp-content/uploads/2019/08/yb2018_full-reportr3wappendices.pdf http://nieer.org/wp-content/uploads/2019/08/yb2018_full-reportr3wappendices.pdf https://nieer.org/wp-content/uploads/2021/08/yb2020_full_report_080521.pdf https://nieer.org/wp-content/uploads/2021/08/yb2020_full_report_080521.pdf http://www.jstor.org/stable/40071173 https://governor.alabama.gov/newsroom/2021/05/governor-ivey-announces-new-first-class-pre-k-classrooms-in-first-round-of-funding/ https://governor.alabama.gov/newsroom/2021/05/governor-ivey-announces-new-first-class-pre-k-classrooms-in-first-round-of-funding/ https://governor.alabama.gov/newsroom/2021/05/governor-ivey-announces-new-first-class-pre-k-classrooms-in-first-round-of-funding/ microsoft word [formatted] 3 learning predictors from multidimensional data with tensor factorizations aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. learning predictors from multi dimensional data with tensor factorizations soo min kwon, anand d. sarwate (faculty advisor) ✵ abstract statistical machine learning algorithms often involve learning a linear relationship between dependent and independent variables. this relationship is modeled as a vector of numerical values, commonly referred to as weights or predictors. these weights allow us to make predictions, and the quality of these weights influence the accuracy of our predictions. however, when the dependent variable inherently possesses a more complex, multidimensional structure, it becomes increasingly difficult to model the relationship with a vector. in this paper, we address this issue by investigating machine learning classification algorithms with multidimensional (tensor) structure. by imposing tensor factorizations on the predictors, we can better model the relationship, as the predictors would take the form of the data in question. we empirically show that our approach works more efficiently than the traditional machine learning method when the data possesses both an exact and an approximate tensor structure. additionally, we show that estimating predictors with these factorizations also allow us to solve for fewer parameters, making computation more feasible for multidimensional data. 1 introduction machine learning classification algorithms are widely used in many applications such as fraud and spam detection. the objective of these algorithms is to model a linear relationship between the independent (e.g. card transactions, amount spent) and dependent (e.g. fraud or not fraud) variables. this relationship is generally modeled by learning a hyperplane that best separates the two classes of data as shown in figure 1. the hyperplane is constructed of weights and biases, which can simply be interpreted as the slope and intercept, respectively. one can solve or estimate these values by learning the parameters that most accurately describe the observed data points. in machine learning, we solve for these parameters (or predictors) through empirical risk minimization (erm). the erm framework tries to estimate the parameters that minimize the ``risk’’ or error of a loss function between the true and computed predictors given data. the minimization of this loss function measures the ``closeness’’ of the predictors, where a smaller objective function value would account for a more accurate model. there are many different machine learning classification algorithms, and each algorithm has a different loss function. however, since many loss functions try to model a linear relationship, there is an implicit need for our figure 1: visualization of learning a line (or hyperplane in higher dimensional space) that best separates two classes of data. machine learning algorithms estimate these weights, 𝑊, and bias, 𝑏, through empirical risk minimization. aresty rutgers undergraduate research journal, volume i, issue iii data to be vectorized. if our data samples were to be multidimensional, vectorization would make estimation of accurate predictors much more challenging. for example, consider a different scenario in which one would like to make movie recommendations for a user given the number of movies watched in a certain genre. this data can easily be stored in the form of a matrix, where the rows represent each user, and the columns represent the movie genre. however, what happens when a user’s movie preferences change over time? as shown in figure 2, we can capture this third variable (and many others) by modelling the observations in the form of a tensor, as it matches the structure of the data. clearly, the structure of this tensor is significant for accurate data analysis. if the orderings of the movies watched were swapped for two given users, incorrect recommendations could be made. vectorizing this data does not account for these types of structures, making inference much more challenging. there are many modern applications that involve analyzing data with intrinsically many more dimensions, including medical imaging,[3,13] image processing,[5,18] and seismic data analysis.[8] in most of these settings, the objective is similar to that of the traditional machine learning goal: to formulate a problem of prediction to establish an association between the tensor covariates (independent variable) and outcomes (dependent variable). however, as previously mentioned, most machine learning frameworks are formulated for vector spaces, making statistical inference challenging for tensor data. in addition, in most of these domains, the tensor data also exhibits high dimensions. for example, in medicine, tensor data samples may be of dimensions 128 × 128 × 128 or greater. naively turning this array into a vector for traditional machine learning would result in solving for 128 = 2,097,152 coefficients. in this scenario, vectorization not only destroys the structure of the data, but also makes computation infeasible. related works recently, work on tensor-based machine learning approaches uses tensor factorizations to reduce the number of coefficients to be estimated.[10,15,21] specifically, tensor decompositions are imposed on the coefficients as a scheme of feature selection or dimensionality reduction. integrating such decomposition structures solves for low rank approximations of the true predictor, rather than the vector counterpart. zhou et al. proposes a tensor regression model with additional independent variables for predicting continuous values given fmri data.[21] for parameter estimation, they propose a maximum likelihood (ml) approach using a block relaxation algorithm, which we adopt to formulate tensor classification models. tan et al. proposes a logistic tensor regression model with a norm regularization to induce sparsity.[15] we observe that this technique efficiently exploits structure, which motivates us to generalize and formulate more classification problems with different regularization (e.g. norm), and on different datasets. our contribution in this paper, we investigate the performance of machine learning classifiers with a candecomp/ parafac (cp) decomposition structure on the coefficients/predictors. we have seen in previous literature that these methods work efficiently for solving linear regression and logistic regression coefficients.[15,21] we solve classification problems, namely support vector machines and logistic regression figure 2: example of modeling observations: left – matrix, right – tensor aresty rutgers undergraduate research journal, volume i, issue iii on both synthetic and real data. the rest of this paper is organized as follows. we first discuss some tensor algebra and notation that will be used throughout this paper. then, we propose the objective functions as well as a short analysis of the cp structured machine learning problems. we motivate and show results of our approach by fixing and solving for the true predictor. we compare the results from the tensor structured algorithm as opposed to the unstructured vector algorithm. our contributions can be summarized as follows:  we perform experiments to show that our structured method works more efficiently than the traditional method when the true predictor exhibits both an approximate and exact low rank structure.  we show that our structured approach solves for fewer coefficients more efficiently than the traditional approach with a dimensionality reduction step (e.g. principal component analysis).  we develop algorithms to solve machine learning problems with decomposition of 𝑛 -dimensional tensors with an alternating minimization scheme. 2 preliminaries we dedicate this section to discuss some of the concepts used throughout this paper. due to the theoretical nature of this work, the technical description may require some mathematical maturity. the reader interested in the empirical findings can skip to section 4. for a complete introduction to tensors, see the comprehensive survey of kolda and bader and rabanser et al.[6,14] tensors are simply defined as multidimensional arrays, and these two terms will be used interchangeably. we will denote vectors with lower case letters (𝑥), matrices with capital letters (𝑋), and tensors as bold capital letters (𝐗). i. tensor reorderings let 𝐗 be a third-order tensor of dimensions 𝐗 ∈ ℝ × × with the two frontal slices defined by 𝑋 , 𝑋 ∈ ℝ × : vectorization we can create a vector from any matrix or tensor by stacking the row or column elements into a row or column vector, respectively. for example, vectorizing the tensor 𝐗 by its columns would yield the following column vector: where we stack the columns from the first frontal slice, 𝑋 and the second frontal slice, 𝑋 . the dimensions of the resulting vector would be 𝑥 ∈ ℝ . matricization the 𝑛-mode matricization (or unfolding) of a tensor 𝐘 ∈ ℝ × ×…× is denoted as 𝑌( ) , where 𝑌( ) has the columns of the 𝑛 -mode fibers. consider the same tensor 𝐗 from the previous example. then the three 𝑛-mode matricizations are the following: one can think of matricization as a generalization of vectorization but to matrices. since our example 𝐗 is a third-order tensor, we have three matrices from matricization, one for each mode. ii. vector & matrix products outer product let 𝑎 and 𝑏 be two vectors of dimensions 𝑎 ∈ ℝ and 𝑏 ∈ ℝ , aresty rutgers undergraduate research journal, volume i, issue iii the outer product of 𝑎 and 𝑏, denoted as 𝑎 ○ 𝑏, is a matrix of dimensions (𝑎 ○ 𝑏) ∈ ℝ × , note that this outer product is not only limited to vectors, and can be generalized to matrices and tensors as well. kronecker product let 𝐴 and 𝐵 be two matrices of dimensions 𝐴 ∈ ℝ × and 𝐵 ∈ ℝ × , the kronecker product of 𝐴 and 𝐵 , denoted as 𝐴 ⊗ 𝐵, is a matrix of dimensions (𝐴 ⊗ 𝐵) ∈ ℝ × , in essence, the kronecker product is computed by multiplying every element in the first matrix, 𝐴, by the entire second matrix, 𝐵. khatri-rao product the khatri–rao product is the columnwise kronecker product. consider two (different) matrices 𝐴 ∈ ℝ × and 𝐵 ∈ ℝ × . the khatri–rao product of 𝐴 and 𝐵, denoted as 𝐴 ⊙ 𝐵, is a matrix of dimensions (𝐴 ⊙ 𝐵) ∈ ℝ × , here, we are taking the kronecker product between every column vector from 𝐴 and 𝐵. note that if 𝐴 and 𝐵 itself were column vectors, i.e. 𝑛 = 1 , then the khatri–rao product is equivalent to the kronecker product, 𝐴 ⊙ 𝐵 = 𝐴 ⊗ 𝐵. iii. tensor decomposition tensor decompositions are generalizations of matrix factorizations to multidimensional arrays.[20] we introduce one tensor factorization scheme that is important in understanding the setting of our algorithm. in the matricized form, we show that this factorization has useful properties to be solved with an alternating minimization scheme. candecomp/parafac (cp) decomposition the objective of the cp decomposition is to express a tensor as the sum of component rank–one tensors, i.e. vectors, as depicted in figure 3. for example, consider a third-order tensor 𝐗 ∈ ℝ × × . we can approximate this tensor as the following where ”○” denotes the outer product, r represents the rank (positive integer), and 𝑎 ∈ ℝ , 𝑏 ∈ ℝ , and 𝑐 ∈ ℝ for 𝑟 = 1, . . . , 𝑅 . we can formalize this decomposition as the following optimization problem: where 𝐗 would represent a low rank approximation of 𝐗. the factor matrices or cp factors are matrices with the rank–one tensors as entries. from the previous three-dimensional case, 𝐴 ∈ ℝ × would be an estimated cp factor with entries figure 3: graphical representation of the candecomp/ parafac decomposition – low rank approximation of a third–order tensor aresty rutgers undergraduate research journal, volume i, issue iii with these definitions and the products defined previously, we can formulate some useful properties for the third–order case: these relationships can easily be generalized to 𝑛mode tensors, but for the purposes of this paper, 𝑛 = 3 will suffice. we will show how we can use these equations for our alternating minimization algorithm in the following sections. there also are other useful tensor factorizations, such as the tucker decomposition, which is explained in detail in the survey paper.[6] iv. machine learning optimization problems many machine learning algorithms can be framed as empirical risk minimization (erm) problems. the empirical risk is defined in terms of a risk, or loss function (·). for linear classifiers, the loss of a linear predictor 𝑤 on the data sample (𝑥 , 𝑦 ) can be written as (𝑤 𝑥 , 𝑦 ) and the average empirical risk as ∑ (𝑤 𝑥 , 𝑦 ). we discuss these loss functions for some common classifiers and how we can use them to solve tensor structured erm problems. support vector machines consider a dataset with 𝑛 samples, i.e. {(𝑥 , 𝑦 )} , where 𝑦 ∈ {−1,1}. support vector machine (svm) or maximum margin linear classifier is a binary classifier that finds a hyperplane to best separate the data, while making minimal margin violations.[4] svm uses a loss function called the hinge loss function, defined by where 𝑤 is the coefficients of the separating hyperplane. with a penalty (or regularizer), we can mathematically formulate svm as the following erm problem: the regularization term, λ, is used to penalize the features, and hence weights, that do not necessarily contribute to the prediction outcome. here, we are considering the penalty, but there are other regularizers such as the penalty. we use these regularization terms in our loss function to estimate a more accurate model. logistic regression similarly, consider a dataset with 𝑛 samples, i.e. {(𝑥 , 𝑦 )} , where 𝑦 ∈ {−1,1}. the objective of logistic regression (logit) is the same as svm, with a different loss function called the logistic loss function, defined by the logistic loss function takes the form of the sigmoid function. with a regularization term, we can define logistic regression as the following erm problem: we only introduce the objective function of these two classifiers, as we will construct the cp structured algorithm with these functions in the following section. note that we do not include the bias term in our hyperplane equation, as it can be modeled in 𝑤 as a column vector. 3 problem formulation in this section, we propose our tensor-based classifiers in the form of an erm framework. in general, we structure our linear predictors (𝑤 ) to admit a cp decomposition, in which we can reconstruct to make classifications. we also discuss the metrics that we will be investigating to evaluate the performance of our models. aresty rutgers undergraduate research journal, volume i, issue iii i. candecomp/parafac structured classifiers support vector machines consider a dataset {(𝐗𝒊, 𝑦 )} , where 𝐗𝒊 ∈ ℝ ×…× denotes a tensor data sample with 𝑦 ∈ {−1,1} . by imposing the constraints from (1) onto the predictors of (3), we can formulate the following optimization problem: the traditional erm problem for svm in (3) solves for one vector predictor of dimensions 𝑤 ∈ ℝ ×…× . the problem in (5) , which we call “cp-svm”, solves for 𝑁 matrix-valued predictors of dimensions 𝑊𝒊 ∈ ℝ × , for 𝑖 = 1, . . . , 𝑁 . as a concrete example, let each tensor sample be dimensions 𝐗𝒊 ∈ ℝ × × and 𝑅 = 3 . the traditional problem would solve for 5 × 5 × 5 = 125 coefficients, whereas the structured problem would solve for 3 × (5 × 3) = 45 coefficients. as the dimensions increase, the structured problem substantially reduces the number of parameters/coefficients to be estimated. logistic regression similarly, consider a dataset {(𝐗𝒊, 𝑦 )} , where 𝐗𝒊 ∈ ℝ ×…× denotes a tensor data sample with 𝑦 ∈ {−1,1} . by imposing the constraints from (1) onto the predictors of (4), we can solve the following erm problem: this new framework, which we call “cplogit”, solves for fewer parameters, similar to cpsvm. in practice, we solve for the weights using numerical optimization methods such as gradient descent. however, solving for the weights in this new cp-structured paradigm is a non-trivial task. in order to solve for the coefficients in (5) and (6), we adopt an alternating minimization algorithm similar to the block relaxation algorithm proposed in zhou et al.[21] at each iteration, we update block 𝑊𝒊, while keeping the rest of the blocks fixed. to see this, when updating 𝑊𝒊 ∈ ℝ × , we can rewrite the inner product in (5) and (6) with the properties mentioned in (2): this alternating minimization algorithm is summarized in algorithm 1, in which (·) represents the erm problem to be minimized, 𝜃 represents a collection of all the parameters, and λ is the regularization parameter. the parameter λ was tuned by hand, but can also be determined through cross validation. to understand the cp structured algorithm, consider the loss function in (5) with 𝑁 = 3 . when updating 𝑊𝟐 , we rewrite the inner product 〈∑ 𝑊 ( ) ○ 𝑊 ( ) ○ 𝑊 ( ) , 𝐗𝒊〉 as 〈𝑊 , 𝐗( )(𝑊 ⊙ 𝑊 )〉 . note that this equation follows from the property of tensor algebra as shown in (2). we perform this algorithm for all the factor matrices until the stopping criteria is met. aresty rutgers undergraduate research journal, volume i, issue iii the alternating minimization algorithm is useful for several reasons. first, in practice, this algorithm almost always converges to at least a local minimum.[1,20,21] to find the best solution, the algorithm can be ran several times with different initial factor matrices. second, the low rank optimization problem over the factor matrices is non-convex.[2] thus, this problem becomes difficult to solve using common unconstrained solvers, such as gradient descent. in literature, there are two ways to handle the non-convexity of this optimization problem. one way is to relax the rank constraint by adding a convex regularization term that induces low rank (e.g. trace norm, nuclear norm).[16,19] the other solution is to employ this alternating minimization algorithm, as the optimization over one matrix, while holding the others fixed is convex. we chose to explore this procedure following zhou et al.,[21] as the algorithm is straightforward to implement using statistical software such as matlab or python. ii. performance metrics we evaluate the performance of our models using several measures with different sample sizes. the following four metrics help determine the measure of “closeness” between the true and estimated predictors. 1. the mean squared error (mse) for 𝑛 data samples and true predictor 𝑊 is computed as where 𝑊 is the estimated predictor from solving the erm problem. 2. the cosine distance (or similarity)[12] for true predictor 𝑊 is computed as where 𝑊 is the reconstructed predictor from solving the erm problem. mathematically, the cosine similarity measures the cosine of the angle between two vectors projected in a 𝑛-dimensional space. as the angle, 𝜃 , between the two vectors becomes smaller, the cosine similarity will approach a value of 1. as the angles become farther apart (perpendicular), the cosine similarity will approach a value of 0. 3. the reconstruction error for true predictor 𝑊 and estimated tensor predictor 𝑊 is defined as where || · || denotes the frobenius norm, a matrix generalization of the norm. 4. the classification accuracy for 𝑛 test samples is simply defined as the following: after solving for 𝑊 , we make predictions on test data and compare 𝑦 to the true 𝑦 . before comparing the labels, we use the sign function to quantize our values to 𝑦 ∈ {−1,1}. 4 experiments we used two types of data for our experiments: synthetic data and the modified national institute of standards and technology (mnist) database.[9] the mnist database is a benchmark dataset used widely in machine learning that consists of 60,000 samples of handwritten digits from 0 to 9. the objective of both experiments is to compare the performance between the cp-structured algorithms and the traditional algorithms, which were implemented using software packages tensorly[7] and scipy.[17] for all experiments, we use a python environment on a macbook pro with 2.2 ghz intel core i7 and 16 gb ram. i. synthetic data for synthetic data, we generated univariate 𝑦 responses with different sample sizes according to the following model: where 𝑋 is drawn independently and identically distributed (iid) from 𝒩(0,1), 𝜖 is a noise term drawn iid from 𝒩(0,1), and 𝑊 is the fixed predictor as shown in figure 4. the objective was to observe if our models defined in (5) and (6) can identify the true signal 𝑊 given (𝑋 , 𝑦 ). aresty rutgers undergraduate research journal, volume i, issue iii performance comparison to measure the “closeness” and classification accuracy between the true model and the predicted model, we use performance metrics defined in (7), (8) , (9) , and (10) , we compute these metrics at different sample sizes and show that as the number of samples increases, the performance of the traditional vector approach converges to the performance of the cp structured model. these results are displayed in figures 5 and 6. in figure 5, we can visually see that the predictors from our method solves for the true predictors more accurately. for example, in the case of 𝑛 = 500 from row 1, the “cross” figure is more accurately portrayed using the cp method (right) than the traditional method (middle). this would allow us to make more accurate predictions, as the estimated weights more closely follow the true weights. in figure 6, we can see that the mse for both algorithms is relatively the same throughout all sample sizes. for the cosine distance, we can see that the cp structured algorithm approaches a value of 1 very quickly, which indicates that there is a strong similarity between the estimated and the true coefficients. the reconstruction error and classification accuracy both generally have gaps in the figures, but lessen as the sample sizes increase. we can conclude that these results depend on the sample size, as more samples can decrease the number of hyperplanes that separates the data, predicting coefficients closer to the true model. based on the trends of the graphs in figure 6, we also hypothesize that if the variance of the noise (𝜖) distribution was higher, the cp structured algorithms would also perform better than the traditional method. results with pca the cp structured algorithm significantly reduces the number of predictors to be estimated. to solve for less coefficients using the traditional method, we can perform principal component analysis (pca) on the dataset before using the algorithm. we use pca on 𝑋 with an energy capture of 95%, which reduces the number of coefficients from 225 to 189. however, even with this minimal reduction, we can see in figure 6 that there is a notable decrease in performance throughout most metrics. the mse seems unaffected, but the other three metrics start to see a gap between the traditional method with no pca and the cp structured algorithm. a possible explanation for this phenomenon is that pca does not capture tensor data efficiently in lower dimensional space. if we were to decrease the energy capture, the gap in performance would grow larger even for a bigger sample size. we predict that as the dimensions of the data increases, pca would not be an efficient feature learning method for parameter reduction, favoring the cp structured methods. figure 4: two 15 × 15 images used as true predictors 𝑊 to generate synthetic data figure 5: reconstructed predictors from both algorithms: left – true predictor, middle – reconstructed predictor from traditional method with increasing sample sizes (𝑛 = 500,1000,1500), right – reconstructed predictor from cp-structured logistic regression with increasing sample sizes (𝑛 = 500,1000,1500) aresty rutgers undergraduate research journal, volume i, issue iii ii. mnist data the objective of the mnist dataset experiment was to observe which algorithm would be more efficient to use when the true predictor exhibited an “approximate” low rank structure. in the previous experiment, the two images used as the true predictor had an exact low rank structure, as it could easily be computed through an outer product of two matrices. similar to the synthetic data setup, we generated univariate 𝑦 responses with sample size 𝑛 = 750 with the model defined in (11). however, for the true predictor, 𝑊, we chose a “1” from the mnist dataset, as it exhibits “approximate” low rank structure. we compared the cp-structued algorithms to the traditional algorithms using different rank values. these results are shown in table 1. performance comparison we use the same performance metrics defined for the previous experiment and display the results in table 1. from this table, we can conclude that both cp structured algorithms gave favorable results when the cp rank was 2. this shows that we can approximate a ”1” from the mnist dataset with matrices of rank 2. however in all cases from rank 1 figure 6: variation of performance (y-axis) with different sample sizes (x-axis) for svm and logit. columns 1-4 represent plots for the mse, cosine distance, reconstruction error, and classification accuracy, respectively. row 1, 2: performance metrics for logit with predictors as cross and square, respectively. row 3, 4: performance metrics for svm with predictors as cross and square, respectively. predictors of cross and square is as shown in figure 4 aresty rutgers undergraduate research journal, volume i, issue iii to 3, the structured algorithms gave more favorable results. this proves to show that if the true predictor exhibits an approximate low rank structure, it may be beneficial to use the structured algorithms for classification. 5 conclusion in this paper, we explored tensor-based classification models using a tensor decomposition. we proposed two algorithms that imposed a candecomp/parafac factorization structure on the predictors of traditional classification algorithms: support vector machines and logistic regression. imposing these techniques on traditional algorithms allowed us to exploit the structure of the data, enabling efficient learning with fewer parameters. we showed with different performance metrics that our proposed method increased accuracy and overall solved a more accurate estimation of the weights. the experiments showed that the cp algorithm performed best when the true predictor had either an approximate or exact low rank structure. we also showed that solving for fewer parameters using pca compromised the performance of the traditional method. we predict that pca would not generalize well to data with multidimensional structure, favoring the cp structured algorithms. however, we believe that it would be interesting if one could show when pca could be better than using cp structure. this could possibly be a case when the data in question is known to be linear, as pca is a linear feature learning method. one potential example is using structured data for prediction when it is known a priori that the features have a linear relationship. however, due to time constraints, we were not able to explore this possibility in detail. we also think it would be interesting to test these algorithms on more datasets. in addition, we believe an exciting direction for future research is to exploit tensor decompositions in other learning problems such as deep learning. however, it is not clear how one would approach this problem, as deep learning algorithms have nonconvex loss functions. we leave this up to the audience to investigate for future exploration.∎ method mse cos distance reconstruction error svm 0.00128 0.51832 0.00053 cp-svm (r=1) 0.00088 0.66727 0.00044 cp-svm (r=2) 0.00026 0.90259 0.00024 cp-svm (r=3) 0.00039 0.85099 0.00029 logit 0.00120 0.54759 0.00051 cp-logit (r=1) 0.00089 0.66483 0.00044 cp-logit (r=2) 0.00028 0.89590 0.00024 cp-logit (r=3) 0.00033 0.87807 0.00026 table 1: performance metrics between the traditional and structured algorithms for the mnist dataset experiment. the bolded values represent the “best” performance throughout each method, where 𝑅 represents the rank of the cp structured algorithm for each experiment. aresty rutgers undergraduate research journal, volume i, issue iii 6 references [1] bezdek, j. & hathaway, r. (2003). convergence of alternating optimization. neural, parallel & scientific computations. 11. 351-368. [2] candès, e. j., & recht, b. (2009). exact matrix completion via convex optimization. foundations of computational mathematics, 9(6), 717–772. [3] de luis-garcía, r., westin, c.-f., & alberola-lópez, c. (2010). gaussian mixtures on tensor fields for segmentation: applications to medical imaging. computerized medical imaging and graphics, 35(1), 16–30. [4] james, g., witten, d., hastie, t., & tibshirani, r. (2013). an introduction to statistical learning with applications in r (1st ed. 2013.). springer new york. [5] jia, c., kong, y., ding, z., & fu, y. (2014). latent tensor transfer learning for rgb-d action recognition. proceedings of the 22nd acm international conference on multimedia, 87–96. [6] kolda, t. g., & bader, b. w. (2009). tensor decompositions and applications. siam review, 51(3), 455–500. [7] kossaifi, j., panagakis, y., anandkumar, a., & pantic, m. (2019). tensorly: tensor learning in python. journal of machine learning research, 20. [8] kreimer, n., stanton, a., & sacchi, m. d. (2013). tensor completion based on nuclear norm minimization for 5d seismic data reconstruction. geophysics, 78(6), v273–v284. [9] lecun, y., bottou, l., bengio, y., & haffner, p. (1998). gradient-based learning applied to document recognition. proceedings of the ieee, 86(11), 2278–2324. [10] li, x., xu, d., zhou, h., & li, l. (2018). tucker tensor regression and neuroimaging analysis. statistics in biosciences, 10(3), 520–545. [11] li, y., zhu, h., shen, d., lin, w., gilmore, j. h., & ibrahim, j. g. (2011). multiscale adaptive regression models for neuroimaging data: multiscale adaptive regression models. journal of the royal statistical society. series b, statistical methodology, 73(4), 559–578. [12] nguyen, h. v., & bai, l. (2011). cosine similarity metric learning for face verification. in computer vision – accv 2010 (vol. 6493, issue 2, pp. 709–720). springer berlin heidelberg. [13] o’donnell, l. j., & westin, c.-f. (2011). an introduction to diffusion tensor image analysis. neurosurgery clinics of north america, 22(2), 185–196. [14] rabanser, s., shchur, o., & günnemann, s. (2017). introduction to tensor decompositions and their applications in machine learning. [15] tan, x., zhang, y., tang, s., shao, j., wu, f., & zhuang, y. (2013). logistic tensor regression for classification. in intelligent science and intelligent data engineering (vol. 7751, pp. 573–581). springer berlin heidelberg. [16] tomioka, r., & suzuki, t. (2013). convex tensor decomposition via structured schatten norm regularization. [17] virtanen, p., gommers, r., oliphant, t. e., haberland, m., reddy, t., cournapeau, d., burovski, e., peterson, p., weckesser, w., bright, j., van der walt, s. j., brett, m., wilson, j., millman, k. j., mayorov, n., nelson, a. r. j., jones, e., kern, r., larson, e., … moore, e. w. (2020). scipy 1.0: fundamental algorithms for scientific computing in python. nature methods, 17(3), 261–272. [18] weiwei guo, kotsia, i., & patras, i. (2012). tensor learning for regression. ieee transactions on image processing, 21(2), 816–827. [19] wimalawarne, k., tomioka, r., & sugiyama, m. (2016). theoretical and experimental analyses of tensor-based regression and classification. neural computation, 28(4), 686–715. [20] wright, s. j. (2015). coordinate descent algorithms. mathematical programming, 151(1), 3–34. [21] zhou, h., li, l., & zhu, h. (2013). tensor regression with applications in neuroimaging data analysis. journal of the american statistical association, 108(502), 540–552. soo min kwon is currently pursuing a m.s. degree in the department of electrical and computer engineering at rutgers, the state university of new jersey. his research interests broadly lie in optimization, multidimensional (tensor) data analysis, differential privacy, and distributed learning. he earned his b.s. degree in electrical and computing engineering from rutgers university in 2020. he completed his undergraduate thesis under supervision of prof. anand d. sarwate on tensor-based machine learning algorithms. soo min hopes to pursue a phd degree upon completion of his m.s. degree. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. an assessment of the food and physical activity environment on a university campus kayla castellitto and nurgul fitzerald kayla catellitto conducted this study as a part of her george h. cook honors thesis research, while she was an undergraduate student at rutgers, the state university of new jersey, school of environmental and biological sciences. dr. nurgul fitzerald is an associate professor / extension specialist at the department of nutritional sciences at rutgers university. abstract large percentages of college students are reported to be overweight and sedentary and do not consume the recommended amounts of fruits and vegetables. these outcomes can be influenced by the students’ environment. the purpose of this study was to determine the level of healthfulness and environmental supports on rutgers university’s cook campus (ru)—one of the five rutgers campuses—by examining campus food and physical activity environment, and related policies. as a part of the nationwide get fruved study on over 90 college/university campuses, the healthy campus environmental audit (hcea) instrument was used to assess dining establishments, vending machines, recreational environment, and policies at ru. ru scores were compared to the original scales and to the average of the other get fruved universities/colleges. ru’s healthfulness scores for dining halls/cafeterias, recreational environment, vending machine supports, and stimulants policy were on the higher end of the scales and above the averages of other get fruved schools. however, ru’s scores indicated limited healthfulness in fast-food/sit-down restaurants; walking/biking supports; availability of healthy snacks and beverages in vending machines; healthy eating policies; and policies encouraging physical activity and chronic disease prevention. this study identified the strengths and weaknesses in ru’s campus environment and in ru’s policies for healthy eating and active living. these results can be used to support a healthier campus environment. key terms: active environment, audit, built environment, environmental assessment, food environment, health/wellness promotion, healthy eating, nutrient density, policy, university/college campus, vending machines introduction emerging adulthood (18 to 25 years of age) is considered an important stage for maintaining long-term health behavior patterns (nelson, story, larson, neumark-sztainer, & lytle, 2008). yet, the lifestyle changes college students experience can make establishing healthy long-term behaviors difficult and put them at risk of gaining weight. these changes may include moving https://creativecommons.org/licenses/by-nc-sa/4.0/ https://creativecommons.org/licenses/by-nc-sa/4.0/ aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 2 out of the family home; changes in financial responsibilities; unhealthy food options on campus; and unhealthy food habits like late-night, alcohol-related, or stress-related eating (nelson, kocos, lytle, & perry, 2009). additionally, large proportions of college students experience stress. in a national survey of college students, 45% reported “more than average” level of stress, and 13% reported high levels of stress (american college health association, 2018). chronic stress may be linked to weight gain (torres & nowson, 2007). gaining weight during college years can have long-term effects: being moderately overweight in early adulthood and gaining weight in adulthood have both been shown to increase risks of developing major chronic diseases such as cardiovascular diseases, diabetes, and certain cancers (de mutsert, 2014). therefore, it is important that college students have easy access to healthy food options and physical activities on campus. previous research has shown that many college students are at high risk of being overweight or obese. according to the american college health association ([acha], 2018) national college health assessment, which surveyed 88,178 students, large proportions of the participants were overweight (23.8%) or obese (16.2%). in a study of 67 first-year college students, it was reported that almost 75% of the students had gained weight during their first year in college. over the seven-month study period, their average weight gain was 3.1 kilograms (7 pounds), and average percentage of body fat gained was 0.9% (hoffman, policastro, quick, & lee, 2006). the exact causes of such weight gain were uncertain, but it was speculated to be because of decreased physical activity (butler, black, blue, & gretebeck, 2004) and lifestyle changes like decreased leisure time and increased alcohol consumption (hoffman et al., 2006). the acha assessment also revealed low levels of fruit and vegetable intake among college students. only 4.8% reported eating the recommended levels of five or more servings of fruits and vegetables a day. college campuses can affect food and beverage intake patterns through dining halls, restaurants, convenience stores, markets, food courts, and vending machines. because the environment plays a role in what students consume (stokols, 1992), it is important to examine the types of foods these venues provide. an assessment of dining establishments on 15 college campuses across the united states indicated that campus restaurants offered healthier food and beverage options in comparison to off-campus restaurants; still, they offered large portion sizes and “combo” meals that encouraged overeating (horacek et al., 2012). campus food environment is also likely to influence students who commute to campus. a study of 1,059 off-campus students in two minnesota colleges found that 45% of the students reported purchasing foods or beverages from at least one campus venue more than three times per week. students who frequently purchased foods from these venues reported buying foods that were high in fat and added sugar; they also reported eating breakfast less frequently and purchasing more fast-food on campus. students who brought their own lunch from home had healthier dietary intake patterns (pelletier & laska, 2013). vending machines are a convenient way for students to obtain snacks throughout their school day. in an assessment of 2607 vending machines at 11 colleges, it was found that the most aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 3 commonly available snacks were chips, pretzels, other salty snacks, and sweet treats like candy (byrd-bredbenner et al., 2012). most of these snacks were low in fiber and high in calories and fat; about half of the snacks were high in sugar. a majority of the available beverages were high in calories and sugar (byrd-bredbenner et al., 2012). these studies suggest that food and beverages available in vending machines limit healthy choices and promote low nutrient, energy-dense snacks as the easy choice. physical activity, a key component of preventing unwanted weight gain, can be impacted by the availability of recreation programs and facilities on campus. the acha (2018) assessment reported low levels of physical activity among college students. more than half (53%) of the students did not meet the physical activity recommendations of moderate-intensity (at least 30 minutes on five or more days/week) or vigorous-intensity exercises (at least 20 minutes on three or more days/week). in a study completed at 13 colleges/universities, researchers concluded that while the programs, classes, and equipment available at the recreation facilities supported healthy lifestyles, the facilities’ policies (equipment accessibility, safety features, weather accommodations, transportation, and fees) and the built environment (bike rack availability and accessibility, presence of health promotion signs, and accessibility and cleanliness of stairwells) did not support healthy lifestyles (horacek et al., 2014). there were very few recreational screening policies for at-risk students and promotional signs for daily activities like taking the stairs. exercise classes were not offered on weekends when students had most of their spare time. exercise spaces were rated insufficient for use during peak hours, and secondary facilities (smaller versions of primary recreation facilities) lacked the quality of primary facilities. researchers also reported good walkability but low bikeability across the campuses. in summary, the literature indicates that college is a period marked by changes, which can put students at risk of becoming overweight or obese. college campus environments can influence students’ eating and physical activity patterns through eating establishments, vending machines, recreation centers, and policies related to these facilities. the objective of this study was to comprehensively evaluate the healthfulness of the food and physical activity environment on a university campus through environmental and policy assessments. these assessments could serve as a basis for identifying policy and environmental changes that promote health on campus. methodology study location and setting this study was conducted at the cook campus of rutgers, the state university of new jersey in new brunswick, nj during the 2017-2018 school year. it was a part of the get fruved study, which included more than 90 universities and colleges across the united states. the original get fruved study was funded by united states department of agriculture (award number 2014aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 4 67001-2185) and administered by university of tennessee (ut) to promote healthier lifestyles on college campuses (get fruved, 2017). the current study location was limited to only one (cook campus) of rutgers’ five campuses. for the environmental audits, the campus area was defined as the space within a 1.5 mile radius of the cook campus center. this was chosen to be the center of the campus because it is a place where many students gather, and it is located near student life facilities (multiple residence halls, dining hall, recreation center, laundry and postal services, and bus stops). the website “draw radius circles on a map” was used to draw the radius from the campus center using google maps (draw radius circles on a map, 2017). assessment methods the healthy campus environmental audit (hcea) was used to help determine aspects of the campus environment that might affect students’ health behaviors. the hcea is composed of five audits focusing on campus environmental demographics (ca), dining environment (fresh audit), recreation facilities (paces audit), policies (points audit), and vending machines (vending audit) (see supplemental materials for details). these audits have been validated in studies completed on various campus environments (horacek et al., 2019a, 2019b, 2019c). the audits were conducted by teams of two student researchers. the student researchers underwent audit-specific online training prior to collecting data. this training included reading the audit-specific protocol, completing the training videos, and passing quizzes with an 80% or higher score. collected data were directly entered into ut’s online survey portal, de-identified by the get fruved administrative team at ut, and sent back to rutgers for use. the study was approved by the institutional review boards of ut and rutgers. statistical analyses the environmental audit data were received from ut in a de-identified format. results from each audit included rutgers’ scores as well as descriptive statistics—minimum, maximum, mean scores and standard deviations—from the national sample (that is, all of the participating get fruved schools that completed the same audits). for the interpretation of rutgers’ data, the results were compared to both the maximum possible audit scores and the national average data. the hcea scores do not currently include specific cutoff points like low versus high or failure versus success. therefore, the interpretations of the scores are based on the comparisons to the national average and the maximum possible score within each scale. results campus environmental demographic audit more than half of the campus population (n=3,428) in this study were female (62.1%; 37.9% male). slightly more than a half of them were reported to be white (51.1%), 14.1% were aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 5 hispanic, 7.0% were black, and 27.8% were listed as other. a small proportion (10.2%) of the students were out-of-state residents. in terms of food access in the city of new brunswick, there were 0.1-0.2 grocery stores per 1,000 population and 0.8-1.0 restaurants per 1,000 population. fresh audit food score, measured on a scale of 60, indicates the variety of the available foods. rutgers received a food score of 20.8 (34%) for restaurants and 45.3 (76%) for dining halls/cafeterias (figure 1). restaurants received lower scores because they had more energy-dense food options, less whole grains, and fewer lean meat options. dining halls, meanwhile, had healthier options such as a salad bar, low-fat desserts, healthier beverages, and lower-fat entrées/sides. rutgers’ food score for restaurants was similar to the national average (20.3), but the food score for dining halls/cafeterias was higher than the national mean score of 35.9. support score, measured on a scale of 40, includes the environmental supports to make healthy food choices. rutgers received a support score of 19.1 (48%) for restaurants and 25.3 (63%) for dining halls/cafeterias. restaurants scored lower because their healthy supports were limited; they charged extra for healthy substitutions, charged higher prices for healthier foods, and did not offer nutrition information at some locations. additionally, some restaurants were found to have practices encouraging overeating. these included larger portion sizes for entrees and beverages; and signs, deals, or promotions for less healthy options. similarly, audited dining halls included all-you-can-eat buffets, which encourages overeating. however, dining halls also offered supportive strategies for healthy eating. they priced healthier and less healthy options similarly. further, the dining halls implemented trayless dining service, where students take their food using one plate at a time rather than fitting multiple items on a tray. this service is designed to promote sustainability and discourage overeating. in addition, dining halls offered smaller cup sizes for beverages, which can help discourage overconsumption. rutgers’ support scores for restaurants were similar to the national mean, while rutgers’ support scores for dining halls/cafeterias were slightly higher than the national average. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 6 paces audit the facility support score indicates a facility’s accessibility and the quality of its staff. rutgers’ score was 9.0 (60%), which was higher than the national sample’s average score of 6.4 (figure 2). recreation facility staff’s availability and competence contributed to the higher scores. the equipment score, measuring the quality of exercise equipment itself, was 16.0 (80%) for rutgers, compared to the national sample average of 12.4. availability of a wide variety of equipment, including aerobic and strength training equipment, contributed to this higher score. for its walk/bike score—a measure of facility support for walking and biking—rutgers received an 8.0 (53%), whereas the national average was 7.0. the audit detected that rutgers recreation facility only had one bike rack available for use, which had about 41-60% of its spots open. the facility total score, a sum of all the subscores, was 33.0 (66%) for rutgers and 25.8 for the national sample. the campus score, which included the quality and extensiveness of health and fitness programs, was 46 (66%) for rutgers in comparison to the national sample’s mean of 45.0. among the contributing factors for higher rutgers campus score were the wide variety of indoor and outdoor amenities (e.g. pool, basketball courts, roller hockey rink, racquetball court), several intramural sports, health/wellness activities and fitness classes. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 7 points audit rutgers received a score of 100% for support to limit drugs and alcohol (stimulants), compared to the national sample mean of 84% (figure 3). rutgers received a 0% for both chronic disease prevention support and active environment support, while the national sample averages were 7.0% and 14.9%, respectively. the audit did not detect any chronic disease prevention and active environment policies, such as chronic disease online education or physical activity during work hours. rutgers’ score for supporting healthy eating was 34%, higher than the national sample score. the overall score for student support (in relation to employees) was also 34%, ranking rutgers slightly above the national sample. this audit also evaluated other policies that were not detected at the audit sites, such as minimum and maximum nutrition standards, healthy food labeling programs, and food taxes/subsidies. rutgers’ score for the comprehensiveness of policy—that is, how well policies are monitored and enforced—was 21% in comparison to 19% for the national sample. this suggests some, but not all, policies were enforced. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 8 vending audit of the items in the audited snack machines at rutgers, 38% were unhealthy, 34% were somewhat healthy, and 28% were healthy (figure 4). the total snack nutrient density score was 3.3 (out of 7), which was similar to the national sample mean of 3.02. snacks that were considered unhealthy included candies, chocolates, regular chips, and cookies. healthy snacks included baked/kettle chips, fruit snacks, granola bars, and unsalted nuts. of the items in the audited beverage machines at rutgers, 55% were unhealthy, 22% were somewhat healthy, and 23% were healthy. rutgers’ total beverage nutrient density score was 1.2 (out of 4), with the national sample mean being 1.09. beverages such as regular sodas, iced tea, lemonade, and energy drinks were considered unhealthy. water (flavored and unflavored) was considered healthy. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 9 the vending supports score for rutgers was 72%, which was higher than the national sample’s score. availability of vending machines and the fact that many of the machines contained specific nutrition information for each product weighed positively on rutgers’ score. on the other hand, healthier products were often more expensive than unhealthy products, which weighed negatively on the score. discussion the purpose of this study was to determine the healthfulness of the food and physical activity environment on rutgers university’s cook campus. compared to the audit scales, cook campus scores showed limited healthfulness in some categories. these included food quality and environmental support at fast-food and sit-down restaurants; healthy snack and beverage availability in vending machines; walking and biking supports; and policy supports (with the exception of policies about stimulants). however, rutgers scored higher than the national sample average for dining hall and cafeteria foods and supports; recreation facility and fitness programs; policies about stimulants and healthy eating; and supportive messaging in vending machines. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 10 the ca audit illustrated the diversity of rutgers’ cook campus. a diverse background among students can play a role in their food and beverage intakes, levels of physical activity, and health outcomes (u.s. department of agriculture & u.s. department of health and human services, 2015). students can have different food preferences, attitudes about their weight, and beliefs about physical activity. in addition, the campus environment is likely to influence the students’ choices, social norms, and values by controlling the availability of food, beverage and physical activity options. the ca audit found that the city of new brunswick had a high number of restaurants per capita when compared to the audit scale, and that many of these restaurants were outside the campus but within the 1.5 mile radius from the campus’ student center. this study analyzed the healthfulness of 22 (close to 30%) of these restaurants in order to determine how they may influence the overall healthfulness of the food environment for the students. thus, it is important to look at the campus demographics and environment to examine the needs that can be addressed in future health promotion interventions. the fresh audit data indicated that rutgers’ dining halls and cafeterias showed a higher level of healthfulness (76% on the food scale and 63% on the support scale) than the fast-food and sit-down restaurants (34% on the food scale and 48% on support scale). the dining halls and cafeterias had significantly higher scores compared to the national sample, while the restaurants’ scores were similar to the national average. dining halls and cafeterias offered healthful food options and provided environmental supports like signage (e.g. vegan, meatless monday), appropriate portion sizes, and nutrition information. still, there appears to be room for improvement in supporting healthier options. placing consistent and attractive signage next to the healthier meal options could make it easier for students to make better food choices; this strategy has been successfully implemented by some of the other schools in the get fruved study. restaurants’ lower healthfulness scores were consistent with previously published research that found restaurants offered large portion sizes and combo meals, thus encouraging overeating (horacek et al., 2012). other get fruved schools marketed healthy foods and costeffective meal deals at point of selection. additionally, they provided tastings of healthy food choices to tempt patrons to buy these products. such strategies could improve the food environment at rutgers as well. the paces audit data indicated that rutgers’ recreational facilities moderately supported healthfulness. the facilities’ staff, equipment, and campus recreation programs showed 60% to 80% healthfulness and environmental support. the facility support and equipment scores were above the national averages; these results concur with previous data that showed high ratings for overall facilities and equipment available to students (horacek et al., 2014). rutgers’ scores can be further improved by adopting strategies from other get fruved schools. for instance, rutgers could increase the number and variety of fitness classes, provide free fitness assessments to encourage the use of programs and facilities, and offer “late night at the gym” (open between 10pm and 2am) as an alternative to partying. in addition, rutgers’ walk/bike features, like bike racks and stairwells, showed room for improvements: its score was 53%, which was similar to the national sample’s score. these findings are somewhat consistent with aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 11 published research, which found very few policies to help prevent obesity and very few promotional signs for daily activities like taking the stairs (horacek et al., 2014). the points audit indicated that rutgers’ drug and alcohol policy was excellent (100% score) and was above the national average. however, this study did not find any policies for chronic disease prevention and active environment, and these results were below the get fruved schools’ averages. other get fruved schools with higher scores have implemented strategies such as advocating for a walkable campus and banning cars from campus. policies for encouraging healthy eating (score of 34%), monitoring/enforcement (21%), and the average student policy scores (34%) were on the lower end of the scales for rutgers, although the healthy eating policy score was above the national average. some of the other get fruved schools have set nutrition guidelines to standardize healthier dining options and have pledged to use local and community-based sources. the vending audit indicated that the snack and beverage machines have limited healthful options but promising environmental supports (e.g., good product pricing, nutrition information). the mean snack and beverage nutrient density scores were only in the “somewhat healthy” category, but the environmental support score was 72%. the snack and beverage nutrient density scores were generally similar to the scores from the national sample. more than half of the beverages in the audited vending machines were considered unhealthy, thus limiting students’ healthy choices. this is consistent with previous research that found majority of the beverages in vending machines were sugary, calorie-dense beverages that limits students’ healthy choices (byrd-bredbenner et al., 2012). to further improve these scores, rutgers could post signage highlighting healthier choices. for example, one of the get fruved schools has been using “healthy-choice selection buttons” for snacks with less than 200 calories, less than 10 percent of calories from saturated fat, less than 60 mg cholesterol, no trans fat, and less than 230 mg sodium. this study presents the results of the comprehensive campus environmental assessments that were not previously published. the study methods, developed by a collaborative team from eight universities, have been validated through testing in multiple university campuses across the country. the original collaborative team collected and managed data from the participating universities and colleges centrally, allowing for comparisons between local data and the national sample’s averages. however, because national data were not collected by the rutgers research team, the authors were unable to conduct further statistical analyses to test whether the differences between rutgers’ and other universities’ scores were statistically significant. therefore, the other universities’ scores were only used as a relative benchmark to narratively compare rutgers’ scores. another limitation of this study was its restriction to only one of rutgers’ five campuses (cook campus) due to time and budget constraints. therefore, this study is not a representative sample of other rutgers campuses or rutgers university as a whole. in the future, including other rutgers campuses could allow researchers to perform statistical analyses on the scores from different campuses. this would help determine the aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 12 differences between various locations on campus. in addition, the beverage machine audits were limited to a small sample size because many of the machines had covered front panels and therefore, accurate data could not be collected from all. lastly, the points audit did not assess university policies about facilitating stress relief, even though stress may have an impact on weight gain. in summary, rutgers’ healthfulness scores for dining halls and cafeterias, recreational environment (facilities, equipment, supports, and campus programs), and the vending machine supports were on the higher end of the environmental audit scale; the drug and alcohol policy scored particularly well. these scores were also above the averages of the national sample of get fruved schools. however, rutgers scores were lower in categories like fast-food and sitdown restaurants; healthy snacks and beverages available in vending machines; walking and biking supports; and the policies to support healthy eating. this study also detected weaknesses in the policies for encouraging physical activity and preventing chronic diseases: these scores were lower compared to the national sample. these audits comprehensively determined areas of strengths and weaknesses that can help direct further efforts toward a healthier food and physical activity environment on campus. these audit scales are not meant to classify campus environments into good versus bad or success versus failure. the audit scores provide a baseline for each school as well as a benchmark average score from a wide variety of schools across the united states. it is possible that additional benchmarks can be developed by gathering the best practices from all participating schools. these results can be good starting points for students, campus organizations, and administrators hoping to implement environmental and policy changes and seeking to explore additional strategies for supporting healthy lifestyles on campus. acknowledgements this material is based upon work that was supported by the national institute of food and agriculture, u.s. department of agriculture, under award number 2014-67001-21851 (get fruved study pi: sarah colby, university of tennessee). any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the u.s. department of agriculture. we would like to thank the original get fruved study team, and especially the administrative team at the university of tennessee and hcea group at syracuse university for their guidance; haifa matos (rutgers institutional research and academic planning), dave donlon (rutgers dinning services), nathan johnson, brain cousin (campus residence departments), and dr. judith storch (cook campus deans office) for their assistance during the environmental audits; kappa omicron nu honor society members radwa nassar, yulissa pereira, nada ismail, jamie schroeder, nadine glowzenski, and kim manna for their assistance during the audits and reporting, and dr. worobey for his review of k. castellitto’s research thesis, which is the basis of this paper. partial aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 13 results of this research have been presented as a poster at the new jersey academy of nutrition and dietetics annual meeting in may 2018. references american college health association. (2018). national college health assessment ii: reference group executive summary spring 2018. handover, md. retrieved from https://www.acha.org/documents/ncha/nchaii_spring_2018_reference_group_executive_summary.pdf. butler, s.m., black, d.r., blue, c.l., & gretebeck, r.j. (2004). change in diet, physical activity, and body weight in female college freshman. american journal of health behavior, 28, 24-32. byrd-bredbenner, c., johnson, m., quick, v.m., walsh, j., greene, g.w., hoerr, s., … horacek, t.m. (2012). sweet and salty. an assessment of the snacks and beverages sold in vending machines on us post-secondary institution campuses. appetite, 58, 11431151. country health rankings & roadmaps (2017). key finding report. retrieved from http://www.countyhealthrankings.org/ de mustert, r., sun, q., willett, w.c., hu, f.b., van dam, r.m. (2014). overweight in early adulthood, adult weight change, and risk of type 2 diabetes,card iovascular diseases, and certain cancers in men: a cohort study. american journal of epidemiology, 179 (11): 1353-1365. draw radius circles on map. (2017). retrieved from http://obeattie.github.io/gmapsradius/?lat=51.500358&lng=-0.125506&z=10&u=mi&r=5. get fruved. (2017). fruved = more than fruits and vegetables. retrieved from http://fruved.com/what-2/. hoffman, d.j., policastro, p., quick, v., & lee, s.k. (2006). changes in body weight and fat mass of men and women in the first year of college: a study of the “freshman 15”. journal of american college health, 55, 41-45. horacek, t.m., erdman, m.b., byrd-bredbenner, c., carey, g., colby, s.m., greene, … white, a.b. (2012). assessment of the dining environment on and near the campuses of fifteen post-secondary institutions. public health nutrition, 16, 1186-1196. horacek t.m., simon m., dede yildirim e., white a.a., shelnutt k.p., riggsbee k.,…brown o. (2019a). development and validation of the policies, opportunities, initiatives and notable topics (points) audit for campuses and worksites. international journal of environmental research and public health, 16, 778. horacek, t.m., white, a.b, byrd-bredbenner, c., reznar, m.m., olfert, m., morrell, j.s., …thompson-synder, c.a. (2014). paces: a physical activity campus environmental supports audit on university campuses. american journal of health promotion, 28, 105117. https://www.ncbi.nlm.nih.gov/pubmed/?term=horacek%20tm%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/?term=simon%20m%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/?term=dede%20yildirim%20e%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/?term=white%20aa%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/?term=shelnutt%20kp%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/?term=riggsbee%20k%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/?term=brown%20o%5bauthor%5d&cauthor=true&cauthor_uid=30836633 https://www.ncbi.nlm.nih.gov/pubmed/30836633 https://www.ncbi.nlm.nih.gov/pubmed/30836633 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 14 horacek t.m., yildirim e.d., matthews schreiber m., byrd-bredbenner c., colby s., white a.a., … kattelmann k. (2019b). development and validation of the vending evaluation for nutrient-density (vend)ing audit. int j environ res public health, 16, 514. horacek t.m., yildirim e.d., seidman d., byrd-bredbenner c., colby s., white a.a., … kattelmann k. (2019c). redesign, field-testing, and validation of the physical activity campus environmental supports (paces) audit. journal of environmental and public health, 2019, 1-13. nelson, m.c., & kocos, r., lytle, l.a., perry, c.l. (2009). understanding the perceived determinants of weight-related behaviors in late adolescence: a qualitative analysis among college youth. journal of nutritional education and behavior, 41, 287-292. nelson, m.c., story, m., larson, n.i., neumark-sztainer, d., & lytle, l.a. (2008). emerging adulthood and college-aged youth: an overlooked age for weight-related behavior change. obesity journal, 16, 2205-2211. pelletier, j.e., & laska, m.n. (2013). campus food and beverage purchases are associated with indicators of diet quality in college students living off campus. american journal of health promotion, 28, 80-87. rutgers, the state university of new jersey. (2017). facts & figures. rutgers university-new brunswick. retrieved from http://newbrunswick.rutgers.edu/about/facts-figures. stokols, d. (1992). establishing and maintaining healthy environments toward a social ecology of health promotion. american psychologist, 47, 6-22. torres, s.j., & nowson, c.a. (2007). relationship between stress, eating behavior, and obesity. nutrition, 23, 887-894. u.s. department of agriculture, economic research service. (2017). go to the atlas. retrieved from http://www.ers.usda.gov/data-products/food-environment-atlas/go-to-theatlas.aspx u.s. department of agriculture, u.s. department of health and human services. (2015). dietary guidelines for americans 2010-2015. retrieved from https://health.gov/dietaryguidelines/2015/guidelines/chapter-3/social-ecologicalmodel/. https://www.ncbi.nlm.nih.gov/pubmed/?term=horacek%20tm%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=yildirim%20ed%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=matthews%20schreiber%20m%5bauthor%5d&cauthor=true&cauthor_uid=30759796 https://www.ncbi.nlm.nih.gov/pubmed/?term=byrd-bredbenner%20c%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=colby%20s%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=kattelmann%20k%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/30759796 https://www.ncbi.nlm.nih.gov/pubmed/?term=horacek%20tm%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=yildirim%20ed%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=seidman%20d%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=byrd-bredbenner%20c%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=colby%20s%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=white%20aa%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/?term=kattelmann%20k%5bauthor%5d&cauthor=true&cauthor_uid=31236118 https://www.ncbi.nlm.nih.gov/pubmed/31236118 https://www.ncbi.nlm.nih.gov/pubmed/31236118 aresty rutgers undergraduate research journal, volume i, issue ii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. minimization of fuel consumption of a swarm of spacecraft through a genetic algorithm approach shane lecompte, dr. annalisa scacchioli (faculty advisor) ✵ abstract as humanity moves closer to forming realistic paths toward space exploration beyond that what we have already accomplished, multiple new challenges have presented themselves. traditional large spacecraft prove to be unfeasible both logistically and economically for missions where a single problem can completely halt operations, especially given that higher reward missions are also of higher risk. a possible alternative to large craft is using a swarm of smaller craft made to accomplish the same goals while mitigating some of the drawbacks large craft face. rockets, space shuttles, and satellites all prove to be too large to navigate areas of space dense with obstacles. smaller craft on the scale of one meter in a large swarm would navigate these regions. due to the decentralized nature of a swarm, any problems faced by one craft do not necessarily affect the others, allowing the swarm to stay operational despite some crafts becoming compromised. this feature means that a problem or miscalculation that could completely derail an entire mission in the context of a large spacecraft would not do the same to a swarm. in the context of exploring dense and/or extreme environments in space, many logistic and economic problems faced by large craft due to their size and centralized nature will not affect a swarm. with an accurate mathematical model of the swarm dynamics from benet et al.[1], a genetic algorithm’s metaheuristic method is utilized[2] to find optimal parameters that yield a minimal fuel consumption value for a given trajectory/mission objective. from this approach, the total fuel consumption was cut in half while retaining desirable characteristics of the trajectory such as collision avoidance and final formation constraints, giving us a similar course that accomplishes the same goal of transporting craft around objects and disturbances while also minimizing economic losses. 1 introduction sending a large spacecraft to a destination far from earth costs more money and has higher risk factors and implications of failure than the alternative presented in this paper. the materials fees, fuel cost for propelling the craft, and opportunity cost associated with large craft failure all heavily outweigh that of a swarm. failure of any portion of the large craft can result in failure of the entire mission in a domino effect, while the swarm approach mitigates this. sending a swarm through a field of densely packed asteroids or ice rocks in planetary rings involves a cooperative intelligence between individual agents that may not be as sophisticated on their own, but as a whole are comparable to any individual many orders of magnitude larger/smarter on its own. this idea is mimicked in nature with swarms of bees together to accomplish great feats of engineering comparable to humans despite their individual intelligence paling in comparison. despite the hackneyed image that space travel is an incredibly advanced and technical area for humankind, controlling a swarm of autonomous agents moving as a single unit was already solved by nature millions of years before humans ever existed. this trend can be seen in flocks of birds, schools of fish, and colonies of ants. they react to disturbances as a whole and can continue to survive even if individual swarm members are compromised. following the biological inspiration for translating of this problem into a mathematical equation/model, a biologically inspired method of solving it seemed to be in order. the genetic algorithm is based on darwin’s theory of evolution. the idea of “survival of the aresty rutgers undergraduate research journal, volume i, issue ii fittest” as it is described in natural selection is utilized in this algorithm to only allow the most optimal members of each generation to survive to the next, thus giving rise to a criterion for convergence to a global optimum if this algorithm is run consecutively and recursively. these “members” are the parameters used in the differential equation model provided by benet et al.[1] and tweaked in this paper. different values of certain parameters give rise to other solutions (trajectories). thus, the genetic algorithm is utilized to find which parameters result in the most optimal trajectory. in their paper, conn et al. detail nasa’s latest developments with autonomous spacecraft swarms.[3] soda, or swarm orbital dynamics advisor, is a computational framework for a control network that governs communications between spacecraft in a swarm. however, a fuel optimization approach is still yet to be fully explored in this context. optimal path planning and fuel minimization approaches for uavs (unmanned aerial vehicles) have been researched for terrestrial environments[4,5], and in space in the context of transferring between multiple orbits[6], but a combination of these approaches has yet to be extensively developed for space exploration in dynamically dense environments. in this paper, calculating the optimal trajectory of a swarm of spacecraft is investigated, minimizing the total fuel consumption in this specific context, thus finding both a logistic and economically viable alternative to that of traditional missions involving a singular large craft. a specific area of space exploration, as mentioned earlier, in which the results of this work may be applicable is in the exploration of saturn’s rings. while we have been able to visit many moons of other planets in our solar system, we have only been able to visualize saturn’s rings from a distance due to the density and unpredictable size (meters to kilometers) of the ice rocks present in them. this environment provides an unacceptable risk for a multimillion-dollar spacecraft, so such missions have been avoided. a swarm of spacecraft with the same instrumental capabilities of a large vessel and collective intelligence equivalent to that of a larger computer system on a traditional craft that can morph around complex obstacles and dynamic obstacle patterns can provide a solution. the total cost of the cassini mission, a probe that orbited saturn retrieving much of the data we have today about saturn’s rings, was about $3.26 billion, with operations, fuel, and communications costing about $760 million.[7] this last figure is for the costs mostly associated with the cassini mission only after it had entered saturn’s orbit and not the costs from the launch from earth. with this being said, costs associated with launch and leaving earth for both the traditional and swarm approach are the same considering the costs of operations in orbit are lessened by the swarm approach, not necessarily launch and manufacturing costs on earth. the purpose of this paper is to highlight the economic viability of the swarm approach in the context of operations post-launch and during the actual mission. the swarm itself will most likely leave earth on a larger craft and then be deployed from the mothercraft once the destination has been reached. only then will the total cost of the entire project start to become much less than that of a traditional mission where a singular large craft leaves earth and carries out the mission at the destination. not only will a swarm approach to this specific problem have the potential to cost much less, but it also has a much greater margin of error, resulting in less risk. there have been multiple studies showing both the economic and lower risk advantages of the swarm approach instead of the traditional large craft approach.[8,9] various work on the validity of swarm approaches to drone applications and general robotics applications has already been explored.[8,10,11] in addition, nasa has considered projects involving “cubesats” that would accomplish similar goals using a swarm guidance framework within earth’s orbit, also including ideas for swarm missions to saturn in particular to individual moons and the upper atmosphere, but never within the dense parts of the rings.[12] in essence, a swarm approach not only can provide a method of getting more data about saturn’s rings to advance the science of planetary rings but can provide an economically viable and less risk-prone alternative to traditional mission outlines. aresty rutgers undergraduate research journal, volume i, issue ii list of relevant terms artificial potential field – the assignment of values to every point in space that corresponds to the willingness of an object to move toward or away from that point in space while not being tied to any actual physical phenomena or interactions such as gravitation. a standard potential field encompasses these fundamental forces.[1] bifurcation – the ability of a dynamic system to drastically change its behavior with the tuning of a single parameter.[1] convergence – when a system tends toward a certain steady value or state after a certain period of time. line integral – the integral of a function evaluated along a path in space.[14] work – the amount of energy transfer associated with a force acting on an object, as calculated via the line integral of the force along a specified path in space.[14] metaheuristic – a computer search algorithm that is more sophisticated than simply checking all possible solutions in order to solve an optimization problem.[2] genetic algorithm – a metaheuristic method of optimizing a function that is otherwise undefinable analytically or extremely difficult to do using methods of differential calculus. a global optimum is approximated numerically by mimicking darwin’s theory of evolution. with each successive generation, a more “fit” (lower fuel consumption) solution is created by taking the desirable characteristics of the previous generation and using them to pseudo-randomly generate the next. each iteration (or generation) has “genes” associated with it that combine in a way with the possibility of mutation to produce the next generation.[2] 2 methods i. mathematical model in this paper, a swarm of twenty spacecraft is considered with dynamics modeled by a system of differential equations provided by bennet et al.[1] the model works for any number of craft, but twenty are used throughout this paper. the swarm is assumed to start stationed at a larger craft or satellite within space. the swarm will then be launched horizontally into space from this dock. this system of equations is derived from the basic principle of a particle’s behavior in a potential field. in this model, at every point in three-dimensional space, a value for potential is defined, and in this case, the particles are individual spacecraft moving through this space. a potential field can be thought of as either a hill or a well where objects (particles) tend to either roll down the hill (move away from a higher potential) or roll into the well (move closer to a lower potential) wherein this analogy the potential is gravity. we can extend this same concept to a more abstract mathematical model by defining points in three-dimensional space that the spacecraft tend to move away from and others where the spacecraft tend to converge to. these artificial potentials depend on both the position of a craft in space as well as its relative positions to all others in the swarm. suppose there are preset positions in space where low potentials occur in addition to high potentials being defined at the positions of each spacecraft and any obstacles. in that case, a mathematical scheme can be created where both external obstacle and inter-craft collision avoidance and convergence to desired final formations are well defined. the following model (figure 1) was based on the theory of bifurcating potential fields with a more rigorous mathematical derivation in the paper by bennet et al.[1] they were able to derive a relation between the velocity of each craft in a conveniently defined (in the sense that it is artificially constructed with the final product in mind during its inception) potential field to get desired final formation and trajectory characteristics. mathematically, the velocity is equal to the negative gradient of the potential field. the effects of these potential fields on the spacecraft aresty rutgers undergraduate research journal, volume i, issue ii only initiate the thrusters to force the spacecraft in a certain direction. in other words, these potentials are not real and are only defined within the computersensor framework of the swarm. the equation above is a modified version of the model from bennet et al.,[1] taking this into consideration. above in figure 1 the system of differential equations modeling the dynamics of the swarm is given. without going into mathematical detail, these are a system of equations relating the velocities, positions, and relative positions of all swarm members to one another. the solution to this equation is the set of all the position functions of time of each member of the swarm, describing the path each craft follows through space. the diagram next to the model is the representation of the position vectors of two arbitrary crafts with respect to the origin and their relative position vector. this model contains four parts: the exponential potential, the hyperbolic potential, the inter-craft repulsive potential, and the modification, which we introduce as the craft-obstacle avoidance potential. the exponential and hyperbolic potentials control the final formation and trajectory, while the repulsive potential experienced by each craft governs obstacle and inter-craft avoidance. from this, a system of 20 first order, nonlinear, highly coupled, vector differential equations governing the motion of the swarm is obtained. relevant variables and parameters are summarized in the following table (table 1). ii. formulation of optimization problem the main goal in this paper is to determine the values of certain parameters in the model that yield a trajectory that minimizes the total fuel consumption of the swarm. fuel consumption in a craft is defined as being equal to the energy loss associated with a trajectory. under the assumption that there is a direct proportion between fuel used and energy lost, the following calculations are all in joules per kilogram of fuel used. in other words, one kilogram of fuel used is equal to one joule of energy. this fact can be changed for any fuel/energy ratio for a real fuel with a simple multiplicative factor. in this case, energy loss is the work done on the craft by the thruster force along its trajectory. since we neglect gravitation and any other forces acting on each craft, the only force acting on one craft is the force of its thrusters acting to either accelerate or decelerate it. thus, the total work done by the thrusters on the craft will be equal to the amount of fuel used under these assumptions. the expression for work then that we are interested in is the following line integral evaluated along the path defined by the trajectory (figure 2). figure 1: mathematical model of swarm dynamics and reference system figure 2: work line integral aresty rutgers undergraduate research journal, volume i, issue ii where 𝒓𝒓 is the displacement of the spacecraft and 𝑭𝑭 is the net force experienced by the craft due to the thrusters along the path it follows to the end goal. we can then see that 𝑑𝑑𝒓𝒓/𝑑𝑑𝑑𝑑 is equal to the velocity of the craft traveling on this trajectory. we then arrive at 𝑑𝑑𝒓𝒓 = 𝒗𝒗𝑑𝑑𝑑𝑑. it is common to use in the field of dynamical systems to denote time derivatives with a small dot above the variable of interest[13]. the only forces exerted on the crafts being considered are the forces due to the thrusters. gravitation from saturn is neglected due to all particles and agents being within saturn’s orbital reference frame, where the force of gravity is already accounted for in defining this frame. gravitation between spacecraft and ring particles is also neglected due to their extremely small masses that would result in small forces compared to the thruster forces. since the variable “𝑥𝑥” denotes the position of a craft, “𝑥𝑥” with one dot denotes the velocity, while “𝑥𝑥” with two dots denotes acceleration. from this, and the fact that force is equal to mass times acceleration (𝑭𝑭 = 𝑚𝑚𝑚𝑚) by newton’s second law, we obtain the total work done by the thrusters in moving the craft along its trajectory as an integral from the starting time (0) to the final time of the force vector. this pattern is expressed as the product of mass and the second time derivative (acceleration) of the i’th craft’s displacement and 𝑑𝑑𝒓𝒓 being equal to 𝒗𝒗𝑑𝑑𝑑𝑑, which is equivalently expressed as the product of 𝑑𝑑𝑑𝑑 and the first time derivative of the i’th craft’s displacement. finally, we sum over all craft to get the total energy lost (fuel consumed) of the swarm over the course of the trajectory (figure 3). a,b,c parameters that are members of the vector k – control the shape of the final formation (circles, spheres, etc.) ch magnitude of the hyperbolic potential term – determines speed of convergence and works together with exponential potential to determine general trajectory shape ce,le magnitude and length scale of exponential potential term – work together with hyperbolic potential to determine general trajectory shapes and provide an upper bound on velocity cr,lr magnitude and length scale of inter-craft repulsive potentials – determine how strong the repulsive force between crafts is. constants of craft geometry and material cdm,ldm magnitude and length scale of craft-obstacle repulsive potentials – determine how strong the repulsive force between crafts and obstacles is. constants of obstacle geometry μ bifurcation parameter – controls how many steady state formations swarm converges to r scalar – determines how spread out the swarm is as a whole by controlling locations of hyperbolic and expotential potentials xi position vector of the i’th craft xij relative position vector between i’th and j’th crafts ux average velocity with which swarm travels in positive x-direction xim relative position vector between i’th craft and m’th obstacle table 1: relevant parameters summary figure 3: total energy expression of swarm along a trajectory aresty rutgers undergraduate research journal, volume i, issue ii where 𝑑𝑑𝑓𝑓 is the final time during the simulation. the absolute value is necessary considering that if the thruster force is ever acting to decelerate a craft (it is acting opposite the direction of the craft’s velocity), then the mathematical expression for work will yield a negative value, representing an energy gain in the system. fuel is not generated in this instance, so the absolute value bars restrict this work to always represent an energy loss in the system. the above expression is equal to the total energy lost during a swarm maneuver, representing the objective function we wish to minimize. since this function is locked within a highly coupled system of differential equations and an integral that has no closed-form expression (an expression involving a finite amount of known algebraic operations and variables without derivatives and integrals), we must use a more advanced method of optimization which in this case is the genetic algorithm. iii. genetic algorithm in the genetic algorithm, a single member of a population is the set of parameters and its associated final fuel consumption value. we only consider the parameters 𝑚𝑚, 𝑏𝑏, 𝑐𝑐, 𝐶𝐶𝑒𝑒, 𝐿𝐿𝑒𝑒, 𝐶𝐶ℎ, and 𝑟𝑟 as the rest are either constants associated with the spacecraft or obstacles, which cannot be changed or are constants associated with a desired final formation type that we wish to keep constant. the algorithm (refer to figure 4) works by taking a set of random initial parameters, running them through a matlab simulink[14] simulation to obtain the fuel consumed in that trajectory, and then selecting the two sets of parameters that result in the lowest fuel consumed out of all of the initial sets. these two sets are chosen as the parents for the next generation. the parameters associated with these are then converted to a type of binary vector of ones and zeroes. the reason for this is because the next step figure 4: genetic algorithm visualization[2] aresty rutgers undergraduate research journal, volume i, issue ii is the exchanging of the “genetic material” of the two parents to produce offspring. the binary representations of each parent’s parameter are lined up, and a random stopping point is then selected. this random stopping point is selected through a random number generator in the matlab code. every value before the stopping point is swapped amongst the two parents, along with a random chance for one of the values to be inverted (1 to 0 or 0 to 1) at a low probability, representing random mutations. this process is continued a number of times until multiple “child” binary vectors are created, which are then converted back to their decimal representation. together with this genetic crossover process and random mutations, there is a chance for every generation for a more optimal set of parameters to arise while not converging prematurely. after the conversion back to decimal form, the children are run through the same fitness test the previous population went through, and two new parents are selected while the other sets of parameters are scrapped. this process continues until a population converges to a global optimum where the fuel consumption values are all equally as fit as each other. of course, other convergence criteria could be used (energy consumption below a certain value across the swarm, total time spent, etc.), but defining convergence in this way where every craft is equally optimal tends to prioritize the crafts expending the most fuel first. this process was implemented in matlab[14] with multiple simulink simulations running per generation to get the final fuel consumption value for each new set of parameters. information and methods followed that were used to code this algorithm are presented in the article by mallawaarachchi.[2] a similar evolution-based optimization method was successfully used in path planning a single uav in a terrestrial environment in the paper by rathbun et al.[4] 3 numerical results in the following simulations, there is a swarm of twenty spacecraft initially arranged in a circle at the vertical launch pad and three static disturbances (ice rocks/asteroids) in space along the trajectory. in the context of saturn’s rings, this simulation represents a swarm being deployed from a mothercraft near saturn into a region of saturn’s rings with ring particles acting as obstacles. it should be noted that this simulation and approach will work for any number of spacecraft and disturbances that may or may not be static. still, for this work, a simpler scenario was chosen, primarily to demonstrate the efficacy of the chosen genetic algorithm in reducing fuel consumption. as described before, the fuel consumption depends on the force exerted on a craft by the thrusters, which is proportional to the acceleration of the craft. whenever crafts are accelerating to form a final formation, avoid an obstacle, or avoid another craft, the thrusters must exert force on the craft to move it, thus spending fuel. smoother trajectories generally use less fuel than trajectories where sharp turns and abrupt accelerations are required to avoid collisions. within this preliminary model, collisions may happen and can be checked for but will not be physically accurate. the agents will move through obstacles or other craft upon collision. this simulation’s main goal is to validate the genetic algorithm before making the mathematical model and environment too complicated. future work on this project will involve the implementation of physically accurate collision physics and the event of a craft colliding and being compromised, leaving the swarm. first, we start with an initial population of five sets of parameters, listed in the chart to the left of figure 5. the trajectory and fuel consumption analysis for one craft is shown below in figure 7. in figure 5 we show the parameters used for this simulation as well as the swarm trajectory. figure 6 presents an alternate view. as seen from figure 5 and figure 6, the swarm starts in an initial circular launch apparatus. it then moves through space approaching the final formation while avoiding the obstacles at locations denoted by large open circles and the other craft. the blue craft on the right provides an example of such aggressive avoidance behavior with an aggressive fuel consumption depends on the force exerted on a craft by the thrusters, which is proportional to the acceleration of the craft aresty rutgers undergraduate research journal, volume i, issue ii maneuver at around 10 meters. the red trajectory next to it also appears to move out of the way of the blue one when it moves in the red trajectory’s path to avoid the obstacle as well. this trend is better visible in figure 5. as the blue craft dips inward toward the left, the red craft above it speeds up (more space in between consecutive dots represents faster movement) to avoid the blue craft. although they are not necessarily on a collision course, they could be as these dots only represent the centers of spherical crafts, which could have varying sizes. the magnitudes of the inter-craft repulsive potentials responsible for this collision avoidance behavior would factor in the craft’s radius when being chosen. in this case, the red and blue crafts became a bit too close to each other, so the avoidance mechanism governed by the potentials was initiated. in figure 7 we show the fuel consumption against the time of this same blue craft described above. a graph of a singular craft’s fuel consumption is presented as it is still indicative of the swarm’s consumption as a whole. this blue craft was the most responsible for increased fuel consumption in this trajectory due to its harsh avoidance maneuver. the fuel consumption spikes during periods of high acceleration and saturates once the swarm converges to the final formation as expected since the craft is no longer accelerating at that point. also noteworthy is that fuel consumption spikes around when the blue trajectory aggressively avoids the obstacle (asteroid) bound for a head-on collision if it did not move out of the way. all of the fuel consumptions summed for each craft for this simulation was 3462.7 j/kg. this value is accurate within the model itself as it is a result of pure mathematical calculations without any data being taken. after running this data set (set of the seven parameters being considered) and four others through the genetic algorithm, it converged to a possible optimal data set after 71 generations (or 355 individual simulations in simulink). this process worked by taking the original population of five sets of the seven parameters being considered and running the simulink simulation and corresponding total fuel usage for a swarm. with the five values of fuel consumption, the two smallest are considered and figure 5: parameters and associated trajectories of swarm traveling in x-direction. different colored trajectories represent different spacecraft. figure 6: top-down view of same trajectory. different colored trajectories represent different spacecraft. figure 7: fuel consumption (j/kg) of one craft against time (s). aresty rutgers undergraduate research journal, volume i, issue ii used as parents for the next generation. using the technique described in the section on the genetic algorithm, five more sets of parameters (that have a higher chance of being optimal) are generated from the parents. one generation consists of the five simulations and calculations of total fuel consumption. so in total, the algorithm converged after 71 of these or 355 individual instances of numerically solving the system of equations and calculating integrals. the algorithm’s optimal parameters output and their corresponding trajectories when run through the same simulation are described below in figures 8 and 9. one of the first things immediately made apparent is that the blue trajectories now no longer need to avoid the obstacle so aggressively, and the swarm as a whole being generally smoother as it converges to a final formation. with the use of the genetic algorithm, the swarm could navigate through the field of obstacles and arrive at the same destination while minimizing the fuel used to arrive there. in figure 10 we display the fuel consumption versus time of the same craft as before, only now after the algorithm has run and the new parameters put in place. the fuel consumption saturates in about half the time as the original simulation reaching a maximum value of a little less than 60 j/kg, which is much smaller than the original in figure 7, being at about 170 j/kg. another notable aspect of this new trajectory is that the fuel consumption curve in figure 10 increases much more smoothly than that of the unoptimized swarm, which has many abrupt spikes. in the context of a real-world scenario, the same mission could be completed for half the fuel cost while also putting less strain on the thrusters. 4 conclusions although these are preliminary results, they show that the genetic algorithm approach can be successfully optimize the fuel consumption of a swarm trajectory of spacecraft defined by this complex nonlinear model. the solution to this problem provides a method of exploring and gaining new information on the rings of saturn that otherwise would not be possible. future work will include making the model more realistic and transitioning the figure 8: parameters and associated trajectories of swarm after genetic algorithm. different colored trajectories represent different spacecraft. figure 9: side view of same trajectory. different colored trajectories represent different spacecraft. figure 10: fuel consumption (j/kg) of one craft against time (s) after genetic algorithm. aresty rutgers undergraduate research journal, volume i, issue ii deterministic problem to a stochastic one; denser obstacle distributions, and obstacles that are moving, will also be considered in future simulations more akin to the real-world density of ice rocks in the rings of saturn. applying the genetic algorithm method to the same spacecraft swarm system described by more accurate stochastic models can lead to future innovative developments in this field. new methods of optimization will also be investigated∎ 5 acknowledgements i would first like to thank dr. annalisa scacchioli for guiding me through this process and opening up this wonderful opportunity to learn as well as create something meaningful. without her, this project would not have come to fruition. i would also like to thank the nasa new jersey space grant consortium for funding and facilitating this research. also, i would like to thank dr. jingang yi as well as dr. enver koray akdogan for providing insight into the inner workings of the mathematical model used in this paper as well as possible avenues to best utilize it for this project. 6 references [1] bennet, derek j., mcinnes, c.r., suzuki, m., and uchiyama k., “autonomous three-dimensional formation flight for a swarm of unmanned aerial vehicles,” journal of guidance, control, and dynamics, vol. 34, no. 6, november-december 2011, 1899-1908. [2] mallawaarachchi, vijini, “introduction to genetic algorithms – including example code,” medium towards data science, 7 july 2017. [3] conn, t., plice l., dono perez, a., and ho, m., “operating small sat swarms as a single entity: introducing soda,” nasa ames research center / mission design division, 31st annual aiaa/usu conference on small satellites, 2017. [4] rathbun, david, kragelund sean, and pongpunwattana, anawat, “an evolution based path planning algorithm for autonomous motion of a uav through uncertain environments,” university of washington, seattle, wa, 2002. [5] zhao, yiyuan j. and celia qi, ying, “minimum fuel powered dynamic soaring of unmanned aerial vehicles utilizing wind gradients,” university of minneapolis, mn, 2004. [6] morgan, daniel, chong, soon-jo, and hadaegh, fred y., “model predictive control of swarms of spacecraft using sequential convex programming,” journal of guidance, control, and dynamics, vol. 37, no 6, november-december 2014, 1725-1740. [7] “cassini-solstice mission faqs,” jet propulsion laboratory. retrieved january 24, nasa, 2014. [8] blocher, a., “alternative mission concepts for the exploration of outer planets,” california polytech state university, master’s thesis 2017. [9] hadaegh, f. y., chung, s., manohara, h. m., “on development of 100-gram class spacecraft for swarm applications,” ieee systems journal, vol. 10, no. 2, pages 673-684, june 2016, 2016. [10] hinchley, m. g., rash, j. l., truszkowski, w. f.,“autonomous and autonomic swarms,” information systems division, nasa goddard space flight center, technical report, 2006. [11] nallapu, r., thangavelautham, j., “design of spacecraft swarm flybys for planetary moon exploration,” structures, structural design and materials conference, technical report, 2019. [12] blocher, a., atkinson, d., freeman, t., “saturn swarm study small probe and cubesat architectures to accompany new frontiers missions at saturn” low cost planetary mission conference, presentation, 2017. [13] matlab simulink, math and computer algebra system, coding interface, and simulation environment, mathworks, 2019. [14] hibbeler, russel c., engineering mechanics: dynamics, 14th edition, 2016. aresty rutgers undergraduate research journal, volume i, issue ii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. particle trajectories for compton scattering in one space dimension adriana scanteianu, xiangyue wang ✵ abstract using a relativistic extension of bohmian mechanics known as multi-time wave function formulation, we examine a two-body, one-dimensional system consisting of one photon and one electron that interact only upon contact. we investigate the effects that various parameters in this theory including momentum of the incoming photon and mass of the electron have on the dynamics of the two interacting bodies with the goal of understanding conservation of momentum and energy in the system. we show that the core principles of compton scattering remain when we use this alternative formulation of quantum mechanics. although a complete relativistic theory of bohmian mechanics has yet to be developed, our work aims to make the ideas in this theory more accessible to a wider audience. 1 introduction mathematical physics aims to provide logically consistent and mathematically rigorous explanations for why nature behaves the way it does. currently, at the fundamental level, there are two experimentally verified and mathematically rigorous theories that seem to be in conflict: namely, quantum mechanics and general relativity. while quantum field theory is, for all practical purposes, highly successful at reconciling quantum mechanical principles with einstein’s theory of light (special relativity), it is incompatible with einstein’s theory of gravitation (general relativity). it is also mathematically non-rigorous due to its inherent infinities[14]. such incompatibilities and lack of rigor have led mathematical physicists to continue exploring other theoretical frameworks. more than half a century ago, louis de broglie and david bohm formulated a non-relativistic theory where elementary particles exist regardless of observations[2,4,10]. de broglie-bohm theory, otherwise known as bohmian mechanics, is a formulation of quantum mechanics in which subatomic particles in a system have definite positions, guided by the wave function of the system[10]. in one of the relativistic extensions of this theory, there is a single multi-time wave function originally proposed by dirac, defined on the configuration space of the system that guides the motion of all the particles[5]. some researchers believe the photon wave function was never explored due to historical reasons: dirac completed his quantization of the electromagnetic field prior to discovering his relativistic electron wave equation, and the world of research acted on the first discovery more than the second[1,5,6]. goldstein and others have developed non-relativistic bohmian mechanics in great mathematical detail[9,10]. the remaining challenge is developing a fully relativistic bohmian mechanical system with a fixed number of particles. some have claimed that this task is downright impossible[15]. nevertheless, kiessling and tahvildar-zadeh successfully developed a relativistic bohmian theory for a single photon and, together with lienert, they have developed an interacting electron-photon theory in one space dimension[11,12]. in this paper, we explore whether their theory is applicable to the study of compton scattering. arthur compton discovered, ninety-six years ago, that when x-rays interact with a graphite plate, the scattered rays have lower frequencies than the incident ones and the difference in the energy of the “x-ray quanta,” now called photons, turns into kinetic energy of the electrons on the surface of the material[3]. with this experiment, compton provided evidence for einstein’s 1905 theory, which stated that aresty rutgers undergraduate research journal, volume i, issue ii light behaves as particles. we now know this phenomenon as compton scattering. quantum field theory provides us with the following narrative about what happens during the interaction: just before the photon smashes into the electron, it abruptly vanishes into nothing. at the same moment, a new photon with a different frequency emerges next to the electron out of thin air, and the two depart as if they have just collided. we would like to explore a different narrative of a photon-electron collision based on bohmian principles, as provided in the paper by kiessling et al [11]. the novelty of our work lies in the actualization of previous theory via our modeling of compton scattering using matlab software[11,13]. furthermore, we are the first to use this novel theoretical framework to analyze the effects of changing the momentum of the incoming photon on the trajectories of the scattered photon and electron. we are also the first to investigate the conservation of momentum and applicability of compton’s scattering formula in this system. the framework we use is compatible with many experimentally verified theories in physics, thus we hope that our work will bring us one step closer to a mathematically rigorous understanding of the world around us on the subatomic level[11]. 2 methodology in this paper, we treat both the electron and the photon as relativistic quantum objects. first, we model the trajectories of a single photon in one space dimension using the photon wave function proposed by kiessling and tahvildar-zadeh[12]. next, we do the same for a single electron using dirac’s relativistic wave equation for the electron[6]. we then examine the two-body quantum-mechanical system of a photon and an electron, first assuming that the two do not interact, and then implementing the photon-electron interaction through a boundary condition, as in the paper by kiessling et al[11]. afterwards, we review the method compton used to compute the formula for the scattering angle in his experiment. we then compare compton’s predicted value for momentum gain of the electron with our computed values for this quantity based on our numerical solutions of the guiding equations for the photon and the electron. we then remark on the obstacles that one needs to overcome if one is to rigorously derive such a formula from our bohmian perspective. note that our unit of time is the attosecond (1 attosecond = 10—18 seconds) and our unit of length is 2.997924 angstroms (1 angstrom = 10—10 meters). this allows for the speed of light in a vacuum to be 1 unit of length over time. i. single photon theory the wave function of a photon is a rank-2 bispinor field defined on the configuration space time of the photon [12]. in one space dimension this means: where 𝑡𝑡 ∈ ℝ is a time coordinate and 𝑠𝑠 ∈ ℝ is a space coordinate. since a photon is a relativistic quantum particle, its wave function must satisfy a relativistic partial differential equation, which was discovered by kiessling & tahvildar-zadeh[12]. the photon wave equation is a dirac-type equation, and in one-dimension it reads: where ℏ is the reduced planck’s constant, 𝑥𝑥0 = 𝑐𝑐𝑡𝑡, 𝑥𝑥1 = 𝑠𝑠 , 𝛾𝛾0 = �0 1 1 0 � , and 𝛾𝛾1 = �0 −1 1 0 � ; and repeated indices are summed over the range 𝜇𝜇 = 0, 1. here, 𝑐𝑐 is the speed of light in a vacuum. we will be working in units where 𝑐𝑐 = 1. the photon wave equation can be solved given an initial wave function 𝛹𝛹𝑝𝑝ℎ(0, 𝑠𝑠) = 𝛹𝛹𝑝𝑝ℎ 0 (𝑠𝑠) . typical initial data corresponding to a photon localized in both position and momentum space (subject to uncertainty principle) can be found in figure 1. in our case, the typical initial data are gaussian distributions in photon position space. since the fourier transform of a gaussian is a gaussian, these also have a gaussian distribution in momentum space. therefore, there are two parameters that we can set: the mean 𝑘𝑘0 ∈ ℝ of the momentum distribution, and the momentum standard deviation 𝛼𝛼𝑝𝑝ℎ. recall that according to the uncertainty principle, aresty rutgers undergraduate research journal, volume i, issue ii 𝛼𝛼𝑝𝑝ℎ𝜎𝜎𝑝𝑝ℎ ≥ ℎ 2 , where 𝜎𝜎𝑝𝑝ℎ is the standard deviation in photon position space, and that this inequality is saturated by the gaussian distribution. furthermore, we recall the einstein-de broglie energy and momentum relations 𝐸𝐸 = ℎ𝜐𝜐, 𝑝𝑝 = ℎ𝑘𝑘, and the photon dispersion relation 𝐸𝐸 = |𝑝𝑝| (recall that 𝑐𝑐 = 1 .) these imply that we may identify |𝑘𝑘0| with the mean of the initial frequency distribution of the photons. the probability current of detecting a photon at a point in one space and one time dimension is[12]: where 𝑋𝑋 = (𝑋𝑋0, 𝑋𝑋1) is a constant vector field computed from 𝛹𝛹𝑝𝑝ℎ 0 , 𝛾𝛾(𝑋𝑋) ≔ 𝛾𝛾0𝑋𝑋0 + 𝛾𝛾1𝑋𝑋1 , 𝛹𝛹� ≔ 𝛾𝛾0𝛹𝛹†𝛾𝛾0 , and tr denotes matrix trace. the current is conserved: 𝜕𝜕𝜇𝜇𝑗𝑗𝑝𝑝ℎ 𝜇𝜇 = 0 , future directed (𝑗𝑗0 ≥ 0) , and timelike (𝑗𝑗0 ≥ |𝑗𝑗1|). thus, the probability density of detecting the photon at event (𝑡𝑡, 𝑠𝑠) is 𝜌𝜌(𝑡𝑡, 𝑠𝑠) = 𝑗𝑗𝑝𝑝ℎ 0 (𝑡𝑡, 𝑠𝑠). the motion of the photon is guided by its wave function through the following system of guiding equations[11]: here, 𝑞𝑞(𝑡𝑡) is the position of the photon at time 𝑡𝑡; 𝑞𝑞0 is the initial position of the photon. all we know about the initial position is that it is randomly distributed according to the initial probability density 𝜌𝜌(0, 𝑠𝑠). by solving the system of guiding equations, we obtain the trajectory of a single photon over time. ii. single electron theory according to paul dirac, the wave function of a single electron is a spin1 2 field defined on the configuration space of the electron[6]. in one space dimension this means: . like in the case of a single photon, the wave function of a single electron also satisfies a relativistic equation. in particular, it satisfies the massive dirac equation: , where 𝑚𝑚𝑒𝑒𝑒𝑒 is the mass of an electron. the probability current of an electron is: , where 𝛹𝛹� ≔ 𝛹𝛹†𝛾𝛾0 is the dirac adjoint for rank-one figure 1: initial data for the photon wave equation is split into real and imaginary components. we assume typical initial data: the photon starts off as a gaussian wave packet both in position and in momentum-space. since the unit of frequency is the reciprocal of time, our unit of frequency is 1018 hz. aresty rutgers undergraduate research journal, volume i, issue ii bispinors. furthermore, as in previous research, we define the guiding velocity field of an electron in the same way as that of a photon[11]: . similarly to the photon case, the guiding equation for the electron is[11]: the equation satisfied by the electron wave function 𝛹𝛹𝑒𝑒𝑒𝑒 has a mass term: 𝜔𝜔 = 𝑚𝑚𝑒𝑒𝑒𝑒𝑐𝑐/ℏ. in addition, the gaussian family of initial distributions has a parameter we can change, namely the standard deviation of the probability distribution of electron’s initial position: 𝜎𝜎. we will examine the effect of these two parameters on the joint dynamics. iii. the two-body system now, we examine the case in which a photon and an electron are both present, but do not interact. in this case, we have one wave function that guides the motion of each particle through its respective system of guiding equations. the wave function, 𝜓𝜓, is a function of four variables, namely the time and position of each particle. to get a wave function that describes both a photon and an electron in a non-interacting system, we take the tensor product (⨂) of the electron and the photon wave functions, giving us a four component object 𝜓𝜓 = (𝜓𝜓−−, 𝜓𝜓−+, 𝜓𝜓+−, 𝜓𝜓++). the guiding equations for photon and electron are derived using the hypersurface bohm-dirac (hbd) theory, which allows us to describe the motion of the photon and electron in a common time coordinate[8]. the tensored wave function satisfies a relativistic wave equation obtained by the tensor product of the photon and electron wave equations: the joint probability current is the following: then, the guiding equation for the photon is given by: while the guiding equation for the electron is: iv. the two-body system to obtain an interacting system from our two-body non-interacting system, it is necessary to add a boundary condition that prevents the particles from simply going through each other. in this is done by adding a boundary condition on the coincidence set: they set the relative velocity of photon and electron to be 0 whenever the two particles are at the same space and time point[11]. in other words: when v. derivation of compton scattering formula to better explain the formula used to calculate predicted momentum of particles in the system, we present a short derivation of the original compton scattering formula in our one-dimensional setting. assume we have positive energy plane waves going into and coming out of the scattering zone. assume that conservation of momentum and energy still hold. (2.1) (2.2) (2.3) (2.4) aresty rutgers undergraduate research journal, volume i, issue ii conservation of momentum means: going into the rest frame of electron, 𝑘𝑘𝑒𝑒𝑒𝑒 𝑖𝑖𝑖𝑖 becomes 0. hence: suppose the photon is approaching the electron from the right. thus 𝑘𝑘𝑝𝑝ℎ 𝑖𝑖𝑖𝑖 > 0 . conservation of energy is given by: (here and henceforth, we have chosen units in which the speed of light in vacuum is 𝑐𝑐 = 1.) substituting equation 2.6 into equation 2.7 and attempting to solve for 𝑘𝑘𝑝𝑝ℎ 𝑜𝑜𝑜𝑜𝑜𝑜 we observe that choosing the plus sign in the right-hand-side of equation 2.7 yields no solution. thus, the only option is to choose the negative sign, which implies that 𝑘𝑘𝑝𝑝ℎ 𝑜𝑜𝑜𝑜𝑜𝑜 < 0 since we have assumed that the plane waves have positive energy. as a consequence, the direction of motion of the photon has changed after the scattering event. solving for 𝑘𝑘𝑝𝑝ℎ 𝑜𝑜𝑜𝑜𝑜𝑜 then yields the scattering formula: we note that this agrees with the three-dimensional compton scattering formula derived in the 1923 paper if the scattering angle in that formula is set to 𝜋𝜋. this formula also suggests that the frequency of the scattered x-ray will be lower than the frequency of the incident ray and that the difference in the momentum and energy is transferred to the electron. more precisely, combining equation 2.8 with equation 2.6 and recalling that 𝜔𝜔 = 𝑚𝑚𝑒𝑒𝑒𝑒𝑐𝑐/ℏ one obtains: 3 results throughout the results section, it is important to note that time is measured in attoseconds (10—18 seconds), and length is measured in units of approximately 3 angstroms (2.997924 × 10—10 meters). in these units the speed of light in vacuum is equal to 1 unit of length over time. i. single photon motion the photon probability density looks like this: https://reu.dimacs.rutgers.edu/~aas377/photon_pdf.mp4 varying the standard deviation of the initial profile gives us the following: https://reu.dimacs.rutgers.edu/~aas377/multiple_photon_pdf.mp4 figure 2 shows multiple trajectories, each with slightly different initial positions. ii. single electron motion the evolution of the probability distribution of the electron’s position can be found here: http://reu.dimacs.rutgers.edu/~aas377/electron_new.mp4 figure 3 shows the trajectories of many electrons, each in a single-body system guided by its single wave function through its guiding equation, with different initial conditions for each electron. unlike the trajectories of the photon, the trajectories of the electron exhibit oscillations and travel slower than the speed of light. iii. non-interacting two-body system since tensor products preserve probability distributions, “multiplying” the probability density movies of the photon and electron gives the joint probability density for the non-interacting system. since there is no interaction, we expect the joint probability density to be simply the product of the probability density of a single electron and that of a single photon. the numerical results confirm our expectation. https://reu.dimacs.rutgers.edu/~aas377/non_int_pdf_mesh_new.mp4 by solving the system of guiding equations, we obtain figure 4, which shows the trajectories of a non-interacting system of one electron and one photon. (2.6) (2.7) (2.8) (2.9) https://reu.dimacs.rutgers.edu/%7eaas377/photon_pdf.mp4 https://reu.dimacs.rutgers.edu/%7eaas377/multiple_photon_pdf.mp4 http://reu.dimacs.rutgers.edu/%7eaas377/electron_new.mp4 https://reu.dimacs.rutgers.edu/%7eaas377/non_int_pdf_mesh_new.mp4 aresty rutgers undergraduate research journal, volume i, issue ii top figure 2: a free photon in one dimension would move either left or right at the speed of light. the linear trajectories in the diagram validates this intuition. trajectories with different initial conditions are differentiated by color. middle figure 3: a free electron in one dimension moves according to its wave function at a velocity slower than the speed of light. trajectories with different initial conditions are differentiated by color. bottom figure 4: non-interacting photon (left) and electron (right) trajectories. since the two particles do not interact, they pass through one another as the diagram indicates. trajectories with different initial conditions are differentiated by color. below figure 5: interacting and non-interacting photon (red) and electron (blue) trajectories. the interaction resembles a collision, after which the two particles bounce away from each other. dotted lines show the non-interacting case, and the solid lines show the interacting case. aresty rutgers undergraduate research journal, volume i, issue ii [a]for reasons having to do with numerical computations, the value we use for ω is numerically equal to the mass of the electron but is not the actual physical value. the actual mass of an electron in our units is 776.343694 units of mass, but we are working on the order of 1 unit of mass. that is to say that we are working with an electron mass that is 700 times smaller. iv. interacting two-body system adding the boundary condition to the wave function gives us a modified probability density function: http://reu.dimacs.rutgers.edu/~aas377/int_pdf_mesh_new.mp4 figure 5 is the contour map. the shape of the interacting probability density is delineated by lines from the perspective of someone looking down from above: http://reu.dimacs.rutgers.edu/~aas377/int_pdf_mesh_new.mp4 these are the trajectories of the interacting electronphoton system with the boundary condition added in. the solid lines show the interacting system trajectories and the dotted lines show the non-interacting trajectories. plotting 100 overlaid trajectories of the interacting system gives figure 6. to make this clearer, here is a video of the collision trajectories being plotted individually on the same set of axes. http://reu.dimacs.rutgers.edu/~aas377/int_trajs_slowmo.mp4 v. varying the electron parameters here, we sought to perform numerical experiments using our model. the units used in calculating the values of our parameters are chosen such that planck’s constant ℏ = 1 unit of mass × length squared over time: ℏ = 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 × 𝐿𝐿𝑒𝑒𝑖𝑖𝐿𝐿𝑜𝑜ℎ 2 𝑇𝑇𝑖𝑖𝑇𝑇𝑒𝑒 . in our units, one unit of mass is 0.117337 × 10—32 kg[a]. the parameter 𝜔𝜔 = 𝑇𝑇𝑒𝑒𝑒𝑒𝑐𝑐 ℏ = 1 𝜆𝜆 , where 𝜆𝜆 is the reduced compton wavelength of the electron and equal to approximately 0.001288 units of length. note that change in 𝜔𝜔 corresponds to a change in mass of electron 𝑚𝑚𝑒𝑒𝑒𝑒 since the units are set such that 𝑐𝑐 and ℏ are set to equal 1. since 𝜔𝜔 is the reciprocal of wavelength, we top figure 6: 100 overlaid interacting photon (red) and electron (blue) trajectories. bottom figure 7: looking across each row, we can see variation in standard deviation of initial distribution of the electron’s position, σ, while the frequency of the electron, ω, is kept constant across each row. looking down each column, we can see variation in the frequency of the electron, ω!, as the standard deviation of the initial distribution of the electron’s position, σ, is kept constant. http://reu.dimacs.rutgers.edu/%7eaas377/int_pdf_mesh_new.mp4 http://reu.dimacs.rutgers.edu/%7eaas377/int_pdf_mesh_new.mp4 http://reu.dimacs.rutgers.edu/%7eaas377/int_trajs_slowmo.mp4 aresty rutgers undergraduate research journal, volume i, issue ii [b]we are of course aware that electron’s rest mass is fixed. the point here is to do a mathematical study of how the trajectories would look like for a hypothetical particle with the same characteristics of the electron but a different rest mass. can refer to 𝜔𝜔 as electron frequency. change in 𝜎𝜎 refers to change in the standard deviation of the initial probability distribution of the electron’s position in one-dimension, i.e. the standard deviation of the gaussian distribution representing initial data. the units of 𝜎𝜎 are the same as the units for position: approximately 3 angstroms. the changes in these parameters are visible in figure 7[b]. vi. varying the phase angle of interaction in the boundary condition that provides the interaction, it is possible to specify a phase angle. it seems there are no changes as a result of variation in this parameter; however, future research is necessary to elucidate whether this parameter truly has no effect on trajectories. it seems to be that the three graphs in figure 8 are the same since they look like one set of trajectories instead of three when overlaid. vii. varying momentum of incoming photon changing the momentum of the incoming photon gives us a different picture and lays the groundwork for analyzing compton scattering in our setting. the units of momentum are mass × velocity. our unit of mass is 0.117337 × 10—32 kg, and our unit of velocity is 2.997924 × 10–8 m/s (the speed of light in a vacuum). hence, one unit of momentum is 0.351767 × 10–24 kg m/s. in order to get a more detailed picture, we varied the mean momentum of the incoming photon between 0 and 10 with mesh of 0.5. the result is figure 10. viii. compton scattering in our model we have found some evidence of the applicability of compton’s scattering formula to our model, as seen in figure 11. the figure corresponds to the parameter values 𝜔𝜔 = 2 and 𝜎𝜎𝑝𝑝ℎ = 𝜎𝜎𝑒𝑒𝑒𝑒 = 0.1 , while the initial photon mean frequency is increased from 𝑘𝑘0 = 1 to 𝑘𝑘0 = 10 , and the resulting electron trajectories are distinguished by color. again, the unit of momentum is 0.351767 × 10–24 kg m/s. top figure 8: the left-most plot shows an angle of interaction of 0. the middle plot shows an angle of interaction of π/2 and the right-most plot shows an angle of interaction of π /4. we can see that as the angle of interaction varies, no change occurs in the particle trajectories. non-interacting trajectories are visible in dotted lines, and all other parameters are kept constant. bottom figure 9: interacting photon (red) electron (blue) trajectories with varying photon momentum. in the left-most plot, the momentum of the incoming photon is 0.1; in the middle plot, the momentum of the incoming photon is 1, and in the right-most plot, the momentum of the incoming photon is 10. thus, we see that a higher energy photon—a photon with higher momentum—gets the electron to bounce away. again, non-interacting trajectories are visible in dotted lines. aresty rutgers undergraduate research journal, volume i, issue ii top figure 10: each pair of trajectories depicts a variation in the momentum of the incoming photon. bottom figure 11: left: change in bohmian trajectories of the two particles due to the increase in incoming photon frequency. right: the corresponding change in electron’s post-scattering momentum, computed from those trajectories. aresty rutgers undergraduate research journal, volume i, issue ii top figure 12: change in electron momentum due to scattering (detail). bottom figure 13: left: dotted lines show the non-interacting case, and solid lines show interacting cases at various incoming photon frequencies. right: solid lines refer to momentum of the outgoing electron at various frequencies of incoming photon. the dotted line shows the momentum of the electron in the non-interacting case, as the two particles “go through” each other. aresty rutgers undergraduate research journal, volume i, issue ii if we zoom in on the interaction region, we clearly see that the gain in electron momentum increases linearly with 𝑘𝑘0 , in qualitative agreement with equation 2.9. see figure 12. in these figures we have calculated the electron’s “post-collision” momentum by where 𝑣𝑣𝑒𝑒𝑒𝑒 is the bohmian velocity of the electron calculated from the wave function, as in equation 2.2, and evaluated at the actual positions of the two particles that have been calculated by solving the two differential equations equation 2.1 and equation 2.2. the graph on the right of figure 13 shows that even in the non-interacting case, the momentum of the electron computed from trajectories is not constant because its bohmian velocity is not constant; instead, it fluctuates in time. finally, we compared our computed electron momentum to the theoretical values computed with compton’s formula. our results show that some frequencies of the incoming photon allow for a better match than others between compton’s formula and computations based on de broglie-bohm theory. further research in this area would be useful to understand which regime of parameters allow for a better fit, and why. 4 discussion in this research, we studied probability density functions and bohmian trajectories for a single electron and a single photon guided by their respective wave functions in one space dimension[10,12]. we then analyzed the probability density functions and bohmian trajectories of both a non-interacting and interacting two-body system consisting of a photon and an electron, in one space dimension, where both particles were guided by the same two-body multi-time wave function. dirac introduced the concept of multi-time wave functions, but the idea that the wave function guided the motion of particles in the system did not arise until bohmian mechanics was developed[2,7]. the idea that multi-time wave functions can guide the motion of relativistic particles is far more recent[8]. we studied how the dynamics of the two particles depended on several key parameters in the problem, such as the frequency of the incoming photon, and the rest mass of the electron. we did a preliminary analysis of compton scattering, and found qualitative agreement between compton’s quasi-classically derived formula for the gain in electron’s momentum, and the corresponding values computed from the trajectories. however, there is much work that remains to be done. in our modeling of the two-body system, the trajectories of the non-interacting system were shown to simply be a superposition of the trajectories of a single photon and a single electron. this was no longer the case once the boundary condition was added. the wave function governing the system indicates that post-interaction, the two particles become entangled even if we initially have a pure product state. in varying the parameters of mass and standard deviation of the initial distribution for the electron, we saw that more massive electrons were likely to have a straighter trajectory, indicating slower figure 14: the red line shows theoretical values of momentum of the outgoing electron computed using compton’s formula. the blue circles indicate the results of our numerical investigation of the outgoing electron momentum computed from bohmian trajectories. aresty rutgers undergraduate research journal, volume i, issue ii movement and less likelihood to recoil away. furthermore, varying the standard deviation of the initial distribution of the electron’s position seemed to alter the trajectories and the nature of the interaction: lower standard deviation values resulted in the electron and photon staying together for longer instead of bouncing off immediately. however, more investigation is required to better understand the interplay between these parameters. in varying the phase angle of interaction, we found evidence that the dynamics of the interaction is independent of that phase angle. our work in varying the momentum of the incoming photon shows that shooting a higher energy photon at the electron will proportionally increase the electron’s post-collision momentum in a manner consistent with compton’s formula. i. analysis of compton scattering results it should be noted that we cannot expect the actual electron momenta to be uniformly close to the values predicted by equation 2.9, since this formula is of a classical nature (apart from using einstein’s suggested values for the momentum and energy of the photon). it has been derived using the rules of classical mechanics, such as exact conservation of a welldefined energy and momentum for both particles. for instance, in the example above, the initial electron distribution has a standard deviation of 0.1; therefore, by the uncertainty principle, the standard deviation of its momentum distribution will be at least 5~. if we randomize the initial position of the electron according to the initial distribution, as we are supposed to do in a quantum theory, we cannot expect the electron momentum at any later time to be close to any one particular value. it is only expected to fluctuate around some mean. this would be equally true for a non-interacting electron. it is a hallmark of the quantum nature of elementary particles. we would nevertheless like to be able to derive a quantum-mechanical (i.e. appropriately statistical) version of formula equation 2.9 within our framework. there are, however, at least two major obstacles on the way to defining a notion of momentum for a quantum particle, particularly for a photon. firstly, our theory does not admit plane waves since these are not square-integrable, and therefore do not belong to our hilbert space. instead, we work with gaussian wave packets. secondly, after interaction, the system is entangled. we may start with a wave function that is a pure product, but we no longer have a pure product after the photon and electron interact. thus, even if we could start with plane waves going in, we would not have plane waves coming out, which violates the underlying assumption used to derive equation 2.9. developing a workable notion of momentum for the particles in our system will be part of our future work. ii. future research we plan to investigate ways of defining a notion of momentum in our framework: one that would allow us to derive the one-dimensional analog of the compton scattering formula in some appropriate classical or semi-classical limit. defining momentum is easy for a single particle using fourier transform. there are, however, many difficulties in isolating photonic and electronic properties from the twobody wave function of the system. a potential way to address this might be to use techniques from linear algebra to decide whether the wave function of the system becomes locally approximately a pure product far away from the interaction zone. naturally, we also plan to expand our analysis to three space dimensions. we hope to eventually arrive at an n-body, relativistic, and quantum mechanical system in three space dimensions from which maxwell’s equations can be derived in the limit as the number of photons goes to infinity∎ 5 acknowledgements we would like to thank the rutgers university dimacs reu program and the rutgers university math department for their generous support. we are grateful to dr. shadi tahvildar-zadeh for his patient guidance, and parker hund for our valuable discussions. we thank erika melder for her technical support. aresty rutgers undergraduate research journal, volume i, issue ii 6 references [1] bialynicki-birula, i. (1996). the photon wave function. in coherence and quantum optics, vii , 313-322. [2] bohm, d. (1952). a suggested interpretation of the quantum theory in terms of hidden variables. part i. phys. rev., 85 , 166-179. [3] compton, a. (1923). a quantum theory of the scattering of x-rays by light elements. physical review , 21 (5), 483. [4] de broglie, l. (1928). la nouvelle dynamique des quanta. in j. bordet (ed.), cinquième conseil de physique solvay. gauthier-villars, paris. [5] dirac, p. (1927). the quantum theory of the emission and absorption of radiation. proc. roy. soc. lond., 14(767), 243265. [6] dirac, p. (1928). the quantum theory of the electron. proc. roy. soc. lond., a 117 , 610624. [7] dirac, p. (1932). relativistic quantum mechanics. proc. roy. soc. lond., 136(829), 453-464. [8] dürr, d., goldstein, s., munch-berndl, k., & zanghí, n. (1999). hypersurface bohm–dirac models. phys. rev. a, 60 , 2729–2736. [9] dürr, d., goldstein, s., & zanghí, n. (2013). quantum physics without quantum philosophy. new york: springer. [10] dürr, d., & teufel, s. (2009). bohmian mechanics: the physics and mathematics of quantum theory. new york: springer. [11] kiessling, m. k. h., lienert, m., & tahvildar-zadeh, a. s. (2019). a lorentz-covariant interacting electron-photon system in one space dimension. preprint [arxiv:1906.03632] . [12] kiessling, m. k. h., & tahvildar-zadeh, a. s. (2018). on the quantum mechanics of a single photon. j. math. phys., 59 (112302). [13] matlab. (2010). version 7.10.0 (r2010a). natick, massachusetts: the mathworks inc. [14] penrose, r. (2016). fashion, faith, and fantasy in the new physics of the universe. princeton university press. [15] weinberg, s. (1995). the quantum theory of fields, vol. i. cambridge univ. press. aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. observing the demographic factors of peer-nominated leaders in urban middle schools simon daniel, angela w. wang maurice j. elias (faculty advisor) ✵ abstract this study investigated the relationship between adolescent students' gender and racial/ethnic backgrounds and their likelihood of being identified by their peers as having leadership qualities. a survey designed to gauge peer perceptions of leadership qualities was administered to 1003 middle school students from three diverse public middle schools in a northeastern us city. the survey asked students to nominate as many students as possible who possess specific leadership characteristics. female students consistently received more nominations across all survey items at two schools. this pattern was observed for five out of the ten survey items at the third school. at a school with a hispanic majority, hispanic students received more nominations for most survey items than asian, black, and white students. additionally, at a school with a black majority, asian students received more nominations for all survey items compared to black and hispanic students and for nine survey items compared to white students. the results indicate that students' gender and schools' racial/ethnic composition may have some influence on peer perceptions of leadership. furthermore, significant differences in how youths perceive leadership among peers of different backgrounds may be indicative of bias. educators and administrators can use this information to make sure that students from marginalized backgrounds have opportunities to grow as leaders. 1 introduction there is an abundance of programming in public schools that aims to help adolescents develop "leadership" skills, whether through community involvement or engaging with peers. despite the number of programs, there is a considerable amount of uncertainty in scientific literature surrounding adolescents' understanding of leadership and whom they identify as leaders amongst their peers. there is also a lack of literature that explores the demographic characteristics of adolescents identified as leaders by their peers. demographic characteristics, such as racial/ethnic background or gender, have been observed in broader leadership studies and should be an area of interest when studying leadership in adolescents. observable demographic patterns within the body of students identified as leaders can help educators offer more ample leadership opportunities to students of all backgrounds. whitehead (2009) proposes a definition for adolescent leadership with self-authenticity, empathy, trust, and community at its core. this study draws from whitehead's definition and, in addition, incorporates central tenets from social and emotional learning, which include responsible decision-making, emotion regulation, and personal and collective goal-oriented behavior (casel). thus, this paper proposes leadership among adolescents to be perceived through strong interpersonal skills, effective emotion management, goal-oriented behavior, and community involvement. using this definition, a classroom leader among peers could be an individual who is compassionate, helpful, communicative with peers, and involved in their school or community. given that this paper's definition of adolescent leadership consists of multiple dimensions, one approach for investigating the construct is to base literature searches around the various facets defined rather than the construct itself. for example, metzger and ferris (2013) found that in a sample of primarily white adolescents, female students were more likely than male students to find community service and prosocial behaviors more socially desirable.[5] addi aresty rutgers undergraduate research journal, volume i, issue iii tionally, a study of coping strategies among german children and adolescents found that girls were more likely to use problem-solving skills in a stressful environment.[3] though these studies do not explicitly explore the demographic differences of adolescent leadership, their implications are relevant given their areas of investigation. certain demographic variables may be related to how students perceive their classmates, whether in terms of perceived leadership qualities or other favorable traits. for example, jackson et al. (2006) identified the importance of the racial composition of the classroom when considering peer nominations of likability.[4] in their study, african american students received more favorable likability nominations as the african american representation in the classroom increased.[4] a school's racial composition of students can also have an impact on peer relationships. while students in the racial minority may receive more nominations given an increased representation in the classroom, interracial peer relationships seem to be most prevalent in a racially balanced setting.[1] barth et al. (2013) also discussed the importance of considering the nominator's race, as they discovered the presence of a positive ingroup bias for favorable traits and a negative outgroup bias for unfavorable traits.[1] in this case, an ingroup bias for favorable traits would present as nominators assigning favorable traits to those of their race, whereas an out-group bias would present as nominators assigning unfavorable traits to those of other races. there is little dissent that opportunities for young students to develop leadership and socialemotional skills should be maximized. in fact, researchers suggest that developing leadership skills and engaging peer leaders may be an effective way to implement interventions that promote socialemotional skills.[6] however, current literature fails to identify what qualities adolescent students consider to constitute a leader as well as any demographic trends among peers identified as possessing those qualities. this information could be used to develop programs that build on skills that students identify as leadership qualities and identify demographic groups that may benefit from leadership opportunities. research questions and hypotheses the current study aimed to explore the demographic characteristics of adolescents nominated for peer leadership. the current study's hypotheses were formed using findings from previous literature on gender differences[3,5] and the influence of a school's racial/ethnic composition in peer nomination surveys.[1,4] research question 1: how does gender relate to peer perceptions of leadership facets? hypothesis 1-3: female students are more likely to receive nominations for being community-service oriented (1), having problem-solving skills (2), and expressing forgiveness (3) than male students. research question 2: how does each school's racial/ethnic composition relate to the race/ethnicity of the students nominated for leadership? hypothesis 4: students who have greater racial/ethnic representation at their school are more likely to receive nominations than students whose racial/ethnic groups are less represented. 2 methods participants a total of 1003 6th-8th students from three public urban middle schools in a northeastern us city made up the sample of this study. the three schools are referred to as school a, school b, and school c. six cases were removed from the final dataset because they were recorded as having two different school ids. data were collected from fall of 2015. materials the youth leadership survey (yls), developed by the rutgers social-emotional and character development lab, is a nomination survey asking students to nominate as many peers as possible on ten facets identified in the literature as related to leadership[7]. since the survey asks students to identify leadership in others, the yls allows for speculation on peer perceptions of leadership rather than self-ratings of leadership. the ten facets include peer perceptions of being a good leader, being a role model, following through with commitments, making the community better, being rarely upset, demonstrating com aresty rutgers undergraduate research journal, volume i, issue iii passion, having communication skills, having problem-solving skills, demonstrating forgiveness, and being inclusive of others (see table 2). items 4, 8, & 9 on the yls correspond to hypotheses 1, 2, & 3, respectively. the yls demonstrated high internal consistency (10 items; α = .96), meaning that the items in the survey were found to be closely related to each other. procedure the study used data collected as part of a socialemotional and character development curriculum implemented at schools selected for their diversity. the curriculum is designed to help build social-emotional skills, promote youth voice, and develop a positive sense of purpose. participants consented to the study through a passive consent process, in which the participants’ guardians indicated if they did not wish for their student to participate in the study, approved by the school district and the university’s institutional review board. students in participating schools were asked to complete nomination surveys. students were asked to nominate as many peers as possible for each aspect of leadership covered in the survey. independent t-tests and analyses of variance (anovas) were conducted to test for significant differences between gender and racial/ethnic backgrounds, respectively, in relation to the number of peer nominations. additionally, post hoc testing using tukey’s honest significant difference (hsd) test was used to determine which differences between racial/ethnic groups were significant. 3 results table 3 contains frequency data from the yls. overall, the dataset demonstrates a right-skewed distribution. most students received zero nominations, and the frequency of students receiving more than one nomination decreases with the number of nominations received. the first research question asked about the relationship between gender and peer perceptions of leadership. specifically, hypotheses 1-3 proposed that female students were more likely than male stu dents to receive nominations for making the community better, having problem-solving skills, and expressing forgiveness. independent samples t-tests were performed for both the whole sample and for each of the three participating schools. in analyses within school b, within school c, and across the entire sample, female students were more likely to receive nominations for each of the ten facets of leadership proposed in the survey. in school a, female students were more likely to receive nominations for five of the survey items: being a good leader, being a role model, showing compassion, having problemsolving skills, and including others. analyses for the remaining five survey items for school a did not yield significant results (see tables 4-7 for means, standard deviations, and 𝑡𝑡-values). the second research question proposed in this study asked how each school's racial/ethnic composition related to peer nominations of leadership. it was hypothesized that students who had greater racial/ethnic representation at their school were more likely to receive nominations than students who were not as represented. anova testing was performed to identify any significant differences between racial/ethnic groups for each of the survey items. the american indian, multiracial, and pacific islander categories were excluded from the analysis because of the small number of students within those groups. analyses were conducted with the asian, black, hispanic, and white groups for each of the three participating schools. analyses for school a indicated a significant difference in the number of nominations students received for being a good role model across the four racial/ethnic groups [𝐹𝐹(3,300) = 2.81, 𝑝𝑝 = 0.04]. in school a (𝑛𝑛 = 310), black students (𝑛𝑛 = 171) represented the majority, followed by hispanic (𝑛𝑛 = 86) students. post hoc comparisons using tukey’s hsd test indicated that the average number of nominations for hispanic students (𝑀𝑀 = 1.90, 𝑆𝑆𝑆𝑆 = 2.40) was significantly different from that of black students (𝑀𝑀 = 1.21, 𝑆𝑆𝑆𝑆 = 1.64). the means for asian (𝑀𝑀 = 1.67, 𝑆𝑆𝑆𝑆 = 2.22) and white (𝑀𝑀 = 1.00, 𝑆𝑆𝑆𝑆 = 1.86) students were not found to be significantly different (see tables 8a and 8b). aresty rutgers undergraduate research journal, volume i, issue iii analyses for schools b and c found a significant difference in the number of nominations students received across the four racial/ethnic groups for each of the ten survey items. post hoc comparisons for school b (𝑛𝑛 = 311), where hispanic students (𝑛𝑛 = 127) represented the majority, indicated that hispanic students received more nominations than white students for all survey items except the survey item for being rarely upset. the comparisons also revealed that hispanic students received more nominations for being good leaders than black students (see tables 9a and 9b). post hoc comparisons for school c (𝑛𝑛 = 382), where black students (𝑛𝑛 = 169) represented the majority, indicated that asian students (𝑛𝑛 = 47) received more nominations for each survey item except the survey item for being rarely upset, compared to each racial/ethnic group. for the survey item that asked to nominate those who are rarely upset, the number of nominations asian students received was significantly different from black and hispanic students, but not from white students (see tables 10a and 10b). 4 discussion & conclusion this study seeks to identify key demographic characteristics of students nominated by peers for possessing leadership qualities. hypotheses 1-3 proposed that female students would be more likely than male students to receive nominations for making the community better (1), having problemsolving skills (2), and expressing forgiveness (3). these hypotheses were supported by results from school b, school c, and the total sample, while results from school a indicated support for only hypothesis 2. additionally, hypothesis 4 proposed that students whose racial/ethnic backgrounds are more represented in their schools would receive more nominations in general compared to other students. the results from comparisons of nominations between racial/ethnic groups indicated support for hypothesis 4 only in school b's analyses. perhaps the most notable result pertaining to the gender-related hypotheses was the consistency between schools b and c: girls from both schools received more nominations for each survey item compared to boys. all three participating schools had about equal gender distributions, with slightly more male students at each school, which rules out the possibility that female students received more nominations because they made up more of the student population. the results from both schools b and c complement existing literature that reports a gender difference in perceptions of community service[5] and the use of problem-solving skills in stressful situations.[3] however, it is important to note that the current study utilizes nomination data, whereas most of the existing literature assess students individually. nomination surveys allow for speculation on peer perceptions of leadership because they are asking students to nominate peers that possess certain characteristics rather than reflecting on their own traits. in addition, girls from schools b and c appeared to receive more nominations than boys for every survey item. a potential explanation for these nomination patterns is that students from schools b and c may attribute the leadership facets represented in the survey more often towards female peers. in contrast, school a's results indicated that female students received more nominations than male students for only half the survey items. leadership among peers may not be conceptualized the same way in school a as it is in schools b and c, which could explain the difference in nomination patterns. for example, students in school a may attribute some characteristics of leadership to one gender but not other characteristics. existing peer-nomination literature seem to be consistent in their goal of identifying problem behaviors and indicators of aggression among adolescents; however, some studies use nomination data to gauge favorable traits and positive peer relationships. findings from such studies include increased favorable nominations as a function of the student's racial representation in the classroom.[1,4] results from the current study show partial support for what is mentioned in the literature. analyses for school b, which had a hispanic student majority, revealed that hispanic students received more nominations for most survey items compared to white students and for a few survey items compared to black students; however, significant differences were not found for aresty rutgers undergraduate research journal, volume i, issue iii any survey item between hispanic and asian students. in contrast, asian students at school c consistently received more nominations than other racial/ethnic groups for all but one survey item, even though asian students represented the smallest of the four racial/ethnic groups used in the analyses. racial/ethnic representation of the school did not appear to have as strong of an effect on peer perceptions of leadership as was hypothesized, considering that black students did not receive more nominations in their favor at schools a and c, where they experienced the most racial/ethnic representation. it is possible that the diverse makeup of these schools encouraged more relationships among peers from different racial/ethnic backgrounds, as mentioned by barth et al. (2013).[1] in the presence of numerous relationships between racial/ethnic groups, perhaps classmates' background mattered less to students when making nominations. limitations of findings and future directions while demographic categories such as race and ethnicity provide insight into peer perceptions of leadership, it is crucial to consider why such categories may fall short as predictors of youth leadership. ethnic and racial groups are often composed of many subgroups, each of which may have its own view on leadership qualities. for example, there was a significant arab population in school b in this study, but they were categorized as "white" in the school district's system of ethnic identification. the arab subgroup, and subgroups within asian and hispanic categories, could not be identified separately for data analyses. although they may be subtle, differing views within groups may influence a student's decision on which peers they nominate for displaying leadership traits. these differences may be of particular interest for future studies that explore the relationship between race/ethnicity and peer leadership perceptions more narrowly. it is also important to mention that hypotheses favoring male students for certain leadership characteristics could not be made given the sparse literature. additionally, another limitation of the current study is its use of survey data. as the data from this study was obtained through nomination surveys, it is possible that responses could differ based on how the survey was worded and presented. for example, students may feel more inclined to list multiple names for the first few items of the survey as opposed to the last few items. investigators in future studies may wish to apply a correction or adjustment to the data when dealing with unequal ethnic group sizes for more precise results on ethnicity-related hypotheses. future studies may also benefit from running nonparametric analyses, given the positive skew in nominations. nonparametric tests allow for more accurate analyses to be performed on data that do not meet the assumptions for parametric tests (i.e., normally distributed, non-skewed data). however, it is worth noting that nonparametric approaches did not offer any differences in overall findings for the current study. implications the current study has implications for educational practice considering the partial support for its hypotheses. educators who seek to improve leadership among adolescents should remain aware of any demographic patterns in perceptions of peer leadership. if consistent patterns are present, educators can target more opportunities to develop leadership skills toward specific student groups. school professionals may also be able to pair groups of students who lack leadership skills with those who are consistently viewed by their peers as leaders. lastly, findings from the present study reinforce the possibility that students' perceptions of leadership and its many facets can be biased towards specific backgrounds. biased perceptions of leadership may prohibit the development of youth leadership skills in marginalized populations. knowing this, educators may help shape students' perceptions of leadership by continuously emphasizing that the ability to lead is not exclusive to those from certain backgrounds; instead, it is fostered through empathy for, communication with, and commitment to others∎ aresty rutgers undergraduate research journal, volume i, issue iii 5 references [1] barth, j. m., mcdonald, k. l., lochman, j. e., boxmeyer, c., powell, n., dillon, c., & sallee, m. (2013). racially diverse classrooms: effects of classroom racial composition on interracial peer relationships. american journal of orthopsychiatry, 83(2–3), 231–243. [2] casel. (n.d.). what is sel? collaborative for academic, social, and emotional learning (casel). https://casel.org/what-is-sel/ [3] eschenbeck, h., kohlmann, c.-w., & lohaus, a. (2007). gender differences in coping strategies in children and adolescents. journal of individual differences, 28(1), 18–26. [4] jackson, m. f., barth, j. m., powell, n., & lochman, j. e. (2006). classroom contextual effects of race on children's peer nominations. child development, 77(5), 1325–1337. [5] metzger, a., & ferris, k. (2013). adolescents' domain-specific judgments about different forms of civic involvement: variations by age and gender. journal of adolescence, 36(3), 529–538. [6] scharf, m., & mayseless, o. (2009). socioemotional characteristics of elementary school children identified as exhibiting social leadership qualities. the journal of genetic psychology, 170(1), 73–96. [7] whitehead, g. (2009). adolescent leadership development. educational management administration & leadership, 37(6), 847–872. simon daniel is an undergraduate student at rutgers university – new brunswick, currently in his senior year, majoring in psychology and cognitive science and pursuing an undergraduate certificate in data science. simon is a senior and administrative research assistant at the rutgers socialemotional and character development (secd) lab, directed by dr. maurice elias. during his time at the secd lab, he has been awarded the cooper summer undergraduate research fellowship and interdisciplinary research team fellowship awards for his research contributions. simon is also a research assistant at the youth anxiety and depression clinic (yad-c), housed within the graduate school of applied and professional psychology. simon’s research interests include the social-emotional development of youth and the use of prevention science to foster positive mental health outcomes. after graduating from rutgers, simon plans on applying to research-centered doctoral programs in clinical psychology. simon can be contacted at: simon.daniel@rutgers.edu https://casel.org/what-is-sel/ aresty rutgers undergraduate research journal, volume i, issue iii 6 tables category total sample (n=1003) school a (n=310) school b (n=311) school c (n=382) gender male (%) 532 (53) 160 (51.6) 167 (53.7) 205 (53.7) female (%) 471 (47) 150 (48.4) 144 (46.3) 177 (46.3) race/ethnicity american indian (%) 4 (0.4) 1 (0.3) 1 (0.3) 2 (0.5) asian (%) 99 (9.9) 27 (8.7) 25 (8) 47 (12.3) black (%) 413 (41.2) 171 (55.2) 73 (23.5) 169 (44.2) hispanic (%) 306 (30.5) 86 (27.7) 127 (40.8) 93 (24.3) multiracial (%) 3 (0.3) 0 (0) 0 (0) 3 (0.8) pacific islander (%) 9 (0.9) 5 (1.6) 1 (0.3) 3 (0.8) white (%) 169 (16.8) 20 (6.5) 84 (27) 65 (17) free or reduced lunch free (%) 696 (69.4) 222 (71.6) 236 (75.9) 238 (62.3) reduced (%) 51 (5.1) 28 (9) 9 (2.9) 14 (3.7) paid (%) 256 (25.5) 60 (19.4) 66 (21.2) 130 (34) variable name question 1. good leader who do you think is a good leader? 2. role model who acts like a role model for other students? 3. follow through who follows through on things they start? 4. community who wants to make your school and community better? 5. rarely upset who rarely gets upset or angry? 6. compassionate who is compassionate and shows concern for others? 7. communication who communicates well with others? 8. problem solver who is helpful in solving a problem or getting something important done? 9. forgiveness who forgives others easily and does not hold grudges? 10. includes you who includes you in what they are doing? table 1: demographic data of study participants table 2: youth leadership survey questions aresty rutgers undergraduate research journal, volume i, issue iii survey item number of students receiving nominations 0 1 2 3 4 5 6+ good leader 412 162 105 77 57 44 146 role model 438 169 120 68 59 42 107 follow through 430 168 129 96 42 39 99 community 496 178 115 70 40 28 76 rarely upset 401 168 141 104 75 29 85 compassionate 452 198 115 73 48 51 66 communication 446 184 119 85 53 34 82 problem solver 479 180 119 59 56 34 76 forgiveness 461 188 131 73 50 29 71 includes you 403 178 146 94 60 47 75 survey item male students (n=532) female students (n=471) t-test m sd m sd good leader 1.77 2.66 3.25 4.53 -6.22*** role model 1.40 2.10 2.74 3.77 -6.84*** follow through 1.42 2.05 2.43 3.28 -5.74*** community 1.16 1.95 2.09 3.47 -5.12*** rarely upset 1.56 1.91 2.04 2.35 -3.54*** compassionate 1.06 1.66 2.33 2.99 -8.17*** communication 1.26 1.77 2.18 2.84 -6.07*** problem solver 1.09 1.66 2.17 3.06 -6.83*** forgiveness 1.17 1.66 1.93 2.51 -5.58*** includes you 1.38 1.78 2.18 2.45 -5.82*** table 3: nomination frequencies by item for total sample (𝑛𝑛 = 1003) note. the "6+" column represents the frequencies of students receiving six or more nominations. table 4: independent t-test comparing nominations between male and female students (whole sample) note. ∗∗∗ 𝑝𝑝 < .001. aresty rutgers undergraduate research journal, volume i, issue iii survey item male students (n=160) female students (n=150) t-test m sd m sd good leader 1.59 1.93 2.35 3.09 -2.58* role model 1.12 1.49 1.79 2.32 -2.99** follow through 1.10 1.28 1.43 1.94 -1.78 community 0.91 1.37 1.05 1.68 -0.84 rarely upset 1.10 1.38 1.25 1.78 -0.85 compassionate 0.77 1.17 1.17 1.57 -2.56* communication 0.93 1.31 1.04 1.41 -0.74 problem solver 0.79 1.19 1.21 1.88 -2.29* forgiveness 0.75 1.16 0.94 1.36 -1.32 includes you 0.93 1.13 1.25 1.61 -2.07* survey item male students (n=167) female students (n=144) t-test m sd m sd good leader 2.27 2.99 4.22 4.71 -4.29*** role model 1.88 2.31 3.58 3.90 -4.59*** follow through 2.03 2.37 3.15 3.17 -3.49** community 1.76 2.57 3.15 3.30 -4.11*** rarely upset 2.09 2.13 2.67 2.28 -2.31* compassionate 1.58 1.83 3.24 3.14 -5.60*** communication 1.86 2.04 3.16 3.06 -4.33*** problem solver 1.63 2.08 3.00 3.05 -4.55*** forgiveness 1.71 1.93 2.70 2.68 -3.71*** includes you 2.06 2.03 2.93 2.52 -3.32** table 5: independent t-test comparing nominations between male and female students (school a) note. ∗ 𝑝𝑝 < .05. ∗∗ 𝑝𝑝 < .01. table 6: independent t-test comparing nominations between male and female students (school b) note. ∗ 𝑝𝑝 < .05. ∗∗ 𝑝𝑝 < .01. ∗∗∗ 𝑝𝑝 < .001 aresty rutgers undergraduate research journal, volume i, issue iii survey item male students (n=205) female students (n=177) t-test m sd m sd good leader 1.51 2.83 3.24 5.23 -3.93*** role model 1.23 2.26 2.86 4.43 -4.44*** follow through 1.19 2.16 2.69 4.01 -4.46*** community 0.88 1.63 2.10 4.36 -3.53** rarely upset 1.49 1.98 2.21 2.65 -2.95** compassionate 0.87 1.75 2.58 3.45 -5.94*** communication 1.03 1.73 2.35 3.22 -4.89*** problem solver 0.88 1.48 2.32 3.62 -4.94*** forgiveness 1.07 1.64 2.15 2.83 -4.46*** includes you 1.19 1.83 2.35 2.71 -4.83*** table 7: independent t-test comparing nominations between male and female students (school c) note. ∗∗ 𝑝𝑝 < .01. ∗∗∗ 𝑝𝑝 < .001 aresty rutgers undergraduate research journal, volume i, issue iii survey item asian black hispanic white f(3, 300) m sd m sd m sd m sd good leader 2.22 3.22 1.71 2.07 2.43 3.17 1.60 2.74 1.70 role model 1.67 2.22 1.21 1.64 1.90 2.40 1.00 1.86 2.81* follow through 1.15 1.63 1.11 1.41 1.64 1.98 0.90 1.37 2.49 community 1.00 1.24 0.83 1.33 1.23 1.86 1.00 1.84 1.33 rarely upset 1.44 1.65 1.08 1.57 1.31 1.60 0.70 1.13 1.32 compassionate 0.89 1.12 0.87 1.26 1.29 1.71 0.50 1.19 2.61 communication 1.00 1.27 0.92 1.30 1.16 1.56 0.55 0.95 1.30 problem solver 1.48 2.06 0.87 1.24 1.17 1.99 0.65 1.35 1.94 forgiveness 0.85 1.10 0.78 1.19 1.00 1.50 0.50 0.83 1.06 includes you 0.93 1.24 1.01 1.37 1.23 1.47 1.05 1.36 0.62 survey item (i) ethnicity (j) black hispanic white role model asian 0.46 -0.23 0.67 black -0.69* 0.21 hispanic 0.90 table 8a: one-way analysis of variance comparing nominations between ethnic groups (school a) note. ∗ 𝑝𝑝 < .05. table 8b: post hoc comparisons for survey items with significant mean differences (school a) note. mean difference values were calculated as i-j. ∗ 𝑝𝑝 < .05. aresty rutgers undergraduate research journal, volume i, issue iii survey item asian black hispanic white f(3, 305) m sd m sd m sd m sd good leader 4.64 4.86 2.47 3.48 4.06 4.23 2.07 3.39 6.37*** role model 4.96 4.46 2.22 2.86 3.30 3.21 1.46 2.70 10.91*** follow through 4.20 3.32 2.11 2.51 3.09 2.75 1.68 2.71 8.15*** community 3.44 3.78 2.12 2.97 2.90 3.04 1.64 2.54 4.25** rarely upset 3.52 2.18 1.97 1.89 2.65 2.09 1.93 2.51 5.01** compassionate 3.56 3.24 1.79 1.94 2.94 2.80 1.58 2.48 7.73*** communication 2.96 2.62 2.15 2.37 3.10 2.61 1.67 2.70 5.93** problem solver 4.08 2.90 1.79 2.08 2.66 2.60 1.60 2.81 7.86*** forgiveness 3.76 2.65 1.74 2.10 2.60 2.31 1.45 2.22 9.28*** includes you 3.40 1.98 2.53 2.09 2.94 2.48 1.39 1.95 9.96*** survey item (i) ethnicity (j) black hispanic white good leader asian 2.17 0.59 2.57* black -1.59* 0.39 hispanic 1.98** role model asian 2.74** 1.66 3.50*** black -1.08 0.76 hispanic 1.84*** follow through asian 2.09** 1.11 2.52*** black -0.98 0.43 hispanic 1.41** continued on next page table 9a: one-way analysis of variance comparing nominations between ethnic groups (school b) note. ∗∗ 𝑝𝑝 < .01. ∗∗∗ 𝑝𝑝 < .001. table 9b: post hoc comparisons for survey items with significant mean differences (school b) aresty rutgers undergraduate research journal, volume i, issue iii survey item (i) ethnicity (j) black hispanic white community asian 1.32 0.54 1.80* black -0.77 0.48 hispanic 1.26* rarely upset asian 1.55* 0.87 1.59** black -0.68 0.04 hispanic 0.73 compassionate asian 1.77* 0.62 1.98** black -1.15* 0.21 hispanic 1.36** communication asian 0.81 -0.14 1.29 black -0.95 0.48 hispanic 1.44** problem solver asian 2.29** 1.42 2.49*** black -0.87 0.20 hispanic 1.07* forgiveness asian 2.02** 1.16 2.31*** black -0.86 0.29 hispanic 1.15** includes you asian 0.87 0.46 2.01** black -0.40 1.14** hispanic 1.54*** table 9b continued note. mean difference values were calculated as i-j. ∗ 𝑝𝑝 < .05. ∗∗ 𝑝𝑝 < .01 ∗∗∗ 𝑝𝑝 < .0001 aresty rutgers undergraduate research journal, volume i, issue iii survey item asian black hispanic white f(3, 370) m sd m sd m sd m sd good leader 5.70 6.91 1.42 2.43 1.83 2.78 2.80 5.60 14.86*** role model 5.26 5.89 1.24 2.02 1.54 2.45 2.20 4.33 18.93*** follow through 4.83 5.45 1.18 1.77 1.40 2.25 2.26 4.07 18.84*** community 4.09 5.17 0.75 1.40 0.95 1.67 2.00 5.18 15.78*** rarely upset 3.02 2.82 1.49 2.05 1.55 2.02 2.18 2.70 6.55*** compassionate 3.64 4.18 1.18 1.99 1.32 2.10 2.05 3.63 11.05*** communication 3.62 4.04 1.18 1.89 1.29 1.97 1.89 3.12 12.40*** problem solver 3.36 4.31 0.94 1.54 1.37 2.31 2.02 3.74 10.95*** forgiveness 3.19 3.06 1.09 1.71 1.38 1.95 1.98 3.02 11.80*** includes you 3.17 3.10 1.33 1.83 1.51 2.05 1.88 2.75 8.57*** survey item (i) ethnicity (j) black hispanic white good leader asian 4.28*** 3.87*** 2.90** black -0.41 -1.38 hispanic -0.97 role model asian 4.02*** 3.72*** 3.06*** black -0.30 -0.96 hispanic -0.66 follow through asian 3.65*** 3.43*** 2.57*** black -0.21 -1.08 hispanic -0.86 continued on next page table 10a: one-way analysis of variance comparing nominations between ethnic groups (school c) note. ∗∗∗ 𝑝𝑝 < .001. table 10b: post hoc comparisons for survey items with significant mean differences (school c) aresty rutgers undergraduate research journal, volume i, issue iii survey item (i) ethnicity (j) black hispanic white community asian 3.33*** 3.14*** 2.09** black -0.20 -1.25* hispanic -1.05 rarely upset asian 1.53*** 1.47** 0.84 black -0.06 -0.69 hispanic -0.64 compassionate asian 2.46*** 2.32*** 1.59* black -0.15 -0.87 hispanic -0.72 communication asian 2.43*** 2.33*** 1.73** black -0.11 -0.71 hispanic -0.60 problem solver asian 2.42*** 2.00*** 1.35* black -0.43 -1.08* hispanic -0.65 forgiveness asian 2.10*** 1.82*** 1.21* black -0.29 -0.90* hispanic -0.61 includes you asian 1.84*** 1.67*** 1.29* black -0.17 -0.55 hispanic -0.37 table 10b continued note. mean difference values were calculated as i-j. ∗ 𝑝𝑝 < .05. ∗∗ 𝑝𝑝 < .01 ∗∗∗ 𝑝𝑝 < .0001 aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. investigating mimetics of a peptide derived from the effector domain of m a r c k s as possible therapeutics for spinal cord injury monica tschang, melitta schachner (faculty advisor) ✵ abstract like other conditions affecting the central nervous system, spinal cord injury (sci) is difficult to treat with molecular therapies because the bloodbrain barrier makes intravenous treatments largely ineffective. for example, a synthetic peptide chain derived from the effector domain (ed) of myristoylated alanine-rich c-kinase substrate (marcks) has been found to improve functional recovery after sci in female mice; however, peptides do not always pass the blood-brain barrier and are easily degraded due to natural proteases and are excreted during kidney filtration. therefore, the ed peptide cannot access the central nervous system to exhibit its effects if administered intravenously. instead of injecting the ed peptide into the bloodstream, we propose to find compounds that can pass the bloodbrain barrier in place of the ed peptide, improving treatment compatibility. to find such alternatives, we screened compound libraries via competitive enzyme-linked immunosorbent assay (elisa) and identified five potential ed peptide mimetics—compounds that mimic the structure and function of the ed peptide. we then used another competitive elisa to verify their structural similarity to the peptide. after performing toxicity tests to determine the appropriate concentrations of the mimetics to use in functional assays, we found that all five mimetics trigger a significant increase in neurite length in neurons from female mice, but not male mice, when compared to the vehicle control solution. although more functional tests are necessary, these results suggest that these mimetics trigger ed peptide functions and may provide a more efficient treatment alternative for sci. 1 introduction spinal cord injury (sci) is a detrimental condition impacting 39 individuals per million in the usa that results in permanent impairment or loss of motor and neurological function.[5] there is evidence that triggering signal cascades mediates neurite outgrowth, cell migration, and synaptogenesis by targeting function-proactive cell adhesion molecules, leading to an improvement of functional recovery and regeneration after sci.[11] one such cell adhesion molecule, neural cell adhesion molecule (ncam) —a cell-surface molecule known to help neurons stick to other cells or the extracellular matrix— binds to extracellular sugar polymer, polysialic acid (psa) (figure 1).[7] psa has several different functions in the mammalian nervous system, including cell migration, axon growth, central nervous system plasticity, and regeneration—all of which are helpful in sci recovery.[2,21] given the benefits of psa, we were interested in identifying and characterizing its receptors; of these, we chose to further analyze myristoylated alanine-rich c-kinase substrate (marcks) due to its high concentrations in the hippocampus and spinal cord, which could prove helpful in developing treatments targeting sci.[16] the effector domain (ed) on marcks acts as the binding site for psa through the cell, regulat aresty rutgers undergraduate research journal, volume i, issue iii ing the beneficial effects of psa (figure 1).[23] a synthetic peptide (ed peptide) containing the same amino acid residues as the ed of marcks was previously used to verify the marcks-psa interaction, but it was unexpectedly found to also trigger downstream effects on its own. the ed peptide triggers neurite outgrowth of neurons, increases functional recovery, and enhances phosphorylation levels of marcks in neurons and spinal cords in female mice but not male mice.[24] although the ed peptide shows promising effects on sci models, it cannot be administered intravenously because the blood-brain barrier protects the central nervous system from some peptides. to allow for an intravenous treatment option for sci using the marcks-psa pathway, we are looking for ed peptide mimetics—compounds that mimic the structure and function of the ed peptide— that also cross the blood-brain barrier. we hypothesized compounds that are structurally similar to the ed peptide are also functionally similar to the ed peptide. the current study tests this hypothesis by identifying mimetics based on structural similarity to the ed peptide via biochemical assays, then evaluates their effects on neurons to determine their functional similarity to the ed peptide, including sex-specific outcomes. 2 materials & methods elisa screening and verification of ed peptide mimetics to identify potential mimetics, we screened for small organic compounds from the national institute of health clinical collection libraries 1 and 2 (nih ccl1 and nih ccl2) and the natural compound library (ncl) via competitive elisa (figure 2a). onto wells of an elisa plate, we coated 1 µg/ml of marcks antibody (marcks ab, sigma aldrich). as a negative control, we treated two wells with phosphate buffered saline (pbs). wells were then incubated overnight at 4oc and then washed with pbs to discard any loose antibodies that were not attached to the bottom of the well. wells were then treated with blocking solution (5% bovine serum albumin [bsa] + 0.1% triton x-100 in pbs) for one hour at room temperature (rt) to prevent any compounds or reagents from adhering to parts of the well that were not coated with marcks ab. after washing, we added individual compounds from the nih ccl1, nih ccl2, and ncl to the wells and incubated them at rt for 30 minutes. biotinylated ed peptide (genscript) was then mixed with the compounds in each well (final compound concentration = 20 µm; final peptide concentration = 5 µm) for one hour at rt. in this step, compounds compete with the peptide for access to the marcks ab, hence the term “competitive elisa.” after washing, wells were treated with streptavidin coupled to horseradish peroxidase (sa-hrp, jackson immunoresearch. pierce™) at 1:5000 in 5% bsa plus pbs for 30 minutes at rt. streptavidin has a high affinity to biotin, so sa-hrp specifically binds to the biotinylated peptide. afterward, wells were washed, then treated with hrp substrate (0.1% ophenylenediamine dihydrochloride [opd, thermo fisher scientific] + figure 1: schematic of marcks-psa interaction. extracellular psa bound to ncam binds to the effector domain (ed) of intracellular marcks through the cell membrane. when the ed peptide is applied, it competes with endogenous marcks and causes psa to dissociate from marcks, therefore initiating a signal cascade that leads to neurite outgrowth. aresty rutgers undergraduate research journal, volume i, issue iii 0.1% hydrogen peroxide [h2o2] in opd buffer). opd reacts with hrp and h2o2 to produce an orange compound; the more peptide, the more sa-hrp, and the more orange the solution appears. color can be quantified by measuring absorbance at 450 nm with an elisa reader (el800; biotek instruments) after stopping the reaction at 20 minutes with 2.5 m sulfuric acid. from this screen, we identified amiodarone, epigallocatechin, nonyloxytryptamine, sertraline, and tegaserod (sigma aldrich) as potential mimetics. to verify these compounds, we repeated this procedure except we added varying concentrations (0, 1, 5, 10, 25, 50, 100, 200 µm) of the select compounds from the libraries during the competition step to see if they inhibit the binding of the ed peptide to marcks ab in a concentration-dependent manner. mimetics were prepared by dissolving the powdered form in dimethyl sulfoxide (dmso) to make 10 mm stocks which were diluted with pbs to the concentrations listed above. tacrine, a compound found during the screen to not have any effect on macks binding to the marcks ab, was used as a negative control. tacrine also binds to an figure 2: schematic of competitive elisas. a) screening compound libraries for marcks ed peptide mimetics. compounds that are not structurally similar to the ed peptide will not bind to the marcks ab, therefore allowing the ed peptide to bind. conversely, compounds that are structurally similar to the ed peptide will bind to the antibody and block the ed peptide from binding. the peptide is detected via streptavidin bound to horseradish peroxidase and opd. b) verifying that mimetics imitate the ed peptide by seeing if they interact with the ed peptidebinding to site on psa. at high concentrations of the mimetic, more psa will bind to the compound, leaving less free psa to bind to the peptide. at low concentrations of the mimetic, more unbound psa will be able to bind to the immobilized peptide and remain in the wells for detection. psa is detected via anti-ncam primary antibody, secondary antibody coupled to hrp, and opd. aresty rutgers undergraduate research journal, volume i, issue iii antibody that activates another cell adhesion molecule, l1; since l1 and marcks are unrelated, this offers further reason to use tacrine as a negative control.[11] to determine if the verified mimetics also inhibit the interaction between the ed peptide to its binding partner psa, we first incubated varying concentrations (0, 1, 5, 10, 25, 50, 100, 200 µm) of each mimetic with 20 µg/ml of colominic acid (psa, sigma aldrich) at 4oc overnight. then we coated 5 µm biotinylated ed peptide on a streptavidin-coated plate (thermo fisher scientific) for one hour at rt. subsequently, unbound ed peptide was washed off with pbs and the mixed solution of psa and small organic compounds were added to the wells to incubate at rt for one hour (figure 2b). afterward, the wells were washed and incubated with 1:200 of antipsa-ncam (primary antibody 735, thermo fisher scientific) in pbs for one hour at rt. subsequently, wells were washed and incubated with 1:1000 of mouse-hrp (secondary antibody) dissolved in blocking solution at rt for 30 minutes. lastly, wells were washed and treated with hrp substrate. after 10 minutes, we stopped the reaction with sulfuric acid and read absorbance at 450 nm on the elisa reader. cerebellar granule cell culture cerebellar granule cells make up the largest population of neurons in the mammalian brain and are relatively simple in structure, making them an ideal model for in vitro neuronal cultures.[13] to obtain these cells, we removed cerebella from 6 to 8-dayold mice that were euthanized in accordance with the iacuc protocol #: 09-051. to begin the culture, 48-well plates are coated with 0.01% poly-lysine-l (pll) for 30-60 minutes in the incubator (37°c). pll is a chain of synthetic amino acids that helps cells adhere to the bottom of the well-plates. then, pll was aspirated off, and wells were washed with sterile distilled water and left to dry. stocks of medium (containing neurobasal a, b-27, penicillin/streptavidin, bsa, na-pyruvate, l-glutamine, transferrin holo, insulin, l-thyroxine, and na-selenite) were prepared, filtered, and warmed in a 37ºc water bath before use. once cerebella were removed using sharp forceps, the meninges and blood vessels were cleaned off. subsequently, each cerebellum was cut into pieces and put into test tubes of hbss (hank’s balanced salt solution). hbss is a buffer that keeps cells in their preferred ph range and helps maintain high viability during short-term incubations. hbss was then removed via an aspirator, and trypsin was added to the tissue for a 15-minute incubation in a 37ºc water bath. trypsin breaks down adhesive proteins that help cells stick together. cell cultures require seeding individual cells for optimal imaging and even treatment stimulation. subsequently, trypsin was carefully aspirated from each test tube, and tissue was then washed with hbss. dnase was then added to each test tube to digest any dna that leaked from damaged cells, which can form aggregates and hinder dissociation. cells were then dissociated via pipette until there were no visible pieces of cerebella left. dissociated cells were incubated with the dnase solution for 5 minutes in a 37ºc water bath and then diluted with hbss before centrifugation at 200 g for 5 minutes at 4ºc. centrifugation is a separation technique that uses centrifugal force to collect heavier particles in a solution (in this case, the cells) at the bottom of a container, therefore creating a cell pellet. afterward, hbss was aspirated off, and the cell pellet was resuspended in a pre-warmed medium. cell solution was combined with fluorophores (acridine orange and propidium iodide, nexcelom bioscience) that stain dead and live cells different colors to count the cells. initial viability and concentration of live cells were measured using the cellometer auto 2000 cell viability counter. cells from males and females were measured and cultured separately. toxicity tests the cerebellar granule cell culture procedure detailed above was carried out, separating male and female mice. after counting cells, cells from each sex were separately diluted to 500,000 cells/ml with a pre-warmed medium. cells were seeded into 24 wells (12 for male cells and 12 for female cells) on 0.01% pll-coated 48-well plates with 250 µl of 500,000 cells/ml solution in each well and maintained at 37°c for 24 hours. afterward, cerebellar granule cells were incubated with 0.1% dmso as a aresty rutgers undergraduate research journal, volume i, issue iii vehicle control, 1, 10, and 100 µm of the mimetic for another 24 hours. for live imaging, propidium iodide and calcein-am solution (each 4 ng/ml) were first added to each well, and the plate was then incubated (37°c) for 20-30 minutes before using the fluorescent microscope. axiovision was used to control contrast and adjust the clarity of the images. four images were taken per well. the experiment was performed three times in triplicate for each compound (𝑛𝑛 = 36). neurite outgrowth the cerebellar granule cell culture procedure detailed above was carried out, subsequently diluting cells from each sex to 100,000 cells/ml with a prewarmed medium. cells were seeded into 4 wells per treatment (two for males and two for females) on 0.01% pll-coated 48-well plates. cells were immediately stimulated with 0.1% dmso for final concentrations in the wells of 0, 0.1, 1, 10, 100, and 1000 nm of each mimetic diluted in dmso. one row of wells was left untreated to see if dmso would affect neurite outgrowth. plates were then incubated for 24 hours at 37°c. each well was treated with 2.5% glutaraldehyde to fix cells and left to incubate for 30 minutes at rt. subsequently, the wells were washed with water. staining solution (1% toluidine blue o, 1% methylene blue in 1% borax) was added to the wells and left to incubate at rt for 30 minutes. wells were then washed with water, and the plate was left to dry so images could be taken. images were taken using axiovision. neurites were traced and measured using imagej. statistical analysis statistical comparisons between groups were performed by one-way anova test using statview for all experiments because each had more than two experimental groups. subsequently, fischer’s plsd post-hoc test is used to determine the statistically significant variance between groups. according to fischer’s plsd test, asterisks in figures signify statistically significant differences between groups wherein, ∗ 𝑝𝑝 < 0.05, ∗∗ 𝑝𝑝 < 0.01, ∗∗∗ 𝑝𝑝 < 0.001. every experiment was performed three independent times and the averages and standard errors of the mean were calculated from all trials on microsoft excel. schematics, graphs, and calculations were done with microsoft office. 3 results five compounds inhibit the binding of ed peptide to the marcks antibody our results show that epigallocatechin, amiodarone, sertraline, tegaserod, and nonyloxytryptamine inhibit the binding of the ed peptide to marcks ab in a concentration-dependent manner. at 25 µm and higher concentrations, all the mimetics inhibit the ed peptide from binding to the antibody significantly more than our negative control, tacrine, another small organic compound from the nih ccl1 (figure 3). ed peptide mimetics bind to psa to further verify the integrity of the structural similarity between the mimetics and the ed peptide, we conducted a competitive elisa to see if the mimetics could inhibit the binding of psa to the ed peptide (figure 2b). from this procedure, we determined that the compounds inhibit the binding of the ed peptide to psa in a concentration-dependent manner, and all compounds significantly inhibit the binding of psa to marcks at six concentrations or more (figure 4). the five mimetics are nontoxic in cerebellar granule cells at 1µm before determining if the identified mimetics trigger ed peptide-mediated functions, we determined the cytotoxicity of the mimetics. in all five marcks ed peptide mimetics, concentrations of 1 µm are nontoxic for cerebellar granule cells in both sexes. a 10 µm concentration of epigallocatechin gallate, is also nontoxic (figure 5). we define the toxic range to be any percent viability equal to or less than 50%. there is no apparent sex-specific difference in toxic ranges shown in the data. based on the data shown in figure 5, we suggest using a maximum of 1 µm concentration in live functional assays. therefore, we chose seven concentrations ranging from 0 µm (dmso) to 1 µm of each compound to stimulate the cerebellar granule cells in our neurite outgrowth assays. aresty rutgers undergraduate research journal, volume i, issue iii figure 3: verifying potential ed peptide mimetics. verifying epigallocatechin gallate, amiodarone, sertraline, tegaserod, and 5-nonyloxytryptamine as potential ed peptide mimetics. error bars indicate standard error of the mean. (three independent trials were carried out in triplicate, 𝑛𝑛 = 9). asterisks signify statistically significant differences between the identified mimetics and the negative control tacrine wherein ∗ 𝑝𝑝 < 0.05, ∗∗ 𝑝𝑝 < 0.01, ∗∗∗ 𝑝𝑝 < 0.001. (one-way anova, 𝐹𝐹 = 63.39, 𝑝𝑝 < 0.0001). figure 4: determining if mimetics inhibit the binding of psa to ed peptide. determining if epigallocatechin, sertraline, amiodarone, tegaserod, and nonyloxytryptamine inhibit psa from binding to ed peptide. error bars indicate standard error of the mean. (three independent trials were carried out in triplicate, 𝑛𝑛 = 9). asterisks signify statistically significant differences between the identified mimetics and the negative control tacrine wherein ∗ 𝑝𝑝 < 0.05, ∗∗ 𝑝𝑝 < 0.01, ∗∗∗ 𝑝𝑝 < 0.001 (one-way anova, 𝐹𝐹 = 12.512, 𝑝𝑝 < 0.0001). aresty rutgers undergraduate research journal, volume i, issue iii the five mimetics promote neurite outgrowth in female cerebellar granule cells at nanomolar concentrations our first functional test was to determine whether ed peptide mimetics promote neurite outgrowth in female cerebellar granule cells and not males, as the ed peptide does. all five mimetics significantly stimulated neurite outgrowth at 0.01-1000 nm, except for amiodarone which promoted neurite outgrowth at only 1-1000 nm (figure 6). in addition, neurite outgrowth was observed only in female cerebellar granule cells, supporting the discovery that marcks promotes neurite outgrowth in a sex-specific manner. there was no significant difference in neurite length between dmso-treated and mimetic-treated male cerebellar granule cells (figure 6). nonyloxytryptamine was the most effective mimetic, promoting the most neurite outgrowth at 10000 nm with an average neurite length of 48.6 µm (sem 1.85 µm) compared with the vehicle control, dmso, with an average neurite length of 31.4 µm (sem 0.77 µm). epigallocatechin, sertraline, and tegaserod were also effective compounds in promoting neurite outgrowth with average neurite lengths of 46.5 µm (sem 1.43), 42.2 µm (sem 1.38), and 41.8 µm (sem 1.27) at 1000, 100, and 1000 nm, respectively. amiodarone was the least effective mimetic in promoting neurite outgrowth with an average length of 41.5 µm (sem 1.36) at 100 nm (figure 6). 4 discussion our results show that nonyloxytryptamine, epigallocatechin, sertraline, tegaserod, and amiodarone mimic both the structure and function of the ed peptide. we determined structural similarity with competitive elisas and functional similarities by measuring neurite outgrowth. the identified ed peptide mimetics may prove to be better treatment options because they are not as easily degraded or blocked by the bloodbrain barrier as the ed peptide. additionally, compounds from the libraries that we screened have a history of use in human clinical trials and sci studies. figure 5: measuring neurotoxicity of marcks ed peptide mimetics in cerebellar granule cells. cerebellar granule cells were dissociated from 6 to 8day-old mice, separated by sex and seeded onto pll-coated plates. after 24 hours, cells were treated with dmso or 1 µm, 10 µm, or 100 µm of each mimetic. cell viability was measured by propidium iodide and calcein-am staining to take the ratio of live cells to total cells. the bar graphs show the average cell viability and error bars indicate standard error of the mean. (four images were taken per well. three independent trials were carried out in triplicate, 𝑛𝑛 = 36). aresty rutgers undergraduate research journal, volume i, issue iii figure 6: measuring neurite outgrowth of cerebellar granule cells treated with different concentrations of ed peptide mimetics. cerebellar granule cells were dissociated from 6 to 8day-old mice, separated by sex and seeded onto pll-coated plates. cells were immediately treated with 0.1% of dmso and each mimetic with final concentrations of 0 nm, 0.1 nm, 1 nm, 10 nm, 100 nm and 1000 nm. the bar graphs show the average neurite length and error bars indicate standard error of the mean (100 neurites were measured in each treatment per trial. three independent trials were carried, 𝑛𝑛 = 300). asterisks signify statistically significant differences between the identified mimetics and dmso wherein ∗ 𝑝𝑝 < 0.05, ∗∗ 𝑝𝑝 < 0.01, ∗∗∗ 𝑝𝑝 < 0.001. (nonyloxytryptamine, oneway anova, 𝐹𝐹 = 21.11, 𝑝𝑝 < 0.0001; epigallocatechin, one-way anova, 𝐹𝐹 = 18.39, 𝑝𝑝 < 0.0001; sertraline, one-way anova, 𝐹𝐹 = 13.064, 𝑝𝑝 < 0.0001; tegaserod, one-way anova, 𝐹𝐹 = 10.789, 𝑝𝑝 < 0.0001; amiodarone, one-way anova, 𝐹𝐹 = 11.968, 𝑝𝑝 < 0.0001). aresty rutgers undergraduate research journal, volume i, issue iii two of the other verified marcks ed peptide mimetics, tegaserod, and nonyloxytryptamine, are also psa mimetics that facilitate nervous system repair and improve recovery after sci in mice.[4,15] epigallocatechin gallate (egcg) is an active compound found in green tea that has been shown to pass through the blood-brain barrier.[9] in some regards, egcg appears to be functionally similar to the ed peptide in that egcg improves functional recovery after chronic and acute sci and sciatic nerve injury.[12,18,19,25] additionally, the ed peptide and egcg increase the phosphorylation levels of marcks in neurons.[6,27] egcg also has neuroprotective, anti-inflammatory, and anti-edema effects after spinal cord and sciatic nerve injury and attenuates neuropathic pain.[1,3,8,14,20,22,26,28] while we have found that the ed peptide improves functional recovery after sci in female mice, the molecular mechanisms underlying sexspecific responses remain unclear. therefore, ed peptide mimetics are not only beneficial in potentially substituting the peptide as a form of treatment but might offer insight into how the ed peptide promotes neurite outgrowth and improves functional recovery in female mice. in the future, we will test if the mimetics enhance marcks phosphorylation levels specifically in female but not male cerebellar granule cells. we predict that the ed peptide mimetics will be effective replacements for the ed peptide in recovery from trauma not only in sci but also in other model systems of trauma in mammals, fundamentally improving recovery response∎ 5 acknowledgements special thanks to dr. thomas theis and dr. melitta schachner for their mentorship, support, and guidance. additionally, many thanks to the w.m. keck center for collaborative neuroscience for the facilities and resources. this paper contains data submitted in several progress reports submitted to the department of cell biology and neuroscience at rutgers university’s school of arts and sciences. the new jersey commission for spinal cord research funded this work (cscr18erg015). 6 references [1] álvarez-pérez, b., homs, j., bosch-mola, m., puig, t., reina, f., verdú, e., & boadas-vaello, p. (2016). epigallocatechin-3gallate treatment reduces thermal hyperalgesia after spinal cord injury by down-regulating rhoa expression in mice. european journal of pain, 20(3), 341-352. [2] bonfanti, l. (2006). psa-ncam in mammalian structural plasticity and neurogenesis. progress in neurobiology. 80(3),129164. [3] bosch-mola, m., homs, j., álvarez-pérez, b., puig, t., reina, f., verdú, e., & boadas-vaello, p. (2017). (-)-epigallocatechin3-gallate antihyperalgesic effect associates with reduced cx3cl1 chemokine expression in spinal cord. phytotherapy research, 31(2), 340-344. [4] bushman j., mishra b., ezra m., gul s., schulze c., chaudhury s., ripoll d., wallqvist a., kohna j., schachner m., & loers g. (2014). tegaserod mimics the neurostimulatory glycan polysialic acid and promotes nervous system repair. neuropharmacology, 79, 456-466. [5] chen, y., tang, y., vogel, l.c., & devivo, m.j. (2013). causes of spinal cord injury. topics in spinal cord injury rehabilitation, 19(1), 1-8. [6] chou, c.w., huang, w.j., tien, l.t., & wang, s.j. (2007). (-)epigallocatechin gallate, the most active polyphenolic catechin in green tea, presynaptically facilitates ca2+-dependent glutamate release via activation of protein kinase c in rat cerebral cortex. synapse, 61(11), 889-902. [7] finne, j., finne, u., deagostini-bazin, h., & goridis, c. (1983). occurrence of α2–8 linked polysialosyl units in a neural cell adhesion molecule. biochemical and biophysical research communications, 112(2), 482–487. [8] ge, r., zhu, y., diao, y., tao, l., yuan, w., & xiong, x.c. (2013). anti-edema effect of epigallocatechin gallate on spinal cord injury in rats. brain research, 1527, 40-46. [9] giunta, b., hou, h., zhu, y., salemi, j., ruscin, a., shytle, r. d., & tan, j. (2010). fish oil enhances anti-amyloidogenic properties of green tea egcg in tg2576 mice. neuroscience letters, 471(3), 134-138. [10] jain, n.b., ayers, g.d., peterson, e.n., harris, m.b., morse, l., o’connor, k.c., & garsick, e. (2015). traumatic spinal cord injury in the united states, 1993-2012. jama, 313(22), 22362243. [11] kataria, h., lutz, d., chaudhary, h., schachner, m., & loers, g. (2016). small molecule agonists of cell adhesion molecule l1 mimic l1 functions in vivo. molecular neurobiology, 53, 4461–4483. [12] khalatbary, a.r., tiraihi, t., boroujeni, m.b., ahmadvand, h., tavafi, m., & tamjidipoor, a. (2010). effects of epigallocatechin gallate on tissue protection and functional recovery after contusive spinal cord injury in rats. brain research, 1306(8), 168-175. [13] krämer, d., & minichiello, l. (2010). cell culture of primary cerebellar granule cells. methods in molecular biology (clifton, n.j.), 633, 233–239. aresty rutgers undergraduate research journal, volume i, issue iii [14] kuang, x., huang, y., gu, h.f., zu, x.y., zou, w.y., song, z.b., & guo, q.l. (2012). effects of intrathecal epigallocatechin gallate, an inhibitor of toll-like receptor 4, on chronic neuropathic pain in rats. european journal of pharmacology, 676(13), 51-56. [15] loers, g., saini, v., mishra, b., papastefanaki f., lutz d., chaudhury s., ripoll d.r., wallqvist a., gul s., schachner m., & kaur g. (2014). nonyloxytryptamine mimics polysialic acid and modulates neuronal and glial functions in cell culture. journal of neurochemistry, 128(1), 88-100. [16] mcnamara, r.k., & lenox, r.h. (1998). distribution of the protein kinase c substrates marcks and mrp in the postnatal developing rat brain. journal of comparative neurology, 397(3), 337-356. [17] michon, f., brisson, j.r., & jennings, h.j. (1987). conformational differences between linear α (28)-linked homosialooligosaccharides and the epitope of the group b meningococcal polysaccharide. biochemistry, 26(25), 8399-8405. [18] renno, w. m., al-khaledi, g., mousa, a., karam, s. m., abul, h., & asfar, s. (2014). (-)-epigallocatechin-3-gallate (egcg) modulates neurological function when intravenously infused in acute and, chronically injured spinal cord of adult rats. neuropharmacology, 77, 100–119. [19] renno, w. m., benov, l., & khan, k. m. (2017). possible role of antioxidative capacity of (-)-epigallocatechin-3-gallate treatment in morphological and neurobehavioral recovery after sciatic nerve crush injury. journal of neurosurgery. spine, 27(5), 593–613. [20] renno, w.m., khan, k.m., & benov, l. (2016). is there a role for neurotrophic factors and their receptors in augmenting the neuroprotective effect of (-)-epigallocatechin-3-gallate treatment of sciatic nerve crush injury? neuropharmacology 102, 1-20. [21] rutishauser u. (2008). polysialic acid in the plasticity of the developing and adult vertebrate nervous system. nature reviews. neuroscience, 9(1), 26–35. [22] ruzicka, j., urdzikova, l. m., svobodova, b., amin, a. g., karova, k., dubisova, j., zaviskova, k., kubinova, s., schmidt, m., jhanwar-uniyal, m., & jendelova, p. (2018). does combined therapy of curcumin and epigallocatechin gallate have a synergistic neuroprotective effect against spinal cord injury?. neural regeneration research, 13(1), 119–127. [23] theis, t., mishra, b., von der ohe, m., loers, g., prondzynski, m., pless, o., blackshear, p. j., schachner, m., & kleene, r. (2013). functional role of the interaction between polysialic acid and myristoylated alanine-rich c kinase substrate at the plasma membrane. the journal of biological chemistry, 288(9), 6726–6742. [24] theis, t., kumar, k., wei, e., nguyen, j., glynos, v., paranjape, n., askarifirouzjaei, h., khajouienejad, l., berthiaume, f., young, w., & schachner, m. (2020). myristoylated alanine-rich c-kinase substrate effector domain peptide improves sex specific recovery and axonal regrowth after spinal cord injury. faseb j, 34(9), 12677-12690. [25] tian, w., han, x. g., liu, y. j., tang, g. q., liu, b., wang, y. q., xiao, b., & xu, y. f. (2013). intrathecal epigallocatechin gallate treatment improves functional recovery after spinal cord injury by upregulating the expression of bdnf and gdnf. neurochemical research, 38(4), 772–779. [26] urdzikova, l.m., ruzicka, j., karova, k., kloudova, a., svobodova, b., amin, a., dubisova, j., schmidt, m., kubinova, s., jhanwar-uniyal, m., & jendelova, p. (2017). a green tea polyphenol epigallocatechin-3-gallate enhances neuroregeneration after spinal cord injury by altering levels of inflammatory cytokines. neuropharmacology, 126, 213-223. [27] westphal, n., loers, g., lutz, d., theis, t., kleene, r., & schachner, m. (2017). generation and intracellular trafficking of a polysialic acid-carrying fragment of the neural cell adhesion molecule ncam to the cell nucleus. scientific reports, 7, 8622. [28] xifró, x., vidal-sancho, l., boadas-vaello, p., turrado, c., alberch, j., puig, t., & verdú, e. (2015). novel epigallocatechin-3-gallate (egcg) derivative as a new therapeutic strategy for reducing neuropathic pain after chronic constriction nerve injury in mice. plos one, 10(4), e0123122. monica tschang (she/her) is a senior majoring in cell biology and neuroscience in the school of arts and sciences. to explore her interests in the gut-brain axis and mental health advocacy, she is also double minoring in psychology and nutrition. with this undergraduate background, she is applying to neuroscience phd programs to study the neurobiological basis of mental health disorders. she started research for the marcks ed peptide mimetic project in the beginning of 2019 under dr. melitta schachner at the w.m. keck center for collaborative neuroscience and will continue this project for her honors thesis. for questions, feel free to contact her at: mat361@scarletmail.rutgers.edu aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. characterizing and evaluating the diverse microbial communities and their mercury resistance potential from hot spring sites representing gradients in temperature and ph in yellowstone national park yelizaveta rassadkina, spencer roth, and tamar barkay department of biochemistry and microbiology, rutgers university abstract yellowstone national park is home to many different hot springs, lakes, geysers, pools, and basins that range in ph, chemical composition, and temperature. these different environmental variations provide a broad range of conditions that select and grow diverse communities of microorganisms. in this study, we collected samples from geochemically diverse lakes and springs to characterize the microbial communities present through 16s rrna metagenomic analysis. this information was then used to observe how various microorganisms survive in high mercury environments. the results show the presence of microorganisms that have been studied in previous literature. the results also depict gradients of microorganisms including thermophilic bacteria and archaea that exist in these extreme environments. in addition, beta diversity analyses of the sequence data showed site clustering based primarily on temperature instead of ph or sample site, suggesting that while ph, temperature, and sample site were all shown to be significant, temperature is the strongest factor driving microorganism community development. while it is important to characterize the microorganism community present, it is also important to understand how this community functions as a result of its selection. along with looking at community composition, genomic material was tested to see if it contained mercury methylating (hgca) or mercury reducing (mera) genes. out of 22 samples, three of them were observed to have mera genes, while no samples had hgca genes. these results indicate that microorganisms in mustard and nymph springs may use mercury reduction. understanding how microorganisms survive in environments with high concentrations of toxic pollutants is crucial because it can be used as a model to better understand mechanisms of resistance and the biogeochemical cycle, as well as for bioremediation and other solutions to anthropogenic problems. key terms chemolithotrophs, mercury methylation, mercury reduction, thermophiles. https://creativecommons.org/licenses/by-nc-sa/4.0/ https://creativecommons.org/licenses/by-nc-sa/4.0/ aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 2 introduction yellowstone national park is well known for its unique aquatic geothermal habitats, which exhibit a wide array of extreme temperatures and ph levels. extreme conditions select for lithotrophic extremophiles, organisms that obtain reducing equivalents from inorganic substrates (cuhel et al., 2002). chemolithotrophs are a specific group of lithotrophic organisms that use inorganic substrates for atp production, as well as inorganic material for carbon compound biosynthesis such as co2 fixation (cuhel et al., 2002). these groups of organisms belong to the archaeal and bacterial domains and are dominant in the extreme environments common in yellowstone national park (cuhel et al., 2002). yellowstone’s geothermal activity leads to the presence of hydrothermal springs, which emit compounds such as sulfide, elemental sulfur, arsenic, and iron (shock et al., 2010). these compounds can then be used as a source of energy for local microorganisms. since most of the microorganisms in hot springs are chemolithotrophic thermophiles, they require specific nutrients and extreme conditions that are not easily replicated in a laboratory setting. therefore, the ecology of microbial communities within yellowstone is not as wellcharacterized as communities from many mesophilic environments (reysenbach et al., 2000). this is why we used a metagenomic sequencing approach—rather than cultures—to better understand the target microorganism community. it is important to understand which microbial guilds dominate these environments, how they live and interact with each other, and how selective environmental pressures influence these interactions. the environmental pressures of hot springs include a range of moderate to high temperatures, extreme ph ranges, and several toxic compounds. mercury (hg) is one such compound that has been observed at measurable concentrations in ~20% of geothermal sites in western north america (geesey et al., 2016). hg is a pollutant and a toxic metal present throughout many natural environments, such as hot springs and sediments. due to its toxicity, hg applies selective pressure on local organisms, especially in high concentrations. mercury-resistant organisms can be found in some geothermal sites in north america, such as yellowstone national park (geesey et al., 2016). microbes that can tolerate high concentrations of hg often have genes that allow them to reduce hg(ii) to its volatile form, hg(0). some prokaryotes have the mer operon, associated with mercury reduction, which encodes for mercury transport and transformations (barkay and wagner-dobler, 2005). there are many different mer genes that participate in the uptake of hg(ii) into the cell. after being reduced, hg(0) is released into the atmosphere in its gaseous state (barkay and wagner-dobler, 2005). it is crucial to understand how organisms withstand toxic pollutants in extreme conditions within a properly functioning ecosystem. in this study, we aim to characterize the microbial community in yellowstone national park and identify the mechanisms of mercury resistance that allow them to survive in these geothermally aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 3 diverse hot springs. for this purpose, dna sequencing was used to identify the mercury reducing organisms or mercury methylating organisms present in the community. in addition, diversity metrics were used to determine if the physical and chemical properties of the spring environment correlate with the structure and function of their microbial communities. our study differs from previous literature in that we have a ph and temperature range as well as a matrix range. this allows us to study the differences between different types of samples taken from the same site, i.e. scum vs. sediment from the sample site. we used dna sequencing and diversity metrics to broaden our understanding of the structure and function of the microbial community in the hot springs of yellowstone national park. this information can help us better understand the impact of microorganism communities and their diversity in nature. methods geochemistry the conditions of the hot springs (i.e. temperature and ph) from which the samples were extracted were recorded by dr. joann holloway and team. sample description the samples were collected in yellowstone national park in september 2017 by dr. joann holloway (usgs). samples were taken from upper geyser basin, nymph basin, norris geyser basin, mud volcano, artist paint pots basin, gibbon canyon, and midway geyser basin. samples were then stored in lifeguard solution (qiagen). for each sample, approximately one gram was collected and placed in a tube with lifeguard solution. scum and water samples were each collected with a sterile syringe and then filtered. filters were placed into a lifeguard solution and then stored on ice. samples were shipped to rutgers university and stored at -80°c until further analysis. extraction dna was extracted using the mobio powerlyzer powersoil dna isolation kit (qiagen) following the manufacturer’s protocol. this kit is intended to be used for the isolation of microbial nucleic acids from soil, environmental samples, and tough microbes. it uses bead-based homogenizers to eliminate humic substances and other inhibitors, which allows for downstream application. dna extracts were stored at -20°c following extraction. pcr pcr (polymerase chain reaction) was performed after every extraction to amplify the 16s rrna gene from the extracts to determine the presence of bacterial and/or archaeal dna. for pcr products obtained with gotaq green master mix (promega), a 515 forward primer and 806 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 4 reverse primer 16s v4 primer set was used (parada et al., 2016; apprill et al., 2015). products of pcr were separated by 1% agarose gel electrophoresis. dna sequencing the v4 region of the 16s rrna gene was sequenced using the illumina miseq platform following the earth microbiome project at the rutgers university genome cooperative (ul-hasan et al., 2019). hgca detection pcr was performed on every sample to see if hgca genes were present within the extracted dna. hgca was targeted using a 261 forward primer and a 912 reverse primer set as previously published (schaefer et al, 2014). products of pcr were separated by 2% agarose gel electrophoresis. mera detection pcr was performed on every sample to see if mera genes were present within the extracted dna. the pcr products were obtained with gotaq green master mix, a mera2 forward primer and a mera2 reverse primer set was used (poulain et al., 2015). products of pcr were separated by 2% agarose gel electrophoresis. microbial ecology analysis (figure 1) the bioinformatics platform for quantitative insights into microbial ecology 2 (qiime2) was used because it includes sequence alignments, quality filtering, phylogeny building, and taxonomic classification (hall et al., 2018). qiime2, as opposed to qiime, allows for microbial marker gene analysis to better quantify and compare the extracted genomic material (hall et al., 2018). for data analysis, the divisive amplicon denoising algorithm 2 (dada2) pipeline was used to detect and correct the processed sequences (callahan et al., 2016). alpha rarefaction and beta diversity metrics were all performed and analyzed (anderson, 2001). taxonomy was assigned using a naïve-bayesian classifier trained on the v4 region of the silva-132 99% 16s database (pedregosa et al., 2011) then, a feature table and a feature data summary were made after sequence filtering. lastly, a taxonomic bar plot was created along with beta diversity metrics such as weighted unifrac, which qualitatively measures community dissimilarity (bolyen et al., 2019; anderson, 2001). aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 5 results samples were taken from various geothermally diverse sediments, waters, mats, and scum sites. a total of twenty-two samples were taken from distinct environments that ranged in ph from 1.71 – 6.65 and in temperature from 54.2°c – 87.8°c. these twenty-two samples came from various springs, geysers, pools, and basins located in yellowstone national park (figure 2). several sample matrices were taken in duplicate. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 6 after collection, the twenty-two samples were sent for sequencing analysis after completion of the genomic extractions. eleven out of twenty-two samples were retained after rarefication at 500 or more sequence reads (table 1). the other eleven samples had very small amounts of biomass and thus resulted in a low number of reads (<500), so these samples were excluded from further analyses. the samples that had a low number of reads consisted mainly of water and scum samples from sample matrices. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 7 weighted unifrac analysis was performed to observe how the various sample matrices cluster together based on several environmental factors that drive community selection. the three main selection factors that were investigated were temperature, ph, and sample matrix. these environmental conditions were tested to isolate the primary factor that influenced sample clustering. the principle coordinate analysis (pca) plot showed that the samples exhibited clustering based on sample matrix (figure 3a). one cluster was composed of sediment and outlet sediment samples, while the other was composed of sediment and water samples. the mat and scum samples, both from mud volcano, did not cluster together with the other samples. in addition, the pool edge sediment from the crystal sister west sample was shown to be distinct from the other sample matrices. the weighted unifrac beta diversity analysis showed that the microbial communities in yellowstone national park clustered primarily based on temperature, rather than ph, which had no clustering pattern (figures 3b and 3c). microbial communities from samples that had a temperature of 50-60℃ tended to cluster together, while communities that were taken from sampling sites with a temperature of 70-90℃ tended to cluster together, with the exception of crystal sister west sediment. beta diversity dissimilarity results showed that ph accounts for 13.9% (p = 0.006) of variance, while temperature accounts for 22.7% (p < 0.001) and sample matrix accounts for 38.1% (p = 0.009). aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 8 sequence taxonomy was assigned to determine the major taxa present in the samples. the archaeal order thermoproteales was the most abundant throughout the samples, except for at the mud volcano mat and nymph water sites, where it was not detected (figure 4). the second most dominant order was candidatus micrarchaeum, which was present in seven of the samples. in addition, the most dominant bacterial taxon was acetobacterales, while the most dominant archaeal taxon was desulfurococcales. overall, there were more archaea than bacteria present in physiologically extreme conditions. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 9 within the eleven samples, those that had more than a 50% relative frequency of archaea included mud volcano mat, scum, pool edge sediment, and water; sediments from site samples crystal sister e, nymph, crystal sister w, mud volcano; and outlet sediments from gibbon canyon. the samples dominated by bacteria (more than 50% relative frequency) were mud volcano scum and mat. within mud volcano, samples collected from the sediment were dominated by archaea, while the mat and scum samples were dominated by bacteria. candidatus micrarchaeum was abundant in samples with a ph range of 1.0-2.32. desulfurococcales and thermoproteales were present in the ph range of 1.71-6.65. archaeal 16s rrna reads were detected in all samples with ph at or below 6.65. lastly, all of the samples were also tested for mercury reduction or mercury methylation genes. the gene encoding a mercuric reductase, mera, was detected in two samples from mustard spring and one sample from nymph. mustard spring has a ph of 8.19 with a temperature of 86.2 ºc, while nymph spring has a ph of 2.32 and a temperature of 59.5 ºc. no copies of hgca, a gene responsible for hg methylation, were detected in the samples. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 10 discussion/conclusion sequences of 16 rrna gene pcr products were obtained from most of the samples that were extracted. samples that had over one thousand sequence reads were mostly sediment samples. fewer reads, above one hundred but below one thousand, were obtained from the mats, water, and scum samples. in addition, the samples excluded from the pca plot had very low number of reads (<500 reads) such as mat and water samples. this is not that surprising because (i) samples were very small—yielding little biomass in all but the sediments and extreme biomass scarcity in water samples—and (ii) the dna extraction kit was most suited for soil and sediment samples. the beta diversity analysis showed that the microbial communities in yellowstone national park cluster primarily by temperature rather than ph. this matches previous literature that states that the main selection factor in hot springs tends to be temperature (uribe-lorio et al., 2019). beta diversity dissimilarity results showed statistical significance of sample separation based on ph, temperature, and sample type. the weighted unifrac pca plot, which measures community dissimilarity while incorporating the relative abundance of the organisms, showed two distinct groups clustered based on the temperatures of the sites (lozupone et al., 2011). in addition to these two grouping clusters, the results also showed a dominance of archaeal species. this is likely due to the extreme thermophilic and ph conditions which allow archaea to outcompete bacteria in these niches (reed et al., 2013). archaea are known extremophiles that have evolved specific proteins that allow them to proliferate in extreme environments and tolerate normally toxic compounds, such as arsenic and mercury (reed et al., 2013). it is not surprising that they dominate these communities in the extreme conditions of some yellowstone springs. in addition to temperature, the microbial taxa also clustered based on sampling matrix. clusters for sediments and outlet sediments, pool edge sediments, water, and scum samples were all observed. the weighted unifrac plot showed two distinct clustering patterns between the different sample matrices. environmental conditions in the different matrices are expected to vary. sediments are expected to contain less dissolved oxygen, less uv radiation, more sulfide, more arsenic, and higher mercury concentrations compared to water, mat, and scum samples, which were taken from the top of the hot spring (jiang et al., 2016). sequence reads were taxonomically classified at the order level for all sites where more than 500 sequence reads were available. around 10% of the obtained reads were unidentified, possibly due to many microorganisms being unculturable and thus not being included in databases. results showed that the archaeal order thermoproteales was present in all but one, crystal sister west, of the sampling sites, indicating that this order thrives in these extreme thermophilic conditions. since thermoproteales was detected in most of the sequenced libraries, it can be inferred that this order is selected for in these niches, and that its growth is not limited to a singular matrix in the hot spring. this aligns with previous literature that aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 11 depicts diverse metabolic capabilities of thermoproteales, which includes sulfur oxidation/reduction, oxygen respiration, h2 respiration, and co2 fixation (jay et al., 2016). several samples from the mud volcano site were collected, extracted, and sequenced. these included three mat samples, one water sample, two pool edge sediment samples, and two scum samples. this allowed us to compare communities from different matrices and locations within one site. the mud volcano sediments (central and pool edge sediment) were dominated by archaea while the mats and scum showed the presence of both bacteria and archaea. mud volcano has moderately hot and acidic water (table 1), which has magnesium, calcium, and sodium along with high concentrations of both organic carbon and sulfur (sherman et al., 2009). these are all necessary trace minerals that can be used for energy and growth, so it is not surprising that we found an abundance of both acidophilic and thermophilic bacteria and archaea. observing the abundance of taxa in different sample matrices and springs suggests a potential gradient. most of the samples dominated by archaea are the sediment samples, while the mats, water, and scum samples are dominated by bacteria (excluding mud volcano). the anoxic sediments of the hot springs, which have a higher concentration of hg and sulfide (krabbenhoft et al., 1999), show an abundance of archaeal species. on the other hand, the mat, water, and scum samples were taken from the water above the sediment and thus exposed to oxygen, sunlight, less mercury, and less sulfide has a higher abundance of bacterial taxa. this is likely due to the innate characteristics of archaea, such as a lipid monolayer and having specific proteins that have evolved to remain stable and active in various harsh conditions that allow them to thrive and proliferate in extreme environments (reed et al., 2013). in addition, such gradients could be due to the competition between the archaea and bacteria. the archaea outnumber the bacteria at the bottom of the springs, thus out-competing them for resources and nutrients. archaea are better adapted to survive in oligotrophic, low nutrient environments. they are able to live in sediments that are energy stressed due to having lower cell permeability than bacteria, thus reducing energy loss (vuillemin et al., 2019). in addition, archaea are also able to outcompete bacteria for urea, which is used as a source of carbon for the cell (seyler et al., 2019). this shows that archaea have developed several adaptations that allow them to utilize nutrients that often cannot be used by bacteria, thus allowing them to grow and survive in more environments and outcompete bacteria for the limited nutrients. at low temperatures, bacterial and archaeal species were found in similar abundances, while archaea were much more dominant at high temperatures. this indicates that higher temperatures allow archaeal dominance within acidic to neutral ph ranges (1.71-6.65). this might be because most alkaliphilic archaea are also considered halophilic due to sharing similar genomic characteristics (reed et al., 2013). the sampled alkaline springs, like octopus spring, did not have a high salt concentration, thus possibly selecting hyperthermophilic and alkaliphilic bacteria instead of archaea. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 12 in yellowstone national park, as well as other geothermal springs, there are often moderate to high levels of toxic compounds, such as hg, or those that may serve as growth substrates, such as sulfur. consequently, organisms often carry genes for resistance to these compounds or may utilize these compounds for better survival and growth. in our study, genes necessary for hg methylation, hgca, were not detected in any of the hot spring samples, and no known hg methylating genera were detected by 16s sequencing. this, however, does not indicate that there is no presence of hg methylation, it just means that the samples that were collected did not contain microorganisms with this gene or that organisms with this gene were detected but not characterized due to the limited amount of information that is known about uncultured extremophiles. when looking at hg resistance, the taxonomy derived from the 16s sequences showed the presence of the archaeon sulfolobus spp., which has a mer operon (barkay and wagner-dobler, 2005), in the infant geyser outlet sediment sample. this suggests that it is possible that infant geyser, which has a ph of 3.05, ~5.5 x 104 ng/l of total mercury, and a temperature of 82.7℃, has thermophilic mercury resistant species (barkay, personal communication). sequence reads in the order aquificales, which also contains the mer operon, were detected in several sampling sites including nymph (~1,000 ng/l total mercury), crystal sister west (>10,000 ng/l total mercury), mud volcano (1 x 104 1 x 105 ng/l total mercury), and gibbon canyon (>1,000 ng/l total mercury) (geesey et al., 2016; barkay, personal communication). aquificales bacteria were most abundant in nymph and mud volcano. this suggests a potential of mercury reduction occurring within those environments, and this makes sense because these four sites all have a ph of around two. we detected the mera gene in dna from three samples from mustard spring and nymph. mustard spring has a ph of 8.19 with a temperature of 86.2 ºc while nymph spring has a ph of 2.32 and a temperature of 59.5 ºc. this data demonstrates that in the springs where the total mercury concentration in unfiltered water is around 1,000 ng/l and above, there are taxa that have the potential for mercury resistance (barkay, personal communication). in hot springs that have a low ph, there is often a higher concentration of total mercury, thus selecting for mercury resistant organisms. while the genes needed for mercuric reduction were detected within the samples, detection of the gene does not indicate activity. further testing is needed to determine if mercury reduction genes are active within these species. it is important to understand hg reduction because it plays a critical role in the global cycling of hg. the reduction of mercury is the only transformation of a toxic metal that can be done on a large scale (barkay et al., 2003). microorganisms that are capable of mercuric reduction are able to transform both organic and inorganic hg during biotic and abiotic processes (barkay et al., 2003). this transformation helps to prevent hg bioaccumulation, which is a serious problem in some parts of the world. for many decades now, hg reducing bacteria have been utilized by humans to remove toxins from human products such as wastewater (barkay et al., 2003). this means that hg reduction is not only necessary for ecosystem homeostasis, but is also fundamental in the reduction of hg through bioremediation, which is the degradation of aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 13 environmental contaminants through the use of organisms, mainly microorganisms (vidali, 2001). the different taxa shown here to live in hot springs with high mercury concentration, such as sulfolobus spp. and the aquificales bacterial order, can be used for further research to support the bioremediation process. our results correspond with previous literature about community structure within hot springs in yellowstone national park. extracting metagenomic information from several springs demonstrated that the microbial communities cluster together based on the temperature of the spring and sample matrix. as one selection factor often dominates in extreme environments (uribe-lorio et al., 2019), it is important to know that the main selection factor for these communities in yellowstone is temperature. a potential gradient based on the abundance of archaeal and bacterial species was observed in samples from different compartments within the same spring. archaeal species were more abundant in the sediment samples and their presence decreased at the surface of the water, while bacterial species exhibited a reverse gradient where they were most abundant in the water and mats at the top of the spring and less abundant down towards the sediment. this shows that there is a negative correlation between the presence of bacterial and archaeal growth within the hot springs. this is likely due to the different environmental conditions and compounds that are present at the top and bottom of the hot spring. this information gives us a better understanding of how microbial communities form, function, and interact with the surrounding environment. even though the data obtained in this study matches the previous literature, there are still several limitations that need to be addressed. first, the sample size of this study is small, so the data presented in this paper cannot be widely generalized. there were originally twenty-two samples, and only eleven of them had over 500 sequences and were kept for further analyses. five different basins were sampled. all sites had at least one water sample, however, only six sites had mat samples and only five sites had sediment samples. it was hard to extract genetic material from the water samples, as they had very low biomass. furthermore, in those water samples, even if material was extracted, the sequence number was low. therefore, the samples could not be considered. this issue caused the number of usable samples for metagenomic analysis to decrease by half. out of the five basins originally sampled, only four (mud volcano, nymph, gibbon canyon, and norris geyser basin) had samples that were used for analysis and data extrapolation. another limitation is that some of the basins used for data analysis are similar in their ph and temperature range. this means that this data cannot be applied broadly; more samples from other basins with varying ph levels must be analyzed before results can be generalized. additionally, this paper used a metagenomic approach to show that mercury reduction genes are present, while mercury methylation genes were not. this, however, does not indicate if the mera genes are active within the community. rather, it only shows that these genes are present. regarding the absence of mercury methylation genes, this might not be representative of all of the sites due to the small sample size that was analyzed. in addition, the hg concentration was not measured at the time of sampling but rather at another time in the same year. overall, one of the more considerable limitations to this study is that many of the aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 14 microorganisms tested were unculturable. this means that there is often little information on these organisms, so it would be impossible to classify genes even if they were detected. this happened during the taxonomic analysis of the eleven samples, which were often categorized with only domain, family classification, or just “uncultured.” these samples therefore did not have enough information for us to determine a bigger picture of the community structure. due to the various limitations that are present in this study, there is more that can be done for future research and analysis. first, more samples need to be taken, specifically more sediment samples, and they should be taken from a wider range of basins located throughout yellowstone national park. this will not only help to increase the sample size but also to search for other mercury reducers and mercury methylators within the different environments. second, other mer operon genes can be tested instead of just mercuric reductase (mera). this can improve understanding of the different capabilities of the microorganisms inhabiting hot springs. lastly, proteomics analyses can be done to see if the mera genes are active within the sampled microorganisms. other research that can further this study is to look at other ways these microorganisms resist toxic mercury concentrations. while some microbes do not have the mer operon, they could have other ways to resist hg. these methods could include influx, accumulation, efflux, and metallothionein, a protein that protects the cell against the toxicity of hg (irawati et al., 2012). it is important to know the many diverse ways that microorganisms resist toxic metals. this new information can give us a more holistic view of the unique environment of the microorganisms in the hot springs of yellowstone national park. acknowledgments we would like to thank dr. joann holloway and her team for collecting all of the samples in yellowstone national park. references anderson, m. j. (2001). a new method for non-parametric multivariate analysis of variance. austral ecology, 26(1), 32–46. doi: 10.1111/j.1442-9993.2001.01070.pp.x apprill, a., mcnally, s., parsons, r., & weber, l. (2015). minor revision to v4 region ssu rrna 806r gene primer greatly increases detection of sar11 bacterioplankton. aquatic microbial ecology, 75(2), 129–137. doi: 10.3354/ame01753 barkay, t., miller, s. m., & summers, a. o. (2003). bacterial mercury resistance from atoms to ecosystems. fems microbiology reviews, 27(2-3), 355–384. doi: 10.1016/s01686445(03)00046-9 barkay, t., & wagner-döbler, i. (2005). microbial transformations of mercury: potentials, challenges, and achievements in controlling mercury toxicity in the environment. advances in applied microbiology advances in applied microbiology volume 57, 1–52. doi: 10.1016/s0065-2164(05)57001-1 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 15 bolyen, e., rideout, j. r., dillon, m. r. et al. (2019) reproducible, interactive, scalable and extensible microbiome data science using qiime 2. nat biotechnol 37, 852–857 doi:10.1038/s41587-019-0209-9 callahan, b. j., mcmurdie, p. j., rosen, m. j., han, a. w., johnson, a. j., & holmes, s. p. (2016). dada2: high resolution sample inference from amplicon data. nature methods, 13(7), 581 doi: 10.1101/024034 cuhel r. l., aguilar c., anderson p. d., maki j. s., paddock r. w., remsen c. c., et al. (2002). underwater domains in yellowstone lake hydrothermal vent geochemistry and bacterial chemosynthesis in: anderson, r.j., and d. harmon, eds. 2002. yellowstone lake: hotbed of chaos or reservoir of resilience? proceedings of the 6th biennial scientific conference on the greater yellowstone ecosystem. mammoth hot springs hotel, yellowstone national park. yellowstone national park, wyo., and hancock, mich. yellowstone center for resources and the george wright society. geesey, g. g., barkay, t., & king, s. (2016). microbes in mercury-enriched geothermal springs in western north america. science of the total environment, 569-570, 321–331. doi: 10.1016/j.scitotenv.2016.06.080 hall, m., & beiko, r. g. (2018). 16s rrna gene analysis with qiime2. methods in molecular biology microbiome analysis, 113–129. doi: 10.1007/978-1-4939-8728-3_8 irawati, w., p., soraya, y., & baskoro, a. h. (2012). a study on mercury-resistant bacteria isolated from a gold mine in pongkor village, bogor, indonesia. hayati journal of biosciences, 19(4), 197-200. doi:10.4308/hjb.19.4.197 jay, z. j., beam, j. p., kozubal, m. a., jennings, r. d., rusch, d. b., & inskeep, w. p. (2016). the distribution, diversity and function of predominant thermoproteales in hightemperature environments of yellowstone national park. environmental microbiology, 18(12), 4755–4769. doi: 10.1111/1462-2920.13366 jiang, z., li, p., nostrand, j. d. v., zhang, p., zhou, j., wang, y., … wang, y. (2016). microbial communities and arsenic biogeochemistry at the outflow of an alkaline sulfide-rich hot spring. scientific reports, 6(1). doi: 10.1038/srep25262 krabbenhoft, d.p., wiener, j.g., brumbaugh, w.g., olson, m.l., dewild, j.f., & sabin, t.j. (1999). a national pilot study of mercury contamination of aquatic ecosystems along multiple gradients: bioaccumulation in fish. environmental science. lozupone, c., lladser, m. e., knights, d., stombaugh, j., & knight, r. (2011). unifrac: an effective distance metric for microbial community comparison. the isme journal, 5(2), 169-172. doi:10.1038/ismej.2010.133 parada, a. e., needham, d. m., & fuhrman, j. a. (2016). every base matters: assessing small subunit rrna primers for marine microbiomes with mock communities, time series and global field samples. environmental microbiology, 18(5), 1403–1414. doi: 10.1111/1462-2920.13023 pedregosa f., varoquaux g., gramfort a., michel v., thirion b., grisel o., blondel m., prettenhofer p., weiss r., dubourg v., vanderplas j., passos a., cournapeau d., brucher m., perrot m., and duchesnay e. (2011). scikit-learn: machine learning in python. journal of machine learning research, 12, 2825–2830. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 16 poulain, a. j., aris-brosou, s., blais, j. m., brazeau, m., keller, w., & paterson, a. m. (2015). microbial dna records historical delivery of anthropogenic mercury. the isme journal, 9(12), 2541–2550. doi: 10.1038/ismej.2015.86 reed, c. j., lewis, h., trejo, e., winston, v., & evilia, c. (2013). protein adaptations in archaeal extremophiles. archaea, 1–14. doi: 10.1155/2013/373275 reysenbach a-l., ehringer m, hershberger k. (2000). microbial diversity at 83°c in calcite springs, yellowstone national park: another environment where the aquificales and "korarchaeota" coexist. extremophiles 4:61–67. schaefer, j. k., kronberg, r.-m., morel, f. m. m., & skyllberg, u. (2014). detection of a key hg methylation gene,hgca, in wetland soils. environmental microbiology reports, 6(5), 441– 447. doi: 10.1111/1758-2229.12136 seyler, l. m., tuorto, s., mcguinness, l. r., gong, d., & kerkhof, l. j. (2019). bacterial and archaeal specific-predation in the north atlantic basin. frontiers in marine science, 6. doi: 10.3389/fmars.2019.00555 sherman, l., blum, j., nordstrom, d., mccleskey, r., barkay, t., & vetriani, c. (2009). mercury isotopic composition of hydrothermal systems in the yellowstone plateau volcanic field and guaymas basin sea-floor rift. earth and planetary science letters, 279(1-2), 86–96. doi: 10.1016/j.epsl.2008.12.032 shock, e. l., holland, m., meyer-dombard, d. a., amend, j. p., osburn, g., & fischer, t. p. (2010). quantifying inorganic sources of geochemical energy in hydrothermal ecosystems, yellowstone national park, usa. geochimica et cosmochimica acta, 74(14), 4005–4043. doi: 10.1016/j.gca.2009.08.036 ul-hasan s, bowers rm, figueroa-montiel a, licea-navarro af, beman jm, woyke t, nobile cj. (2019). community ecology across bacteria, archaea and microbial eukaryotes in the sediment and seawater of coastal puerto nuevo, baja california. plos one 14(2): e0212355. doi: 10.1371/journal.pone.0212355 uribe‐lorío, l., brenes-guillén, l., hernández‐ascencio, w., mora‐amador, r., gonzález, g., ramírez‐umaña, c. j., … pedrós‐alió, c. (2019). the influence of temperature and ph on bacterial community composition of microbial mats in hot springs from costa rica. microbiologyopen, 8(10). doi: 10.1002/mbo3.893 vidali, m. p. (2001). bioremediation. an overview. pure and applied chemistry, 73(7), 1163 1172. doi:https://doi.org/10.1351/pac200173071163 vuillemin, a., wankel, s. d., coskun, ö k., magritsch, t., vargas, s., estes, e. r., . . . orsi, w. d. (2019). archaea dominate oxic subseafloor communities over multimillion-year time scales. science advances, 5(6). doi:10.1126/sciadv.aaw4108 yellowstone national park – official park map. (n.d.). in yellowstone maps. retrieved from https://yellowstone.net/maps/yellowstone-park-map/ https://dx.doi.org/10.1371/journal.pone.0212355 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. the role of eicosanoids in regulating the ubiquitin proteasome system and proteostasis amanpreet kaur abstract the ubiquitin proteasome system (ups) is a protein degradation mechanism in eukaryotes crucial to maintaining protein homeostasis, or proteostasis. there are tissue-specific differences in ups activity and proteostasis, but the intercellular signalling mechanisms that mediate these differences are not well understood. this work examines eicosanoid signalling molecules—which are derived from polyunsaturated fatty acids (pufas)—and their role in proteostasis regulation, particularly the ups. a reporter transgene that expresses the ubg76v-gfp chimeric protein, a metastable substrate for the ups, is used in caenorhabditis elegans epithelial cells to monitor the level of ups activity. in wild-type nematodes, ubg76v-gfp levels remain high through 24 hours post l4 stage (l4+24). then, levels decrease significantly due to increased ups activity as the animals age and develop 48 hours past l4 (l4+48). mutants for fat-1, a desaturase enzyme that converts ω-6 pufas to ω-3 pufas, exhibited elevated ubg76v-gfp turnover in the hypodermis even at the l4+24 stage, suggesting that either ω-6 pufas (or their eicosanoid derivatives) promote ups activity or ω-3 pufas (or their eicosanoid derivatives) inhibit ups activity. in the intestine, mutants for fat-1 showed reduced ubg76v-gfp turnover at the l4+24 and l4+48 life stages. additionally, mutants for emb-8—an nadph reductase needed to convert pufas into eicosanoids—also showed reduced ubg76v-gfp turnover in the hypodermis even at the l4+48 stage. these results suggest that elements of the eicosanoid signalling pathway, including ω-6 pufas and their derivatives, significantly contribute to regulation of the ups and proteostasis. introduction in eukaryotic organisms, protein homeostasis involves the cellular control of protein synthesis, folding, trafficking, and degradation in order to properly control biological processes—specifically, dna replication, transcription, and mitosis. rapid cell division during these biological processes can cause unfolded or damaged proteins. the initiation of refolding these proteins, along with the prevention of protein aggregation, is controlled partly by protein degradation mechanisms. many human diseases are characterized by an inability to degrade damaged or unfolded proteins, particularly those caused by neurodegeneration in the brain. the resultant protein aggregation is often responsible for the neuronal cell death seen in diseases such as alzheimer’s, parkinson’s, and huntington’s, which predominantly affect older populations. a major mechanism of protein degradation is the ubiquitin proteasome system (ups). the ups removes damaged and unfolded target proteins by tagging them with ubiquitin, a small https://creativecommons.org/licenses/by-nc-sa/4.0/ https://creativecommons.org/licenses/by-nc-sa/4.0/ aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 2 protein, resulting in their recognition and digestion by a large proteasome protease complex (figure 1). in the ups, an initial atp hydrolysis step adenylates a molecule of ubiquitin. the ubiquitin is then transferred to the target protein to signal for further polyubiquitination. once the chain of ubiquitin molecules on the target protein is at least four units long, that substrate becomes a target for degradation by the proteasome. there is a reduction in proteostasis that occurs with age, but the specific changes that occur and the pathways that affect these changes are currently unclear. figure 1. schematic representation of ubiquitin proteasome system (ups) activity in the hypodermis. proper ups function shows that the unstable ubg76v-gfp (green) transgene is quickly polyubiquitinated and degraded whereas mrfp (red) remains stable. when not functioning properly, ubg76v-gfp is either not polyubiquitinated or not degraded and thus remains stable. adapted from “dopamine signaling promotes the xenobiotic stress response and protein homeostasis,” by k.k. joshi, t. l. matlack, and c. rongo, 2016, the embo journal, 35(17), 1885-1901 (doi: 10.15252/embj.201592524). copyright 2016 by embo press. adapted with permission. cytochrome p450s (cyps) are reduction-oxidation (redox) enzymes at the end of the electron transfer chain.1 loss-of-function mutants for various cyps have previously been shown to stabilize polyubiquitinated substrates, suggesting decreased ups activity.2 certain cyps also metabolize polyunsaturated fatty acids (pufas) to eicosanoid signalling molecules (figure 2). pufas relevant to eicosanoid signalling can be divided into ω-3 pufas, which have their first double bond at the third carbon, and ω-6 pufas, which have their first double bond at carbon 6. 1 lamb et al., 2009 2 joshi, matlack, & rongo, 2016 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 3 figure 2. schematic representation of ω-6 and ω-3 pufas and the cyp-eicosanoid derivative signaling pathway. ω-6 pufas are converted to ω-3 pufas via fat-1 and pufa-derived eicosanoids are produced via cyps and emb-8. thus, the eicosanoid synthesis pathway is a worthwhile target for study of ups regulation. here, various aspects of the eicosanoid synthesis pathway are studied in order to shed light on ups regulation. methods ups activity can be monitored in caenorhabditis elegans, a roundworm species, via a transgenic approach using the ubiquitin fusion degradation reporter transgene ubg76vgfp, which is an unstable protein covalently bonded to a green fluorescent protein (gfp). it mimics a monoubiquitinated protein that can be further polyubiquitinated and degraded. once degraded, gfp is no longer detected (figure 1). therefore, gfp fluorescence in animals expressing ubg76v-gfp is inversely proportional to the level of ups activity. the ubg76vgfp is expressed in epithelial cells from tissue-specific promoters—col-19 in the hypodermis and sur-5 in the intestine. two red fluorescent proteins, mrfp in the hypodermis and mcherry in the intestine, may be monitored alongside ubg76v-gfp and used as an internal standard for gene expression. neither mrfp nor mcherry is ubiquitinated, so they do not get degraded. but, they are both expressed from the same promoter as gfp, allowing the quantification of gene expression by analyzing the ratios of gfp/mrfp and gfp/mcherry (figure 1). essential fatty acids, specifically the ω-3 pufa alpha-linolenic acid (ala), are important eicosanoid precursors that cannot be synthesized in the human body from monounsaturated fatty acids (mufas). however, c. elegans is able to convert other ω-6 pufas to ω-3 pufas using the fat-1 gene, making it a model organism for eicosanoid study (figure 2). mutants of fat-1 fail to make ω-3 pufas and subsequently accumulate ω6 pufas. studying these mutants can help determine whether either type of pufa distinctly contributes to regulating the ups. aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 4 first, the odis77 transgene expressing the ubg76v-gfp and mrfp reporters in the hypodermis was introduced into a fat-1(wa9) mutant strain and a control odis77 (wild-type) strain (table 1). c. elegans undergoes four larval stages starting from l1 and ending at l4, the final 24-hour larval stage before the worm enters adulthood. a wild-type worm expressing ubg76v-gfp and mrfp in the hypodermis quickly accrues both proteins between the l4 larval stage and the l4+24 adult stage. upon entering the l4+48 adult stage, which is associated with peak fertility, ubg76v-gfp is swiftly degraded while mrfp remains relatively stable (figure 1). both the wild-type and fat-1 strains were imaged at the l4+24 and l4+48 stages in order to determine ubg76v-gfp and mrfp levels through fluorescence analysis (table 1). to establish if ubg76v-gfp turnover with a fat-1 mutant is as quick in the intestine as it is in the hypodermis, a fat-1(wa9); psur5 mutant strain homozygous for ubg76v-gfp was generated (table 1). ubg76v-gfp expression in the intestine is less stable than expression in the hypodermis, suggesting earlier ups activity in the intestine. this is possibly due to tissue-specific differences in proteasome demand, activity, or expression. thus, there was a possibility of mistaking typical unstable ubg76v-gfp levels for a heterozygous genotype. furthermore, in the intestine, ubg76v-gfp is on chromosome ⅲ; the rfp control, mcherry, is on chromosome ⅴ; and the fat-1 mutation is located on chromosome ⅳ. so, there was a possibility of losing either fluorescent tag. generating the intestinal fat-1 strain in the lab allowed greater control over ensuring fluorescence. a pp608 (wild-type) strain, also homozygous for ubg76v-gfp and expressing ubg76v-gfp (chromosome iii) and mcherry (chromosome v) from the sur-5 promoter, was used as a control to determine the impact of a fat-1 mutation in the intestine on ups activity. both strains were imaged and analyzed at the l4+24 and l4+48 life stages (table 1). in c. elegans, proper cyp function requires the emb-8 gene, which codes for a reductase enzyme that catalyzes electron transfer from nadph to the cyp.3 cyps require emb-8 to metabolize eicosanoids from pufas, so emb-8 mutants fail to produce pufa-derived cypeicosanoids (figure 2). studying emb-8 mutants will help characterize the role of emb-8 eicosanoid derivatives in ups regulation. because eicosanoids are essential signalling molecules in c. elegans, the emb-8 null mutation is a lethal one. as a result, it is difficult to analyze the function of this gene in adult animals via a knockout study. to get around this issue, a temperature sensitive allele of emb-8 called emb-8(hc69) was synthesized to determine how an emb-8 mutation in the hypodermis impacts ups activity. specifically, the emb-8(hc69); odis76 mutant strain and the control odis76 (wild-type) strain were generated and examined, both with the ubg76vgfp and mrfp reporter transgenes again expressed in the hypodermis (table 1). both strains were imaged and analyzed for fluorescence at the l4+24 and l4+48 stages. the l4+72 stage was also imaged for emb-8 mutants to determine the extent of the stability of ubg76v-gfp (table 1). 3 benenati, penkov, muller-reichert, entchev, & kurzchalia, 2009 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 5 *only for the emb-8 mutant strain table 1. experimental setup. ubg76v-gfp/mrfp ratios were measured in the hypodermis at l4+24 hrs and l4+48 hrs for the fat-1 mutant and odis77 wild-type strains. ubg76vgfp/mrfp ratios were measured in the hypodermis at l4+24 hrs, l4+48 hrs, and l4+72 hrs for the emb-8 mutant and pp608 wild-type strains. ubg76v-gfp/mcherry ratios were measured in the intestine at l4+24 hrs and l4+48 hrs for the fat-1 mutant and odis76 wild-type strains. results accelerated ubg76v-gfp turnover in fat-1 mutants in the hypodermis as expected, we observed that the wild-type strain exhibited accumulated ubg76v-gfp at the l4+24 stage, with similar levels of both ubg76v-gfp and mrfp, giving a mean normalized ratio close to 1 (figures 3 & 5; table 2). this is consistent with accelerated protein accumulation seen at the end of the final larval stage as a wild-type worm enters adulthood, likely due to reproduction-associated protein synthesis.4 the wild-type strain exhibited reduced ubg76v-gfp levels at the l4+48 stage compared to the internal control mrfp, with a mean ubg76v-gfp/mrfp ratio closer to 0 (figures 7 & 9; table 2). this is expected in the hypodermis of wild-type worms that have reached peak fertility, possibly as a protective mechanism against damage caused by protein aggregation.5 the wild-type strain expressed mrfp to the same intensity at both stages, which indicates that non-ubiquitinated proteins in the hypodermis of wild-type worms are stable (figures 1, 5, & 9). by contrast, the fat-1 mutant strain showed increased ubg76v-gfp turnover, showing reduced ubg76v-gfp levels and a mean ubg76v-gfp/mrfp ratio close to 0 at both stages (figures 4 & 8; table 2). again, in this strain, mrfp remained stable at both stages, which is standard of non-ubiquitinated proteins (figures 6 & 10). 4 joshi et al., 2016 5 joshi et al., 2016 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 6 l4+24 hrs odis77 (wt) fat-1 (wa9); odis77 l4+48 hrs odis77 (wt) fat-1 (wa9); odis77 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 7 figures 3-10. ubg76v-gfp turnover is accelerated in the hypodermis of fat-1 mutants. 3 expression of ubg76v-gfp (green) and 5 mrfp (red) from the col-19 promoter at l4+24 hrs in wild-type animals. 4 expression of ubg76v-gfp (green) and 6 mrfp (red) from the col-19 promoter at l4+24 hrs in fat-1 mutants. 7 expression of ubg76v-gfp (green) and 9 mrfp (red) from the col-19 promoter at l4+48 hrs in wild-type animals. 8 expression of ubg76v-gfp (green) and 10 mrfp (red) from the col-19 promoter at l4+48 hrs in fat-1 mutants. table 2. fat-1 mutants show accelerated ubg76v-gfp turnover in the hypodermis at l4+24 hrs. mean ubg76v-gfp/mrfp ratios at l4+24 hrs and l4+48 hrs in the hypodermis of fat-1 mutant and wild-type strains. at the l4+24 stage, the variation in ubg76v-gfp expression between the wild-type strain and the fat-1 mutant strain was significant (p<0.0001) (figure 11). ubg76v-gfp levels in the wild-type strain were noticeably higher than in the fat-1 mutant strain at this stage. at the l4+48 stage, both strains expressed low ubg76v-gfp levels (p<0.01) (figure 12). taken together, these results suggest that the ups is activated earlier the hypodermis of fat-1 mutants than in the wild-type strain. as discussed, fat-1 mutants are unable to convert ω-6 pufas to ω-3 pufas, which leads to an accumulation of the former and a corresponding deficit in the latter (figure 2). therefore, there are two mechanisms by which a fat-1 mutation can cause earlier ups activation. first, ω-6 pufas or their eicosanoid derivatives could promote the ups, so their accumulation would cause premature ups activity. alternatively, ω-3 pufas or their eicosanoid derivatives could inhibit the ups, so their deficiency would accelerate ups activity. 11 l4+24 hrs 12 l4+48 hrs figures 11 and 12. graphical representation of accelerated ubg76v-gfp turnover in the hypodermis of fat-1 mutants. 11 quantified fluorescence ratios of ubg76v-gfp to mrfp in the hypodermis of 20 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 8 animals of specified strains at l4+24 hrs. ****p<0.0001. error bars denote sem. 12 quantified fluorescence ratios of ubg76v-gfp to mrfp in the hypodermis of 20 animals of specified strains at l4+48 hrs. **p<0.01. error bars denote sem. reduced ubg76v-gfp turnover in fat-1 mutants in the intestine as predicted in the intestine, the wild-type strain exhibited a mean ubg76v-gfp/mcherry ratio close to 0 at both the l4+24 and l4+48 stages (table 3); ubg76v-gfp fluorescence and expression was low in wild-type worms at both stages (figures 13 & 17). this is consistent with the intestinal ubg76v-gfp instability and potentially high underlying ups activity observed in wild-type worms in prior studies.6 as the internal control, mcherry remained stable at both stages (figures 15 & 19). in the fat-1 strain, ubg76v-gfp levels in the intestine remained stable (and clearly higher than the wild-type strain) at both the l4+24 and l4+48 stages with slightly reduced levels at the latter stage (figures 14 & 18). once again, the mcherry standard remained stable at both stages (figures 16 & 20). thus, the fat-1 mutant strain displayed a ubg76v-gfp/mcherry ratio close to 1 at the l4+24 stage. this ubg76vgfp/mcherry ratio was only moderately reduced at the l4+48 stage, further suggesting increased ubg76v-gfp stability in the intestine of fat-1 mutants (table 3). given previous results with the fat-1 mutant in the hypodermis, it is possible that ups activity and its regulation are subject to tissue-specific differences. l4+24 hrs pp608 (wt) fat-1(wa9); psur-5 6 keith et al., 2016 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 9 l4+48 hrs pp608 (wt) fat-1(wa9); psur-5 figures 13-20. ubg76v-gfp turnover is reduced in the intestine of fat-1 mutants. 13 expression of ubg76v-gfp (green) and 15 mcherry (red) from the sur-5 promoter at l4+24 hrs in wild-type animals. 14 expression of ubg76v-gfp (green) and 16 mcherry (red) from the sur-5 promoter at l4+24 hrs in fat-1 mutants. 17 expression of ubg76v-gfp (green) and 19 mcherry (red) from the sur-5 promoter at l4+48 hrs in wild-type animals. 18 expression of ubg76v-gfp (green) and 20 mcherry (red) from the sur-5 promoter at l4+48 hrs in fat-1 mutants. table 3. fat-1 mutants show reduced ubg76v-gfp turnover in the intestine at l4+48 hrs. mean ubg76vgfp/mcherry ratios at l4+24 hrs and l4+48 hrs in the intestine of fat-1 mutant and wild-type strains. at both stages, the variation in the ubg76v-gfp fluorescence and expression between the two strains was significant (p<0.0001) (figures 21 & 22). ubg76v-gfp levels in the intestine of the wild-type strain were lower compared to the fat-1 mutant strain at both stages. these results suggest that, in the intestine of fat-1 mutants, ups activity is decreased compared to aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 10 wild-type animals. this contrasts with the hypodermis of fat-1 mutants, which exhibits increased ups activity relative to the wild-type strain. therefore, pufas (or their eicosanoid derivatives) in the intestine may have pathways inverse to those in the hypodermis. these results are consistent with a mechanism in which either the ω-6 pufa line inhibits ups activity, or the ω-3 pufa line promotes ups activity (figure 2). interestingly, wild-type worms also appear to differ in ups activity in the hypodermis and the intestine, with the intestine experiencing earlier ups activity than the hypodermis.7 in total, these results suggest that pufas or their eicosanoid derivatives regulate ups activity in a tissue-specific fashion. 21 l4+24 hrs 22 l4+48 hrs figures 21 and 22. graphical representation of reduced ubg76v-gfp turnover in the intestine of fat-1 mutants. 21 quantified fluorescence ratios of ubg76v-gfp to mcherry in the intestine of 20 animals of specified strains at l4+24 hrs. ****p<0.0001. error bars denote sem. 22 quantified fluorescence ratios of ubg76v-gfp to mcherry in the intestine of 20 animals of specified strains at l4+48 hrs. ****p<0.0001. error bars denote sem. reduced ubg76v-gfp turnover in emb-8 mutants in the hypodermis in the hypodermis, wild-type c. elegans expressed ubg76v-gfp at elevated levels with a mean ubg76v-gfp/mrfp ratio close to 1 at the l4+24 stage, which dropped to near 0 at the l4+48 stage (figures 23 & 27; table 4). as expected increased protein accumulation was seen in the hypodermis of wild-type worms when they first enter adulthood at l4+24 hours, and reduced protein levels were seen in the hypodermis of wild-type worms that reached peak fertility at l4+48 hours.8 the internal standard, mrfp, remained stable at both stages (figures 25 & 29). similarly, emb-8 mutants also exhibited accumulated ubg76v-gfp levels at the l4+24 stage with a mean ubg76v-gfp/mrfp ratio close to 1. however, emb-8 mutants exhibited only slightly reduced ubg76v-gfp levels and ubg76v-gfp/mrfp ratios at the l4+48 and l4+72 stages (figures 24, 28, & 31; table 4). at both of the later stages, ubg76v-gfp levels were stable compared to the wild-type strain at the l4+48 stage. once again, as a control, mrfp remained stable in the emb-8 mutant strain at all three stages (figures 26, 30, & 32). 7 keith et al., 2016 8 joshi et al., 2016 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 11 l4+24 hrs odis76 (wt) emb-8 (hc69); odis76 l4+48 hrs odis76 (wt) emb-8 (hc69); odis76 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 12 l4+72 hrs: emb-8 (hc69); odis76 figures 23-32. ubg76v-gfp turnover is reduced in the hypodermis of emb-8 mutants. 23 expression of ubg76v-gfp (green) and 25 mrfp (red) from the col-19 promoter at l4+24 hrs in wild-type animals. 24 expression of ubg76v-gfp (green) and 26 mrfp (red) from the col-19 promoter at l4+24 hrs in emb-8 mutants. 27 expression of ubg76v-gfp (green) and 29 mrfp (red) from the col-19 promoter at l4+48 hrs in wild-type animals. 28 expression of ubg76v-gfp (green) and 30 mrfp (red) from the col19 promoter at l4+48 hrs in emb-8 mutants. 31 expression of ubg76v-gfp (green) and 32 mrfp (red) from the col-19 promoter at l4+72 hrs in emb-8 mutants. table 4. emb-8 mutants show reduced ubg76v-gfp turnover in the hypodermis at l4+24 hrs. mean ubg76v-gfp/mrfp ratios at l4+24 hrs and l4+48 hrs in the hypodermis of emb-8 mutant and wildtype strains. mean ubg76v-gfp/mrfp ratio at l4+72 hrs is also shown for the emb-8 mutant strain. at the l4+24 stage, the variation in ubg76v-gfp fluorescence and expression between the wild-type and emb-8 mutant strains was not significant (p>0.05) (figure 33). these results are unsurprising in the hypodermis, given that ubg76v-gfp accumulated at l4+24 hours in both strains. and, this aligns with increased protein levels and low ups activity seen in the hypodermis of wild-type animals when they initially enter adulthood.9 however, at the l4+48 stage, the difference in ubg76v-gfp fluorescence and expression was highly significant (p<0.0001) (figure 34). although ubg76v-gfp levels were reduced in the wildtype strain at l4+48 hours, ubg76v-gfp levels in emb-8 mutants remained stable at l4+48 hours.10 in the eicosanoid synthesis pathway, the emb-8 gene is involved in the eventual conversion of both ω-6 and ω-3 pufas to their eicosanoid derivatives (figure 2). thus, if an emb-8 mutant stabilizes ubg76v-gfp in the hypodermis by decreasing ups activity, then 9 joshi et al., 2016 10 joshi et al., 2016 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 13 these results support an eicosanoid-regulated pathway of the ups, as opposed to a pufaregulated pathway. it is likely that either ω-6 pufa-derived eicosanoids promote the ups or ω-3 pufa-derived eicosanoids repress it. as a whole, these results indicate that ups activity is suppressed in the hypodermis in emb-8 mutants. 33 l4+24 hrs 34 l4+48 hrs figures 33 and 34. graphical representation of reduced ubg76v-gfp turnover in the hypodermis of emb-8 mutants. 33 quantified fluorescence ratios of ubg76v-gfp to mrfp in the hypodermis of 20 animals of specified strains at l4+24 hrs. lack of a p-value indicates p>0.05. error bars denote sem. 34 quantified fluorescence ratios of ubg76v-gfp to mrfp in the hypodermis of 20 animals of specified strains at l4+48 hrs. ****p<0.0001. error bars denote sem. discussion in order to characterize the role of eicosanoids in the regulation of ups activity and proteostasis, a ubg76v-gfp transgene was introduced into various c. elegans mutants to monitor changes in ups activity, which is inversely related to ubg76v-gfp intensity. in wildtype c. elegans strains, ubg76v-gfp begins to accrue at the l4+24 stage and is swiftly degraded when worms enter peak fertility at the l4+48 stage. in this work, it was confirmed that several constituents of the fatty acid-eicosanoid signalling pathway, namely fat-1 and emb-8, are involved in mediating the ups. in the hypodermis, fat-1 mutants experienced earlier ups activation relative to wild-type animals, while emb-8 mutants exhibited repressed ups activity. in the intestine, however, fat-1 mutants displayed ups repression. these findings suggest that eicosanoids function in a tissue-specific manner to regulate the ups and maintain proteostasis. two mechanisms within the fatty-acid-eicosanoid pathway can be proposed for the increased ubg76v-gfp turnover observed in the hypodermis of a fat-1 mutant. mutants for fat-1 accumulate ω-6 pufas because they cannot convert them to ω-3 pufas. therefore, one possible explanation is that ω-6 derived cyp-eicosanoids promote ups activity. this means a fat-1 mutation leads to enhanced ups activity in the hypodermis due to high levels of one or more ω-6 pufas or their eicosanoid derivatives. another potential explanation is that ω-3 pufas or their eicosanoid derivatives inhibit ups activity, so their absence due to aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 14 fat-1 results in the de-repression of hypodermal ups activity. since the results presented here do not suggest one pathway over the other, further studies are required to determine which one of these two pathways—if not a combination of both—is responsible for the effects of fat-1 on ups activity. curiously, fat-1 in the intestine increases ups activity, which is the opposite of its inhibitory role in the hypodermis. thus, it is likely that eicosanoids exert their effects on the ups in a cell and tissue-specific manner. this is supported by previous studies that found the effects of ω-3 pufas varied depending on the tissue in question, with reduced ups activity in cachexia cancer cells and increased ups activity in hpv-infected cancer cells.11 in the intestine, ω-3 pufas or their derived eicosanoids may promote ups activity, or ω-6 pufas or their derived eicosanoids may inhibit it. the exact purpose of tissue-specific differentiation in ups regulation is unknown, although it is possible that there is a greater demand to break down proteins in the intestine than in the hypodermis. together with cyps, emb-8 is responsible for the direct production of eicosanoids. the stabilization of ubg76v-gfp resulting from an emb-8 mutation in the hypodermis likely suggests a mechanism in which eicosanoids promote ups activity. this is further supported by the observations in the hypodermis of a fat-1 mutant. to determine whether it is ω-6 pufas or ω-6 derived eicosanoids regulating the ups in the hypodermis, it would be worthwhile to study a fat-1 and emb-8 (or even a cyp) double mutant and observe whether the emb-8 mutation blocks the increased ubg76v-gfp turnover induced by the fat-1 mutation. the same experiment could be repeated in the intestine to determine the role of eicosanoids in regulating the ups in a different type of cell. double mutations in fat-1 and other fat enzymes would also be valuable in determining the contribution of different ω-6 pufas to ups regulation and if the effect arises from their derived eicosanoids instead. studying mutations in ceeh-1 and ceeh-2—which convert nonclassical, short-lived eicosanoids to classic eicosanoids—would also be helpful in the latter case (figure 2). although the results in this paper have been discussed in the context of eicosanoids regulating the ups, it is also possible that they function through regulating other protein degradation mechanisms, such as autophagy or even proteasome activity itself. one study observed that supplementing growth media with ω-6 pufas activates autophagy in c. elegans, which serves to increase starvation resistance and elongate their lifespans.12 another publication found that ω-3 pufas increased the concentration of ubiquitinated proteins in adipocytes, crediting it to reduced expression of proteasome subunits.13 although previous studies have examined mutations in possible regulators of the ups in conjunction with direct disruption of the proteasome, the exact role of eicosanoids in modulating proteostasis remains unclear.14 regardless of their precise function, eicosanoids are key components in eukaryotic proteostasis, including protein degradation. beyond c. elegans, understanding the role of 11 whitehouse, smith, drake, & tisdale, 2001; jing et al., 2014 12 o’rourke, kuballa, xavier, & ruvkun, 2013 13 wójcik et al., 2014 14 joshi et al., 2016 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 15 eicosanoids synthesized from essential fatty acids in proteostasis will help to elucidate their biological importance. improving our understanding of the ups system may help us to better comprehend and prevent age-associated neurodegeneration and disease. materials and procedures strains used the following strains were used: odis77 (wt)[pcol-19::ubg76v-gfp, pcol-19::mrfp], fat-1(wa9); odis77[pcol-19::ubg76v-gfp, pcol-19::mrfp], odis76 (wt)[pcol-19::ubg76v-gfp, pcol-19::mrfp]; emb8(hc69); odis76[pcol-19::ubg76v-gfp, pcol-19::mrfp], dpy-13[psur-5::ubg76v-gfp, unc-119(+) ⅲ, psur-5::mcherry, unc-119(+) v], and hhis64[psur-5::ubg76v-gfp, unc-119(+)] ⅲ; hhis73[psur5::mcherry, unc-119(+)] v. growth conditions under standard conditions, worms were grown at 20℃ on nematode growth media (ngm) plates seeded with op50 escherichia coli. the temperature-sensitive emb-8(hc69); odis76 strain was grown at 15℃ until the post-l4 stage, when it was shifted to 20℃, along with the wild-type control, odis76. to generate the fat-1(wa9); psur-5 strain, the dpy-13; psur-5 strain was used as a balancer to introduce psur5 into the fat-1(wa9) strain. 15 hermaphrodites of the dpy-13; psur-5 strain were first crossed to 20 laboratory n2 (wt) males. males of the heterozygous f1 generation were crossed to fat-1(wa9) hermaphrodites. 20 f2 progeny were separated onto 20 plates and self-crossed. from f3 plates that produced both wild-type and dpy worms, 40 phenotypically wild-type worms were separated onto 40 plates and self-crossed. 25 f4 progeny from f3 plates that produced only phenotypically wild-type progeny were separated onto 25 plates. all worms separated displayed ubg76v-gfp and mcherry fluorescence. to test if ubg76v-gfp could be observed in the fat-1(wa9); psur-5 strain, 5 f5 progeny from each of the 25 f4 plates were separated onto 25 ufd-1 rnai plates to knock down ubiquitination in the worms. five worms for the odis77 (wt) strain were also separated onto a ufd-1 rnai plate as a control. worms from the pp608 (wt) control strain were confirmed to be homozygous for ubg76v-gfp by growing on a single ufd-1 rnai plate. statistics a t-test with welch’s correction was used in graphpad prism to analyze ubg76v-gfp and mrfp or mcherry fluorescence values for twenty worms per strain per life stage. each strain’s values were then graphed in graphpad prism. bleaching and synchronization gravid adult worms were harvested with m9 buffer and lysed with modified bleaching solution prepared in distilled water containing 10% 10 n naoh and 10% commercially aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 16 available bleaching solution. eggs were resuspended in 20-70 μl of m9 buffer and spotted onto ngm plates seeded with e. coli. fluorescence microscopy, imaging, and intensity analysis twenty worms per strain per life stage were imaged using a 5x objective to detect ubg76vgfp, mrfp, and mcherry fluorescence with an epifluorescence microscope and ivision software. the worms were first mounted and paralyzed on slides with 2% agarose and 10mm tetramisole. the exposure times were 50 ms for l4+24 hrs and 25 ms for l4+48 and l4+72 hrs. imagej was used to outline individual worms in the images and quantify fluorescence. python software was used to calculate mean fluorescence. rnai screen five individuals from each strain to be screened, including the wild-type control, were put on ngm plates with 25 μg/ml carbenicillin and 0.2% lactose. the plates were seeded with e. coli that produced dsrna to knock down expression of ufd-1 and allow assessment of ubg76v-gfp homozygosity. a second assessment was conducted by first synchronizing worms from each strain through bleaching, including the wild-type control, and spotting the eggs onto the rnai plates. the worms were grown at 20℃ and assessed for ubg76v-gfp homozygosity at the l4+24 stage. acknowledgements i am grateful towards dr. christopher rongo for overseeing and supervising this research project and for providing valuable suggestions and directions for the research question. i would also like to express my gratitude towards dr. kishore joshi for his direct supervision of the experiments and for his aid in learning laboratory techniques and background information. i also thank nanci kane, mehul vora, and nathaly salazar for their help with laboratory equipment and procedures. references benenati, g., penkov, s., müller-reichert, t., entchev, e. v., & kurzchalia, t. v. (2009).two cytochrome p450s in caenorhabditis elegans are essential for the organization of eggshell, correct execution of meiosis and the polarization of embryo. mechanisms of development, 126, 382-393. doi: 10.1016/j.mod.2009.02.001 ciechanover, a., & kwon, y. t. (2015). degradation of misfolded proteins in neurodegenerative diseases: therapeutic targets and strategies. experimental & molecular medicine, 47, 1-16. doi:10.1038/emm.2014.117 jing, k., shin, s., jeong, s., kim, s., song, k-s., park, j-h., … lim, k. (2014). docosahexaenoic acid induces the degradation of hpv e6/e7 oncoproteins by activating the ubiquitin–proteasome system. cell death and disease, 5. doi: 10.1038/cddis.2014.477 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 17 joshi, k. k., matlack, t. l., & rongo, c. (2016). dopamine signaling promotes the xenobiotic stress response and protein homeostasis. the embo journal, 35(17), 1885-1901. doi: 10.15252/embj.201592524 keith, s. a., maddux, s. k., zhong, y., chinchankar, m. n., ferguson, a. a., ghazi, a., &b fisher, a. l. (2016). graded proteasome dysfunction in caenorhabditis elegans activates an adaptive response involving the conserved skn-1 and elt-2 transcription factors and the autophagy-lysosome pathway. plos genetics, 12(2). doi: 10.1371/journal.pgen.1005823 lamb, d. c., lei, l., warrilow, a. g. s., lepesheva, g. i., mullins, j. g. l., waterman, m. r., & kelly, s. l. (2009). the first virally encoded cytochrome p450. journal of virology, 83(16), 8266-8269. doi:10.1128/jvi.00289-09 liu, g., rogers, j., murphy, c. t., & rongo, c. (2011). egf signalling activates the ubiquitin proteasome system to modulate c. elegans lifespan. the embo journal, 30(15), 2990-3003. doi: 10.1038/emboj.2011.195 lockhart-jamieson, k., endo, t., darwesh, a. m., samokhvalov, v., & seubert, j. m. (2017). cytochrome p450-derived eicosanoids and heart function. pharmacology & therapeutics, 179, 47-83. https://doi.org/10.1016/j.pharmthera.2017.05.005 o’rourke, e. j., kuballa, p., xavier, r., & ruvkun, g. (2013). ω-6 polyunsaturated fatty acids extend life span through the activation of autophagy. genes & development, 27(4), 429-440. doi: 10.1101/gad.205294.112 powers, e. t., morimoto, r. i., dillin, a., kelly, j. w., & balch, w. e. biological and chemical approaches to diseases of proteostasis deficiency. annual review of biochemistry, 78, 959-991. doi: 10.1146/annurev.biochem.052308.114844 rappleye, c. a., tagawa, a., le bot, n., ahringer, j., & aroian, r. v. (2003). involvement of fatty acid pathways and cortical interaction of the pronuclear complex in caenorhabditis elegans embryonic polarity. bmc developmental biology, 3(8). doi: 10.1186/1471-213x-3-8 whitehouse, a. s., smith, h. j., drake, j. l., & tisdale, m. j. (2001). mechanism of attenuation of skeletal muscle protein catabolism in cancer cachexia by eicosapentaenoic acid. cancer research, 61(9), 3604-3609. wójcik, c., lohe, k., kuang, c., xiao, y., jouni, z., & poels, e. (2014). modulation of adipocyte differentiation by omega-3 polyunsaturated fatty acids involves the ubiquitin-proteasome system. journal of cellular and molecular medicine, 18(4), 590-599. doi: 10.1111/jcmm.12194 aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 18 appendix a definitions of terms ala. ω-3 pufa alpha-linolenic acid. essential fatty acid that cannot be synthesized in the human body. important eicosanoid precursor. ceeh-1 and ceeh-2. catalyze the hydrolysis of short-lived nonclassic eicosanoids to eicosanoids. coded by the ceeh-1 and ceeh-2 genes, respectively. c. elegans. caenorhabditis elegans. free-living nematodes. model species for many experiments. col-19. promoter from which ubg76v-gfp and mrfp are expressed from in the hypodermis. cyp. cytochrome p450. oxidation-reduction enzyme at the end of the electron transport chain. facilitates the conversion of pufas to eicosanoids. dhet and dheq. dihydroxyeicosatrienoic acid and dihydroxyeicosatrienoic acid. derived from ω-6 pufas and ω-3 pufas, respectively. eicosanoid signalling molecules. (fig 1) eet and eeq. epoxyeicosatrienoic acid and epoxyeicosaquatraenoic acid. derived from ω-6 pufas and ω-3 pufas, respectively. short-lived nonclassic eicosanoid signalling molecules converted to eicosanoids by ceeh-1 and ceeh-2. (fig 1) emb-8. reductase enzyme required by cyps to convert pufas to eicosanoids. coded by the emb-8 gene. emb-8(hc69); odis76. emb-8 mutant strain used in the hypodermis. contains a temperature-sensitive allele. typically a lethal mutation. fat-1. desaturase enzyme that converts ω-6 pufas to ω-3 pufas. coded by the fat-1 gene. fat-1 (wa9). fat-1 mutant strain used in the hypodermis. fat-1(wa9); psur5. fat-1 mutant strain used in the intestine. gfp. green fluorescent protein. used as a reporter so that molecules can be imaged and analyzed. ubg76v-gfp. ubiquitin protein, composed of 76 amino acids, with the first being glycine and the final being valine, covalently attached to a gfp to construct an unstable transgene that can be introduced into research specimens. l4 stage. final larval stage of nematodes. l4+24 stage. first adult stage of nematodes. 24 hours since nematode entered l4. l4+48 stage. 48 hours since nematode entered l4. l4+72 stage. 72 hours since nematode entered l4. mufa. monounsaturated fatty acid. fatty acids containing one double bond. precursors to pufas. odis76 (wt). wild-type strain used as the control for the emb-8 mutant in the hypodermis. odis77 (wt). wild-type strain used as the control for the fat-1 mutant in the hypodermis. pp608 (wt). wild-type strain used as the control for the fat-1 mutant in the intestine. pufa. polyunsaturated fatty acid. fatty acids containing more than one double bond. precursors to eicosanoids. ω-3 pufa. contains first double bond at carbon 3. converted from ω-6 pufas by the fat-1 desaturase enzyme. ω-6 pufa. contains first double bond at carbon 6. converted to ω-3 pufas by the aresty rutgers undergraduate research journal, vol. 1, issue 1, spring 2020 19 fat-1 desaturase enzyme. rfp. red fluorescent protein. used as a control reporter so that molecules can be imaged and analyzed. mcherry. rfp reporter used in the intestine. mrfp. rfp reporter used in the hypodermis. sur-5. promoter from which ubg76v-gfp and mcherry are expressed from in the intestine. ups. ubiquitin proteasome system. mechanism through which damaged or unfolded proteins targeted for degradation by the proteasome via ubiquitin tagging. aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. ethical implications of biohacking as activism: democratized health care, danger, or what? julia zheng ✵ abstract biohacking refers to optimizing one’s body through modifying biology. in the 20th century, do-ityourself (diy) biology emerged as a type of biohacking involving biotechnology. current highhealthcare costs promote diy biology insulin and epipens as ways to challenge norms in healthcare, thus serving as forms of activism. biohacked insulin is part of the #wearenotwaiting movement to support improved treatment of type 1 diabetes, whereas biohacked epipens allow people to make lifesaving autoinjectors at low costs. social media acts as a catalyst and aids in the spread of insulin and epipen biohacking as activism. in 1979, principles of biomedical ethics by beauchamp and childress proposed four principles that continue to guide decision-making in clinical medicine: beneficence, nonmaleficence, autonomy, and justice. this paper applies these principles to explore whether the benefits of performing diy biology outweigh the potential health risks. examining biohacking with a biomedical ethics frame, as outlined by beauchamp and childress, reveals that biohacking acts as a response to current issues but cannot serve as a solution in its current form. however, biohacking can grant patients more power in their relationship with the healthcare system, therefore lessening the domi nance of formal institutions. out of the four princi ples, autonomy applies most differently when regarding biohacking than traditional medicine. accordingly, a model of ethics for biohacking, such as of beauchamp and childress’ with the autonomy altered to acknowledge the additional implications of biohacking, should be developed in the future. 1 introduction recent rising healthcare costs and lack of insurance have jeopardized access to health care for many americans. this problem has sparked a movement towards open-source medicine, leading to increased interest in biohacking.[5] biohacking is a broad term that refers to modifying one’s biology in an informal setting. examples of biohacking range from dieting to implanting computer chips in one’s body. specifically, do-it-yourself (diy) biology is a type of biohacking that expands access to individuals, communities, and small organizations studying biology with the same biotechnology as formal research institutions. biohacking has made biotechnology financially and intellectually accessible to those without proper training, who often work out of makeshift labs in their garages or kitchens. pharmaceutical companies and the lucrative nature of the healthcare system have driven many americans to take the issue into their own hands, forcing people to use diy biology to synthesize their medicines or conduct genetic testing. biohacking serves as an act of activism (which can be defined as campaigning to bring about change) against rising healthcare costs, providing otherwise powerless individuals a voice against the unfair practices of the healthcare system. biohacking responds to high healthcare costs, but its viability from the perspective of biomedical ethics must be analyzed when determining if it can be a real solution to current issues. applying a biomedical ethics framework, like beauchamp and childress’ prominent principles of biomedical ethics, serves as a valuable context to explore the ethical implications of biohacking activism. the authors indicate four principles of biomedical ethics: beneficence, nonmaleficence, autonomy, and justice. these principles were first proposed in 1979 and are currently used widely by medical pro aresty rutgers undergraduate research journal, volume i, issue iii fessionals and ethical review boards when making healthcare decisions. beneficence refers to health care providers preventing harm and working to benefit the patient. nonmaleficence holds healthcare professionals to a “standard of due care…taking sufficient and appropriate care to avoid causing harm, as the circumstances demand of a reasonable and prudent person,” in which “reasonable” refers to one who approaches a situation with caution and sensibly takes action[1]. nonmaleficence requires medical professionals to prevent harming patients purposefully or negligently. autonomy allows patients the ability to make informed, voluntary decisions without controlling influences. justice in healthcare refers to the equal treatment of patients, regardless of money, age, or race. regarding biohacking, however, one must note that the biohacker is the patient and the medical professional, further complicating this issue. this paper will apply the bioethical framework of beauchamp and childress to biohacked insulin and epipens to weigh the benefits and risks of biohacking as a potential solution to high costs imposed by pharmaceutical companies. first, it will discuss the ethical implications of biohacking activism concerning insulin for type 1 diabetes. next, it will examine the diy epipen as a form of biohacking activism for those with allergies. then, it will compare autonomy in its traditional sense with autonomy in biomedical ethics. continuing the discussion of autonomy, it will further explore its applications and nuances regarding parents and children before concluding with a proposed plan to promote future ethical activity. 2 insulin activism for patients with type 1 diabetes and those with type 2 diabetes who do not produce enough insulin, insulin therapy is necessary to keep blood sugar within a target range, prevent hypoglycemia and ketoacidosis, as well as ultimately sustain and enhance life. however, continuously rising insulin costs have presented a barrier to patients with diabetes from receiving this medication. in the past twenty years, the price of insulin has increased from $21 per vial to $332 per vial in 2019, a more than 1000% increase. dr. s. vincent rajkumar of the mayo clinic comments on this issue: “there is limited innovation when it comes to insulin; the more pressing need is affordability.”[12] gallegos agrees with rajkumar that these high costs present a concern and underscores instances of people “rationing treatments, using expired products, fasting, and even intentionally inducing diabetic ketoacidosis in order to obtain insulin from hospital emergency rooms.”[5] these examples show that current healthcare conditions paradoxically push people to harm themselves in order to save their own lives. beauchamp and childress would view the issue presented by rajkumar and gallegos as a violation of the principle of justice, particularly distributive justice.[1] rajkumar agrees, stating that “insulin pricing in the united states is the consequence of the exact opposite of a free market: extended monopoly on a lifesaving product in which prices can be increased at will…”[12] all four of these authors would argue that pharmaceutical companies behave unethically regarding the distribution of insulin, creating an environment that pushes people to engage in biohacking as activism. although insulin is not currently successfully synthesized using diy biology, multiple organizations conduct experiments to move towards this goal. for instance, the open insulin project is a collaboration of community biolabs worldwide that is currently attempting to produce insulin more cheaply than pharmaceutical companies. once widespread, biohacked insulin could act as a life-saving alternative for patients who cannot afford traditional insulin. one main difference between formally produced and biohacked insulin is the rigorous testing a new drug must undergo in the united states, which “cost[s] between $30 and $250 million.” thus, regulatory costs create a major barrier to entry for potential biosimilar producers and necessitate high drug prices to recoup investments spent on clinical trials”.[5] biohacked insulin can avoid extensive testing, but this lack of assurance results in risks of unintentional poisoning for consumers. the lack of required training for people to perform diy biology also presents a risk. for instance, a biohacker working for the open insulin project was described as “frequently check[ing] the instructions on his aresty rutgers undergraduate research journal, volume i, issue iii smartphone, because he is not too familiar with this kind of work yet…[he] has a major in business economics.”[10] this behavior implies an unfamiliarity with the science, further highlighting the dangers of being unvetted and underprepared. gallegos and osterath demonstrate the trade-off between safety and financial accessibility that biohackers must consider. the problematic situation that people with diabetes find themselves in jeopardizes their ability to obtain a life-sustaining medication, resulting in desperation and fear of death that enables immense psychological discomfort. this situation prevents biohackers from acting as “reasonable” people and thus fully demonstrating nonmaleficence. furthermore, trained professional workers in healthcare fulfill the principle of nonmaleficence by relying on their extensive training, following strict protocols, and practicing wise judgment to ensure nonmaleficence for their patients; this also includes obtaining medications and supplies from a trusted, formal source such as a pharmaceutical company. one could also argue that beauchamp and childress’ “due care” refers to taking the proper, rather than biohacked, medication. due to the lack of testing and regulation of diy insulin in its current state, people cannot ensure the same degree of nonmaleficence as when taking a formally produced drug, presenting significant risks. on a small scale, those who synthesize and take biohacked insulin aim to benefit themselves, disqualifying these actions from serving as activism. however, the diy synthesis of insulin, automated insulin delivery (aid) devices, and the spread of the movement using social media allow biohacking to serve as a form of activism and as a tool to promote change in diabetes treatment. there is a distinction between type 1 and type 2 diabetes. type 2 diabetes often develops due to lifestyle factors, whereas type 1 diabetes is mainly caused by genetics and typically affects people throughout their lives, starting from an early age. these conditions force type 1 diabetics to continuously rely on the healthcare system and create a vulnerability that allows pharmaceutical companies to capitalize on patients with no control over their lifelong disease. frustration in the healthcare system and the status of diabetes treatment has led to the #wearenotwaiting movement, which consists primarily of type 1 diabetic individuals who advocate for increased access to insulin delivery technology through making their own aid devices. one advocate, timothy omer, has had type 1 diabetes for more than 22 years and currently works to develop a novel artificial pancreas system. he describes the reality of being a type 1 diabetic: the most modern accessible technology for type 1 diabetes management is an insulin pump, which provides a constant supply of insulin, as well as a self-funded continuous glucose monitor, which provides real-time feedback of the patient’s blood sugar levels. these devices provide many functions and high volumes of data, all of which are very welcome and useful, but such systems always fail with regards to patient expectations to understand and process all of this information. as a result, patients become overwhelmed by a feeling of judgement by healthcare professionals, the vast amounts of ensuing information, and alert and alarm “shouts” from their devices when they have failed at being a “good diabetic”, as well as with their own disappointment of their body letting them down.[9] omer demonstrates that he and others are tired of waiting for a change in how type 1 diabetes treatment is approached. this frustration has pushed them to engage in biohacking in order to regain agency over their health condition. the physical act of making the aid device empowers patients to feel as though they are regaining power after years of feeling controlled by the healthcare system, and the low prices of this equipment make it increasingly financially accessible and allow it to act as a catalyst for the spread of the movement. one prime example is one mother’s model of a diy artificial pancreas for her daughter: “sydney, now 15, is still using an updated version of that diy system, which, because a fellow diyer donated the pump, cost only $250 to make…apple inc. and eli lilly & co. have hired diyers, and medtronic’s latest fda-approved product can now do most of the things the farnsworths’ system can—for $7,000, before insurance…”[6] not only is the act of making one’s device (previously reserved for healthcare professionals) shocking, but the significantly lower prices allow the #wearenotwaiting movement to draw media attention and gain support. although only a tiny portion of the population lives with type 1 diabetes, almost aresty rutgers undergraduate research journal, volume i, issue iii everyone can relate to feeling frustrated or upset at high medical costs. also, regardless of social class and safety factors, people would favor lower healthcare costs, making it easy for the #wearenotwaiting movement to draw support. previously, pharmaceutical companies held power to impose high prices, knowing that patients would likely try to obtain these devices or medications at all costs. however, if biohacking continues to grow, it could provide an alternative way for people to obtain medicines, threatening the long-held domination of the healthcare system. this change would represent a revolution in the way people view healthcare, allowing biohacking insulin on a large scale to serve as activism. further, biohacking not only democratizes medicine by making it increasingly financially accessible but also simplifies science so that people without advanced degrees can intellectually comprehend the processes, fulfilling beauchamp and childress’ principle of distributive justice. they indicate: “distribution of all rights and responsibilities in society, including, for example, civil and political rights.”[1] perhaps it is time to provide patients who suffer from these lifelong conditions, rather than the monopolistic healthcare system, a greater say in their treatment; biohacking offers them the opportunity to do so. 3 diy epipen although biohacking insulin and epipens are similar in that they are critical to the lives of those with diabetes and life-threatening allergies, epipens differ from insulin regarding cost, safety, and the main ethical principles that pertain to situations in which one requires these medications. biohacking insulin requires the actual medicine to be synthesized while biohacking epipens does not. in contrast, only the injector is made for epipens. this cost difference would lead to varying levels of accessibility. for instance, synthesizing insulin requires not only chemicals that could prove unattainable for many but it also requires the intellectual and scientific knowledge and ability to synthesize this medication. this process can be contrasted with that of the diy epipen, which can be made using “off-the-shelf” parts for as low as $30, making the process intellec tually more accessible to those without a scientific background.[4] one prominent distinction between diy insulin and epipen is that the epinephrine in a diy epipen can still come from a formal supplier, resulting in differences regarding safety. while diy insulin contains a risk of poisoning oneself, the use of formally produced epinephrine in diy epipens reduces the risk of physical harm to individuals in comparison. although diabetes and allergies are both chronic conditions, there is the distinction of epipens as necessary upon exposure to an allergen to prevent immediate death. beauchamp and childress highlight this concept in their principle of beneficence, which includes “rescue persons in danger.”[1] people with diabetes need insulin over time to prolong their lives, but the immediate necessity of an epipen upon exposure to an allergen makes the danger of being without the life-saving device more imminent. the low costs, intellectual accessibility, and relative safety of these diy epipens combined with the inherent, natural compulsion for humans to help other humans in cases of immediate need promote the diy epipen movement as a form of activism. in 2016, the price increase in the mylan epipen received media attention and backlash from consumers. the increasing prices had been an issue, increasing from “…$103.50 for a set. by july 2013, the price was up to $264.50, and it rose 75 percent to $461 by last may. this may the price spiked again to $608.61…”[11] as a response to these prices, four thieves vinegar, a biohacking group founded in 2015, posted a youtube video demonstrating how to inject oneself with a homemade epinephrine auto-injector and published a list of materials on how to do so. donovan states that “shortly after the release of the video, other biohacker groups and diyers began offering epinephrine auto-injector alternatives online. progressth (an international design lab) released a statement announcing the development of a 3d concept for an at-home alternative auto-injector, which would be printed for as little as $3 in materials.”[2] although these biohackers primarily publicize their methods to help those with lifethreatening allergies stay safe, they expose the healthcare industry’s issues and demonstrate the aresty rutgers undergraduate research journal, volume i, issue iii dire need for change, qualifying them as accidental activists. the cost of biohacked epipens is meager compared to traditional epipen costs, ranging from $600 to $700; even those without financial strain would likely consider it an option to save money. however, these videos demonstrate how cheap the materials in the epipen are. suppose an average person were to figure out a way to produce the injector as cheaply as $3. in that case, corporations almost certainly have ways to make them even cheaper, exposing the inflated prices. companies that manufacture these devices could charge less and still make a profit, but they choose not to as a way to capitalize on people’s vulnerability. instead, this selfishness forces people to risk harming themselves using a biohacked injector to save their lives. the shocking idea of producing one’s own epipen resulted in the idea’s spread through the media, which would only result in positive reinforcement for the movement’s growth. increasing numbers of posts and videos on the media pertaining to diy epipens normalizes the concept and reduces its stigma. as a result, more people are likely to try to make their own epipens, further propagating the movement. for the rest of society, this movement’s media attention raises the question of what other medications corporations are overcharging for, sparking frustration against the healthcare industry, and fueling a more significant activist movement. in financial terms, saving money through producing one’s epipens would prevent a person from potentially overspending beyond his or her resources, fitting under the bioethical principle of beneficence. beauchamp and childress also emphasize the physical aspects of beneficence: …a person x has a determinate obligation of beneficence towards person y if and only if each of the following conditions is satisfied (assuming x is aware of the relevant facts): 1. y is at risk of significant loss of or damage to life or health or some other major interest. 2. x’s action is needed (singly or in concert with others) to prevent this loss or damage. 3. x’s action (singly or in concert with others) has a high probability of preventing it. 4. x’s action would not present significant risks, costs, or burdens to x. 5. the benefit that y can be expected to gain outweighs any harms, costs, or burdens that x is likely to incur.[1] in biohacking, however, x and y would usually be the same person, and in certain instances, a parent and child (a concept that i will later explore). number 1 is satisfied by a person being in anaphylactic shock, and number 2 represents an epipen’s nature. yet, the other points must be examined. regarding numbers 3 and 5, one could argue that diy epinephrine injectors have high risks compared to those formally produced. for instance, willingham states, “a syringe doesn’t offer the benefit and safety advantage of a well-calibrated dose, and it carries the risk of injection into a vein, instead of muscle, which can be fatal.” with an untrained individual administering a medical technique that he or she does not have formal training in, there are risks. yet, the necessary medication must be administered to patients with diabetes or allergies in the case of imminent death. strikingly, the fourth idea is one of the main benefits of biohacking. for one person, the cost of epipens for her son was “more than her mortgage payment,” and “her older son…just carries around expired epipens.”[13] parents are forced to make the difficult decision of risking their child’s life or overstretching their financial needs, and biohacking allows them to prevent both. the high and unattainable costs of insulin and epipens create desperation in patients, resulting in the sense of hopelessness that prices will become unattainable and ultimately an overwhelming fear of death without the medication. desperation often pushes people to do things they would not otherwise consider, such as synthesizing their own medicines. biohacking serves as a form of activism: while trying to save their own lives, biohackers also promote change in the way people view the severity of high healthcare costs. biohacking’s attention, especially when spread using social media, allows others to support the movement and promote social and economic change, qualifying biohackers as activists. in the hands of the healthcare system, people often have no choice but to follow the rules and policies put in place. however, biohacking allows these individuals to act and change the status quo. therefore, biohacking serves as a physical representation of the changes people have wanted to make in the healthcare system. although the act of biohacking does not directly result aresty rutgers undergraduate research journal, volume i, issue iii in fair healthcare practices and prices, perhaps its shocking nature is enough to generate attention and support to move towards change. 4 autonomy: two different types? autonomy, among beauchamp and childress’ four principles of biomedical ethics, requires more examination due to the distinction between the traditional definition of autonomy and medical autonomy. biohackers’ ability to synthesize their insulin and epipens grants them a powerful sense of independence that they did not previously have; they are autonomous in its traditional sense, meaning independent. however, beauchamp and childress identify medical autonomy: “…in terms of normal choosers who act (1) intentionally, (2) with understanding, and (3) without controlling influences that determine their action…a broad continuum exists on which autonomy goes from being fully present to being wholly absent.”[1] in medicine, autonomy must involve the patient demonstrating understanding of a treatment or procedure, and medical professionals must adequately inform their patients. however, when people become biohackers due to desperation, they face no obligation to understand the risks of their actions thoroughly. without being medically autonomous, biohackers are inhibited from acting ethically when examined with beauchamp and childress’ model. the nature of biohacking facilitates impulsivity: “the ways in which diy biology is ‘governed’ or ‘regulated’ takes a distinctive form: rather than being top-down it is bottom-up; rather than being defined by institutions or policymakers, it is collectively and openly negotiated by a large group of people…”[7] because biohacking relies on the public to regulate themselves rather than formal rules, it can be easily misused, proving extremely dangerous. on the other hand, though, many advancements in science have begun with experimentation, leading to ideas that may take many years for the public to accept. perhaps biohacking is just one of these ideas that currently seems outlandish. as the concept continues to gain momentum and more safety measures are enacted, biohacking can be the next stage of scientific development. 5 applications of autonomy the high costs of necessary technology combined with the monopolistic nature of diabetic health care can cause people to feel powerless at the hands of the healthcare system. as a result, making their own technology grants them a feeling of financial and personal autonomy. currently, there is little to no formal regulation preventing this self-experimentation. although there are no proper laws against biohacking insulin devices in the united states, the german model serves as an example of the current regulations: “healthcare professionals must point out the dangers that may arise when using a diy aid system and should clearly distance themselves from the use of an open system and not encourage patients to use the system.”[8] of course, healthcare professionals are worried about their patients, but they also benefit from higher medications and treatment costs. healthcare professionals fail to advocate for cheaper alternatives, leaving patients with no one to fight for them. patients must engage in activism to promote the change they want, and biohacking is one way for them to do so. the lack of limitation from governments, healthcare professionals, and finances grants patients autonomy from a biomedical ethics perspective. many patients who feel “overwhelmed” by the health care system are now free from this feeling. beauchamp and childress state: “we encounter many problems of autonomy in medical contexts because of the patient’s dependent condition and the medical professional’s authoritative position… in these instances, the patient’s autonomy may be compromised because the physician has assumed an unwarranted degree of authority over his or her patient.”[1] not only do health care professionals hold authority over patients under the traditional health care system, but the healthcare industry controls the price and distribution of insulin and devices. however, health care professionals lack the right to prevent patients from making and using diy aid systems. therefore, when juxtaposed with the legal inability of healthcare professionals to prevent patients from using diy biology to create their aid systems, patients are now free from what they feel is domination by the health care system. aresty rutgers undergraduate research journal, volume i, issue iii 6 autonomy of parents & children the issue of people using biohacking on themselves is less controversial than people performing biohacking on others. specifically, the case of parents using biohacked medicine and devices on their children sparks controversy. because the children are dependent, they do not have the authority to refuse their parents nor the resources to obtain the medication for themselves through formal means. the extent to which parents must follow the principles of biomedical ethics must be examined. for instance, beauchamp and childress identify negligence as a subcategory of nonmaleficence: “negligence is the absence of due care. in the professions, it involves a departure from the professional standards that determine due care in a given set of circumstances. the term negligence covers two types of situations: (1) intentionally imposing risks of harm that are unreasonable (advertent negligence or recklessness) and (2) unintentionally, but carelessly, imposing risks of harm (inadvertent negligence).”[1] if parents do not attempt to obtain the insulin or epipen their child needs through formal means or biohacking, this could be regarded as negligent. just as medical professionals are required to follow the widely agreed-upon rules and procedures to ensure the safety of their patients, a parallel can be drawn to the parents, in which they are expected to provide their children with the safest medications possible—those produced by formal institutions. therefore, parents synthesizing pharmaceuticals using biohacking could also be considered negligent, where “due care” includes obtaining official medicine. the barriers to this situation should not be disregarded. perhaps it is the pharmaceutical or insurance companies that are fostering this negligence, and they should therefore bear more responsibility. biohacking also introduces an additional risk of harm for the children. for instance, the german laws for diy aid systems indicate that “people who build diy aid systems and transfer them to other patients are liable to prosecution under the medical devices act in germany. the placing on the market and commissioning of such a system are prohibited. the person who builds and transfers the system is responsible under the product liability act.”[8] when biohacking for their children, parents take on additional ethical responsibility that, not being healthcare professionals, they may not be able to uphold. these parents likely never wanted this responsibility, but healthcare conditions forced them to act. these drastic measures show the public how dire the situation is and characterize these parents as activists. 7 conclusion the spread of biohacking techniques through social media and the internet can alter science and medicine, resulting in biohacking serving as a form of activism. one primary example of change brought about after the increase in biohacking is the reduced costs of the epipen, including the “release of a generic version of its device at about half the cost (about $340) of the brand name epipen…cvs also announced that the manufacturer would provide a $100 coupon for much of its financially insecure population to reduce the out-ofpocket cost to about $10 per prescription.”[2] it is possible that the media attention received by biohackers threatened pharmaceutical companies, as these companies feared losing profit as more people attempted to biohack. nevertheless, the biohackers who spearheaded the diy epipen movement achieved social and economic change, although they risked their lives in the process, representing activism. perhaps these corporations will, in the future, be wary before imposing high and unattainable prices on consumers for fear of driving more people to biohack. if increasingly accepted by results indicate that most statefunded preschool programs have a long way to go if increasingly accepted by the public, biohacking can drastically change the way people view the relationship between patients and the healthcare system. aresty rutgers undergraduate research journal, volume i, issue iii the public, biohacking can drastically change the way people view the relationship between patients and the healthcare system. dyson proposes the future of biohacking: there will be do-it-yourself kits for gardeners who will use genetic engineering to breed new varieties of roses and orchids. also kits for lovers of pigeons and parrots and lizards and snakes to breed new varieties of pets. breeders of dogs and cats will have their kits too…few of the new creations will be masterpieces, but a great many will bring joy to their creators and variety to our fauna and flora. the final step in the domestication of biotechnology will be biotech games, designed like computer games for children down to kindergarten age but played with real eggs and seeds rather than with images on a screen. playing such games, kids will acquire an intimate feeling for the organisms that they are growing. the winner could be the kid whose seed grows the prickliest cactus, or the kid whose egg hatches the cutest dinosaur.[3] dyson emphasizes the idea that biohacking will “bring joy” and that eventually, people will perform biohacking for pleasure, in contrast to something that is currently performed out of desperation. although biohacking is still a new topic, the media attention it receives (coupled with its low costs) can lead to the democratization of healthcare and science. in its current state, biohacking presents a tempting option for those who seek to avoid the high costs of the traditional healthcare system. yet, the risks of this unregulated experimentation deter many rationally thinking people. however, in the future, if ethical and safety regulations are put in place, people begin devising new ways to use biohacking and posting these ideas on the media; biohacking can break financial and intellectual bounds. although the first amendment protects free speech, sites can employ community rules or rules of conduct to promote the safety of biohacking. no longer would people feel helpless under a system that controls their health. if perfected and made applicable to more situations, biohacking could challenge the long-standing dominance of formal institutions and grant more power to the average person. in activism, people often seek the traditional definition of justice, which typically means fairness. biohackers fall under this category and aim to achieve fair prices and access to medications. however, just as there are multiple definitions of autonomy, there are also different forms of justice; beauchamp and childress define the concept of justice in a biomedical ethics context slightly differently. they describe their idea of material justice: “…primarily our obligations are limited to fundamental needs. to say that someone has a fundamental need is to say that the person will be harmed or detrimentally affected in a fundamental way if that need is not fulfilled.”[1] biohacking finally allows ordinary people to bring attention to the message that their needs are not met. although this purpose was likely not intentional, by democratically synthesizing one’s medication, biohackers also fulfill the idea of justice from a biomedical ethics perspective, aside from simply seeking justice in its traditional sense. the idea of biohacking, or performing experiments once only reserved to formal institutions, can be shocking and disturbing to many. biohacking has the potential to go wrong and harm people. yet, in the case of chronic medical conditions such as diabetes and allergies, biohacking offers a sense of control and a glimmer of hope for those put in a challenging financial and medical situation. when examining types of activism from rallies to boycotts to hunger strikes, all of these examples share one common feature with biohacking: they are striking to the public. real, lasting change rarely comes from mundane activities, so perhaps the dangerous nature of biohacking is critical to it serving as activism. also, low costs and media attention aid in biohacking spreading. when examining these applications of biomedical ethics using beauchamp and childress’ model, many nuances prevent biohacking from serving as a straightforward solution to high healthcare costs. however, because their principles apply to traditional medicine rather than biohacked treatments, these principles should be updated to reflect the ethical concerns over the increasing popularity of biohacking. without an exact code of instructions, biohackers are left to their own devices. although biohackers have developed their code of ethics consisting of broad principles such as responsibility and transparency, this unofficial code resembles more of a list than rules for biohackers to follow. when juxtaposed with beauchamp and childress’ model, the biohacker code of ethics is concise and nonspecific. aresty rutgers undergraduate research journal, volume i, issue iii therefore, perhaps it is time to develop a model of ethics for biohackers. out of the four principles, beneficence, nonmaleficence, and justice can remain largely the same. however, due to the freedom granted by individuals performing biohacking on themselves, the principle of autonomy should be altered to emphasize the importance of a complete understanding of risks and rewards before one performs biohacking to ensure its ethical soundness as increasing numbers of people engage in biohacking in hopes of saving lives and promoting change∎ 8 references [1] beauchamp, tom l., and childress, james l. principles of biomedical ethics. 8th ed. oxford university press, 2019. [2] donovan, matthew c. j. pop-up maktivism: a case study of organizational, pharmaceutical, and biohacker narratives. 2019. arizona state university. phd dissertation. [3] dyson, freeman. “our biotech future.” new york review of books, vol. 54, no, 12, 2007. [4] four thieves vinegar, accessed: 6 dec. 2020, https://www.fourthievesvinegar.org [5] gallegos, jenna e., et al. "the open insulin project: a case study for 'biohacked' medicines." trends in biotechnology, vol. 36, no. 12, 2018, pp. 1211-1218. [6] kresge, naomi and cortez, michelle. “the $250 biohack that’s revolutionizing life with diabetes.” bloomberg, 2018. [7] landrain, thomas et al. "do-it-yourself biology: challenges and promises for an open science and technology movement." systems of synthetic biology, vol. 7, no. 3, 2013, pp. 115-126. [8] oliver, nick et al. “open source automated insulin delivery: addressing the challenge.” digital medicine, no. 124, 2019. [9] omer, timothy. “empowered citizen ‘health hackers’ who are not waiting.” bmc medicine, vol. 14, no. 118, 2016. [10] osterath, brigitte. “do-it-yourself insulin: biohackers aim to counteract skyrocketing prices.” dw, 2019, www.dw.com/en/do-it-yourself-insulin-biohackers-aim-tocounteract-skyrocketing-prices/a-48861257. [11] parker-pope, tara and peachman, rachel rabkin. “epipen price rise sparks concern for allergy sufferers.” the new york times, 2016. https://well.blogs.nytimes.com/2016/08/22/epipen-price-rise-sparksconcern-for-allergy-sufferers [12] rajkumar, s. vincent. “the high cost of insulin in the united states: an urgent call to action.” mayo clinic proceedings, vol. 95, no. 1, 2020. [13] willingham, emily. “why did mylan hike epipen prices 400? because they could.” forbes, 2016. https://www.forbes.com/sites/emilywillingham/2016/08/21/whydid-mylan-hike-epipen-prices-400-because-they-could/ julia zheng is a junior at rutgers university-new brunswick pursuing a bachelor of science in cell biology and neuroscience in the school of arts and sciences honors program with a minor in psychology. she conducted research for this paper in her research in the disciplines: science, medicine, and society course in fall 2020. julia is interested in science and medicine and has the goal of pursuing a career in the medical field as a physician. at rutgers, julia is president of the bioethics society, conducts research in dr. alexander kusnecov's lab, and works as a part-time lecturer for the physics department. outside of school, she works as a medical assistant and receptionist at a pediatrician's office and enjoys volunteering in her community. https://www.fourthievesvinegar.org/ http://www.dw.com/en/do-it-yourself-insulin-biohackers-aim-to-counteract-skyrocketing-prices/a-48861257 http://www.dw.com/en/do-it-yourself-insulin-biohackers-aim-to-counteract-skyrocketing-prices/a-48861257 https://well.blogs.nytimes.com/2016/08/22/epipen-price-rise-sparks-concern-for-allergy-sufferers https://well.blogs.nytimes.com/2016/08/22/epipen-price-rise-sparks-concern-for-allergy-sufferers https://www.forbes.com/sites/emilywillingham/2016/08/21/why-did-mylan-hike-epipen-prices-400-because-they-could/ https://www.forbes.com/sites/emilywillingham/2016/08/21/why-did-mylan-hike-epipen-prices-400-because-they-could/ aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. pulmonary inflammation & injury in a mouse model of non-alcoholic steatohepatitis tanvi banota, alexa murray, laura e. armstrong, bo kong, grace l. guo, andrew j. gow, debra l. laskin (faculty advisor) ✵ abstract non-alcoholic fatty liver disease (nafld) is a chronic liver condition that affects millions of individuals in the united states, of which approximately twenty percent of cases progress to non-alcoholic steatohepatitis (nash). nash is characterized by macrovascular steatosis and persistent inflammation in the liver, which can lead to fibrosis. evidence suggests potential effects of nafld and nash on the development of pulmonary pathologies, but the interaction between the liver and the lung is not well understood. in this study, we assessed the impact of nash development on lung inflammation and fibrosis over time. male c57bl/6j mice were fed control (10% kcal) or high-fat (hfd) (60% kcal) diets. liver tissue, lung tissue, and bronchoalveolar lavage (bal) fluid were collected after 1, 3, and 6 months of feeding. histopathologic evaluation of livers from hfd-fed mice at 6 months confirmed the development of nash. in the lung, we observed histopathologic alterations, including inflammatory cell infiltration, lipid-laden macrophages, septal damage, and epithelial thickening at 6 months. gene expression analysis of whole lung tissue revealed changes in genes related to inflammation (il-1b), fibrosis (ctgf), and lipid metabolism (apoa1). these results characterize an association of pulmonary complications during simple steatosis to nash transition, suggesting lung-liver crosstalk. 1 introduction non-alcoholic fatty liver disease (nafld) is a chronic liver condition that is estimated to affect upwards of thirty percent of individuals in the united states, with even more at risk due to the rising obesity epidemic.[3,5,19] nafld is characterized by the accumulation of fat in the liver, known as steatosis. it is estimated that twenty percent of patients diagnosed with nafld progress to non-alcoholic steatohepatitis (nash).[21] nash is a severe, chronic liver disease characterized by persistent inflammation and immune cell infiltration. nash may also progress to fibrosis and cirrhosis of the liver—this irreversible scarring of tissue can further lead to uncontrolled cell growth, cancer, and death. nash is becoming the leading indication for liver transplantation in both the u.s. and worldwide.[2,14,16] both nafld and nash are associated with systemic effects that manifest in pathologies across the body, including cardiovascular disease, metabolic syndrome, and chronic kidney conditions.[1] emerging evidence also suggests that nafld and nash may impact the development of pathologies in the lung. several longitudinal observational studies have detailed an association between nafld and decreased measurements of lung function (e.g. forced expiratory volume and vital capacity).[7,11,12,15] in addition, patients with chronic obstructive pulmonary disease (copd) showed increased incidence of both nafld and nash.[20] although there is rising clinical evidence of a relationship between nash and reduced pulmonary function, the interplay between the liver and the lung remains largely unexplored. a central aspect of nafld and nash that may contribute to lung injury is inflammation. key players in this response are macrophages, phagocytic cells of the innate immune system. in the liver, macrophages are known to take up surrounding fat through endocytosis. this causes the macrophages to become activated and release pro-inflammatory mediators such as cytokines, small proteins important in inflammatory cell signaling.[14] these mediators enter the bloodstream and can exert effects in other tissues of the body through systemic circu aresty rutgers undergraduate research journal, volume i, issue iii lation, leading to observed comorbidities.[1,9,14] we hypothesize that these inflammatory mediators accumulate in the lung, causing pulmonary injury, inflammation, and the disruption of key signaling pathways by dysregulating genes related to lipid metabolism and inflammation, including il-1b, ctgf, lxr, apoa1, and abca1. the present study was designed to test this hypothesis and characterize the development of lung injury in a high fat diet mouse model of nash. the results of our study provide data on lungliver crosstalk, which may be useful for the development of new approaches for clinical management of pulmonary pathology related to nash. 2 methodology animals & treatments wild type male c57bl/6j mice (6-8 weeks, 𝑛𝑛 = 5 − 9/group) were fed control (10% kcal) or a high-fat diet (hfd) (60% kcal) for 1, 3, and 6 months. food consumption was monitored weekly. animal protocols were approved by rutgers university iacuc. histological analysis liver and lung tissue were collected, fixed in 10% formalin or inflated and fixed in 4% paraformaldehyde, respectively, and cut into 5-μm sections and stained with hematoxylin & eosin (h&e). lung sections were evaluated for characteristics of lung injury and inflammation, including inflammatory cell infiltration, septal damage, and epithelial thickening. liver sections were examined for the histopathological characteristics of nafld (e.g. fat accumulation) and nash (e.g. fat accumulation and inflammatory cell infiltration) based on established criteria.[2,16] bronchoalveolar lavage (bal) cell and protein measurement bal fluid was collected by slowly instilling and withdrawing 1 ml of ice-cold (4°c) pbs into the lungs of mice through a cannula in the trachea (figure 1). this fluid was centrifuged at 300xg for 8 minutes. cell pellets were resuspended in 1 ml of pbs. viable cells (10 μl) were counted on a hemocytometer using trypan blue dye exclusion. cytospins were prepared by centrifugation of 104 cells bal fluid onto microscope slides using a shandon cytospin (thermo scientific). cells were fixed in methanol and stained with giemsa to visualize bal cell populations. total protein content in cell free bal was quantified using a bca protein kit (pierce biotechnologies inc.) with bovine serum albumin as the standard. all samples were assayed in triplicate at 562 nm using a spectrophotometer. mrna isolation and rt-qpcr analysis total rna was extracted from lung tissue using trizol reagent and bead tissuelyser lt (qiagen). cdna was generated using high capacity cdna reverse transcription kit (applied biosystems). realtime quantitative pcr (rt-qpcr) was performed on a quantstudio 6 system using commercially available power sybr® green gene expression assays (applied biosystems). data were normalized to β-actin and presented as fold change relative to 1 month ctrl mice. fold changes in gene expression were calculated using ∆∆𝐶𝐶𝐶𝐶 method where 𝐶𝐶𝐶𝐶(𝐶𝐶𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝐶𝐶 𝑡𝑡𝑡𝑡𝑛𝑛𝑡𝑡) – 𝐶𝐶𝐶𝐶(𝛽𝛽 − 𝑡𝑡𝑎𝑎𝐶𝐶) = ∆𝐶𝐶𝐶𝐶; ∆𝐶𝐶𝐶𝐶 – 𝑡𝑡𝑎𝑎𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ∆𝐶𝐶𝐶𝐶(1 𝑀𝑀𝑀𝑀𝑛𝑛𝐶𝐶ℎ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶) = ∆∆𝐶𝐶𝐶𝐶; and 2 − ∆∆𝐶𝐶𝐶𝐶 = 𝑓𝑓𝑀𝑀𝑓𝑓𝑓𝑓 𝑎𝑎ℎ𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡. statistical analysis data are presented as mean + se and were analyzed using 2-way anova and sidak’s multiple comparisons test. a p-value ≤0.05 was considered statistically significant. 3 results high-fat diet induces nash and causes histopathological changes in the lung to confirm the development of nash, we assessed histopathological changes in the liver. livers from mice fed a hfd for 1 and 3 months exhibited steatosis, or the accumulation of lipid droplets in the liver, but no inflammatory cell infiltration, indicating nafld (data not shown). the most prominent changes in the liver were observed 6 months following consumption of a hfd; these included more severe steatosis and infiltration of inflammatory cells figure 1 aresty rutgers undergraduate research journal, volume i, issue iii (figure 2a). significant increases in body weights, increases in total serum cholesterol, and decreased glucose tolerance in hfd-fed mice confirmed this was metabolic syndrome-related nash (data not shown). we next assessed alterations in lung histology. after 6 months, inflammatory cell infiltration, septal damage, and epithelial thickening were observed in hfd-fed mice relative to mice fed the con trol diet (figure 2b). the accumulation of large macro phages in the lung that appeared lipid-laden was also noted in the histology. further examination of cell cytospins also revealed the presence of large, vacuolated, and potentially lipid-laden macrophages (figure 2b inserts). we further investigated lung injury and inflammation by quantifying levels of bal protein and cells, respectively (figure 2c). alt figure 2: representative images of h&e stained sections of liver (panel a) and lung (panel b) from mice fed control (ctrl) or high fat diets (hfd) for 6 months. panel a, left arrow indicates steatosis, right arrow and inset indicates inflammation. panel b, top arrow indicates inflammatory cell infiltration and bottom arrow indicates epithelial thickening. panel b insets highlight representative macrophages from giemsa-stained cytopsins, including macrophages with a large, lipid-laden appearance in hfd-fed mice. original magnification (a) 4x or (b) 20x, inserts 40x. (c) bal collected from mice 1, 3, and 6 months after a control (ctrl) or high fat diet (hfd) was assessed for protein and cell content. bars, mean + se (n=5-10). *significantly different from ctrl fed mice. #significantly different from 1 month. asignificantly different from 3 month. aresty rutgers undergraduate research journal, volume i, issue iii hough increased lung injury was noted in hfd-fed mice at 6 months as measured by bal protein, this may be due to a reduction in bal protein in control mice at this time. surprisingly, a time related increase in lung inflammation was observed in control mice at 3 and 6 months. cell counts were unaffected by administration of the hfd. high fat diet disrupts expression of inflammatory and lipid metabolism related genes to better understand degrees of inflammatory changes in the lung following a hfd, we analyzed expression of inflammatory and lipid-related genes. expression of interleukin 1 beta (il-1b), a key early response proinflammatory gene, was significantly increased in mice fed a hfd when compared to con figure 3: lung tissue collected 1, 3, and 6 months after control (ctrl) or high fat diet (hfd) from mice were analyzed for gene expression by rt-qpcr. data were normalized relative to ß-actin and presented as fold change relative to 1 month ctrl fed mice. bars, mean + se (n=3-7). *significantly different from ctrl fed mice. #significantly different from 1 month. asignificantly different from 3 month. aresty rutgers undergraduate research journal, volume i, issue iii trol mice at 3 months. il-1b expression returned to control levels by 6 months. the expression of connective tissue growth factor (ctgf), a gene indicative of fibrosis and tissue remodeling, was significantly decreased 1 month following hfd-feeding. interestingly, control mice displayed a significant decrease in ctgf at 3 and 6 months when compared to 1 month. similarly, expression of apolipoprotein a1 (apoa1), a lipid chaperone, was decreased 1 month following hfd when compared to control-fed mice. apoa1 expression was similarly reduced in the 6month control-fed mice when compared to 1-month control mice. there were no changes in expression of liver x receptor (lxr), a nuclear receptor involved in lipid homeostasis, and its target atp-binding cassette transporter 1 (abca1), a lipid transporter. 4 discussion the inflammatory and fibrogenic effects of nash in the liver have been well characterized.[2,14] emerging clinical evidence suggests that nafld and nash are associated with pulmonary injury, but crosstalk between the lung and the liver in nash has not been investigated.[7,11,12,15] in these studies, we used a mouse model of hfd to induce nash and investigate associated injury and inflammation in the lung. we found that nash was associated with histopathological alterations in lung tissue 6 months post hfd feeding; moreover, expression of genes related to inflammation and lipid metabolism was dysregulated throughout nafld development, in cluding the progression to nash. these results provide insights into the interplay between liver and lung inflammation and highlight potential inflammatory pathways for crosstalk between the tissues. based on established criteria, we confirmed the development of nash in mice fed a hfd as we observed increased steatosis and inflammatory cell infiltration at 6 months.[2,16] this was correlated with pulmonary histopathological changes in the hfdfed mice at this time. further assessment by a pathologist will be completed to confirm these histopathological changes. although there was no evidence of pulmonary fibrosis in these animals, we noted the appearance of lipid-laden macrophages in the lung. these cells have been shown to be asso ciated with fibrosis in other disease states, and may contribute to the development of lung fibrosis at later time points in nash.[18] interestingly, histopathological changes in the lung were not reflected by increases in bal protein or cell counts, which are markers of pulmonary alveolar epithelial damage and leaky vasculature, or infiltration of immune cells into the lung in response to injury.[17] it may be that nash is not associated with epithelial barrier dysfunction and that pathologic alterations involve other mechanisms of injury. for example, it is possible that resident macrophages present in the lung are activated following hfd and that they drive lung inflammation. further studies are needed to explore this possibility. we also noted a decrease in bal protein content at 6 months and increases in bal cells in control mice; this may be indicative of age-related changes in tissue structure or vasculature, or in basal inflammatory activity.[4] hfd-fed mice seem to also mimic this age-related trend of increasing inflammation, although not significantly. we speculated that histopathological changes in the lung of animals fed a hfd might be driven by differential expression of genes related to inflammatory proteins and cytokines. in this context, our gene expression analyses revealed increases in il-1b at 3 months in hfd-fed mice, which suggests that during the development of nash, the lung responds to hepatic inflammation by upregulating inflammatory gene expression. il-1β is an early response cytokine known to promote inflammation; thus, increases in inflammatory cells in the lung of mice fed a hfd may be mediated in part by this cytokine. we also observed early downregulation of apoa1 in hfd-fed mice at 1 month. apoa1 is a lipid chaperone that promotes the efflux of cholesterol; it has been shown to have anti-inflammatory and antifibrotic effects in the lung.[6,8] the observed decrease in apoa1 may further exacerbate inflammation in the lung in response to a hfd. although we did not observe lung fibrosis during the histopathological analysis, we assessed changes in ctgf gene expression as a measure of fibrotic extracellular matrix tissue remodeling.[10,13] ctgf expression was decreased in hfd-fed mice at 1 month, suggesting that a hfd may suppress fibrotic mechanisms in the lung early aresty rutgers undergraduate research journal, volume i, issue iii in the process of nash development. this might be a compensatory response to prevent fibrosis induced by other growth factors generated in the lung in response to the hfd. a similar decrease in ctgf expression at 3 and 6 months in the control mice indicates that the hfd may be mimicking age-related effects in mice as early as 1 month. while these data provide preliminary characterization of pulmonary changes in a mouse model of nash, experiments to confirm the presence of lipid-laden macrophages by staining for lipids as well as an assessment of histopathological changes by a pathologist will be conducted. future studies will also be performed to further elucidate mechanisms of high fat diet-associated effects on the lung, including assessment of systemic markers of lipid dysregulation, liver inflammation and dys function, and other inflammatory signaling pathways in the lung. developing our understanding of the interplay between the lung and the liver can help identify the mechanisms by which disease can influence distant pathologies, and how inflammation in particular can be controlled to limit pathological comorbidities in patients. 5 conclusion overall, these data demonstrate that hfdinduced nash leads to pulmonary histopathological changes. moreover, these changes may be driven by the dysregulation of key mediators involved in inflammation and lipid metabolism. this analysis of lung-liver crosstalk in nash highlights potential for the clinical management of pulmonary complications associated with nash∎ 6 acknowledgements i would like to acknowledge everyone in dr. debra laskin’s lab for their guidance and support, especially dr. alexa murray and dr. debra laskin. i would also like to acknowledge dr. bo kong for his expertise and guidance with the animals used in this project, dr. laura armstrong for providing liver images, and dr. grace guo for her continued scientific guidance. this research was supported by nih grants es029258, es005022, and es004738. 7 references [1] armstrong, m.j., adams, l.a., canbay, a., & syn, w.k. (2014). extrahepatic complications of nonalcoholic fatty liver disease. hepatology, 59(3), 1174–1197. [2] benedict, m., & zhang, x. (2017). non-alcoholic fatty liver disease: an expanded review. world journal of hepatology, 9(16), 715–732. [3] fabbrini, e., sullivan, s., & klein, s. (2010). obesity and nonalcoholic fatty liver disease: biochemical, metabolic and clinical implications. hepatology (baltimore, md.), 51(2), 679–689. [4] franceschi, c., garagnani, p., parini, p., giuliani, c., & santoro, a. (2018). inflammaging: a new immune-metabolic viewpoint for age-related diseases. nature reviews. endocrinology, 14(10), 576–590. [5] jackson, s.e., llewellyn, c.h., & smith, l. (2020). the obesity epidemic – nature via nurture: a narrative review of high-income countries. sage open medicine, 8, 2050312120918265. [6] kim, c., lee, j.m., park, s.w., kim, k.s., lee, m.w., paik, s., jang, a.s., kim, d.j., uh, s., kim, y., & park, c.s. (2016). attenuation of cigarette smoke-induced emphysema in mice by apolipoprotein a-1 overexpression. american journal of respiratory cell and molecular biology, 54(1), 91–102. [7] kwak, m.s., kim, e., jang, e.j., & lee, c.h. (2018). the association of non-alcoholic fatty liver disease with lung function: a survey design analysis using propensity score. respirology (carlton, vic.), 23(1), 82–88. [8] lee, e.h., lee, e., kim, h.j., jang, a.s., koh, e.s., uh, s., kim, y.h., park, s., & park, c. (2013). overexpression of apolipoprotein a1 in the lung abrogates fibrosis in experimental silicosis. plos one, 8(2), e55827. [9] lim, s., taskinen, m.r., & borén, j. (2019). crosstalk between nonalcoholic fatty liver disease and cardiometabolic syndrome. obesity reviews, 20(4), 599–611. [10] lipson, k.e., wong, c., teng, y., & spong, s. (2012). ctgf is a central mediator of tissue remodeling and fibrosis and its inhibition can reverse the process of fibrosis. fibrogenesis & tissue repair, 5(1), s24 [11] mantovani, a., lonardo, a., vinco, g., zoppini, g., lippi, g., bonora, e., loomba, r., tilg, h., byrne, c.d., fabbri, l., & targher, g. (2019). association between non-alcoholic fatty liver disease and decreased lung function in adults: a systematic review and meta-analysis. diabetes & metabolism, 45(6), 536–544. [12] peng, t.c., kao, t.w., wu, l.w., chen, y.j., chang, y.w., wang, c.c., tsao, y.t., & chen, w.l. (2015). association between pulmonary function and nonalcoholic fatty liver disease in the nhanes iii study. medicine, 94(21). [13] ponticos, m., holmes, a.m., shi-wen, x., leoni, p., khan, k., rajkumar, v.s., hoyles, r.k., bou-gharios, g., black, c.m., denton, c.p., abraham, d.j., leask, a., & lindahl, g.e. (2009). pivotal role of connective tissue growth factor in lung fibrosis: mapk-dependent transcriptional activation of type i collagen. arthritis and rheumatism, 60(7), 2142–2155. aresty rutgers undergraduate research journal, volume i, issue iii [14] schuster, s., cabrera, d., arrese, m., & feldstein, a.e. (2018). triggering and resolution of inflammation in nash. nature reviews gastroenterology & hepatology, 15(6), 349–364. [15] song, j.u., jang, y., lim, s.y., ryu, s., song, w.j., byrne, c.d., & sung, k.c. (2019). decreased lung function is associated with risk of developing non-alcoholic fatty liver disease: a longitudinal cohort study. plos one, 14(1). [16] spengler, e.k., & loomba, r. (2015). recommendations for diagnosis, referral for liver biopsy, and treatment of nafld and nash. mayo clinic proceedings, 90(9), 1233–1246. [17] sunil, v.r., patel, k.j., shen, j., reimer, d., gow, a.j., laskin, j.d., & laskin, d.l. (2011). functional and inflammatory alterations in the lung following exposure of rats to nitrogen mustard. toxicology and applied pharmacology, 250(1), 10–18. [18] venosa, a., smith, l.c., murray, a., banota, t., gow, a.j., laskin, j.d., & laskin, d.l. (2019). regulation of macrophage foam cell formation during nitrogen mustard (nm)-induced pulmonary fibrosis by lung lipids. toxicological sciences, 172(2), 344–358. [19] vernon, g., baranova, a., & younossi, z.m. (2011). systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. alimentary pharmacology & therapeutics, 34(3), 274– 285. [20] viglino, d., plazanet, a., bailly, s., benmerad, m., jullian-desayes, i., tamisier, r., leroy, v., zarski, j.p., maignan, m., joyeux-faure, m., & pépin, j.l. (2018). impact of non-alcoholic fatty liver disease on long-term cardiovascular events and death in chronic obstructive pulmonary disease. scientific reports, 8. [21] wong, r.j., aguilar, m., cheung, r., perumpail, r.b., harrison, s.a., younossi, z.m., & ahmed, a. (2015). nonalcoholic steatohepatitis is the second leading etiology of liver disease among adults awaiting liver transplantation in the united states. gastroenterology, 148(3), 547–555. tanvi banota is a senior in the honors college at rutgers university majoring in cell biology and neuroscience and minoring in linguistics. she has been conducting research in the laskin lab since high school when she was first matched with the lab through the liberty science center partners in science program. her research is on the mechanisms of inflammation following toxicantinduced pulmonary injury. tanvi plans to pursue an md/phd after rutgers and hopes to have a career running her own lab, seeing patients, and mentoring students. outside of academics and research, tanvi is an epee fencer on the rutgers fencing team, is highly involved with the aresty research center, and enjoys watching sports, spending time with her friends, and playing and listening to music. tanvi was also a part of the founding team for the aresty rurj in 2019 and has served as a peer reviewer, senior peer reviewer, and editor for the journal. she’s excited to see where the journal will go next and how it will continue to uplift and represent the undergraduate research community at rutgers! for any questions, please contact tanvi at: tanvi.banota@rutgers.edu aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. vitamin d receptor binding with dna in duodenal crypt, duodenal villi, & colonic epithelial cells of mice dennis a. aldea, rohit aita, sohaib hassan, evan s. cohen, joseph hur, oscar pellón-cárdenas, lei chen, michael p. verzi (faculty advisor) ✵ abstract vitamin d receptor (vdr) is a transcription factor that mediates calcium absorption by intestinal epithelial cells. although calcium absorption is canonically thought to occur only in the small intestine, recent studies have shown that vdr activity in the colon alone is sufficient to prevent calcium deficiency in mice. here, we further investigate vdr activity in the colon. we assess vdr-dna binding in mouse duodenal crypt, duodenal villi, and colonic epithelial cells using chromatin immunoprecipitation sequencing (chip-seq). we find that most vdr-responsive elements are common to all intestinal epithelial cells, though some vdr-responsive elements are regionally-enriched and exhibit greater vdr-binding affinity in either duodenal epithelial cells or colonic epithelial cells. we also assess chromatin accessibility in the same three cell types using assay for transposase-accessible chromatin sequencing (atacseq). by integrating the vdr chip-seq and atacseq data, we find that regionally-enriched vdr-responsive elements exhibit greater chromatin acces sibility in the region of their enrichment. finally, we assess the transcription factor motifs present in vdrresponsive elements. we find that duodenumand colon-enriched vdr-responsive elements exhibit different sets of transcription factor motifs other than vdr, suggesting that vdr may act together with different partner transcription factors in the two regions. our work is the first investigation of vdr-dna binding in the colon and provides a basis for further investigations of vdr activity in the colon. 1 introduction intestinal anatomy the small and large intestines are a pair of digestive organs that form the lower half of the vertebrate gastrointestinal tract following the stomach. the small intestine is divided into the duodenum, the jejunum, and the ileum (figure 1). the large intestine follows the small intestine and is divided into the cecum, the colon, and the rectum. both the small and large intestines are lined by the intestinal epithelium—a layer of cells separating the contents of the intestines from the rest of the body. in the small intestine, the intestinal epithelium is a tightly folded structure consisting of villi—finger-like protrusions, and crypts—thin invaginations between villi.[6] the folded structure of the intestinal epithelium increases the surface area available for nutrient absorption in the small intestine. villi are composed of differentiated epithelial cells, and crypts are composed of undifferentiated intestinal stem cells. intestinal stem cells in the crypts differentiate and migrate to the villi, maintaining the intestinal epithelium even as epithelial cells are constantly shed from the villi.[5] villi are not present in the colon; instead, the colonic epithelium is a heterogeneous mixture of differentiated epithelial cells and undifferentiated stem cells.[6] vitamin d receptor vitamin d receptor (vdr) is a transcription factor present in intestinal epithelial cells. transcription factors are proteins that regulate the transcription of genes from dna to rna. transcription factors bind to particular genomic regions near their target genes, inducing conformational changes in the dna aresty rutgers undergraduate research journal, volume i, issue iii that either promote or inhibit transcription of the target genes. most transcription factors can only bind to genomic regions located in open chromatin— dna that is not condensed by histone proteins. a transcription factor’s responsive elements are the open genomic regions to which the transcription factor binds. a transcription factor’s binding specificity results from the presence of particular nucleotide sequences, called transcription factor motifs, in the transcription factor’s responsive elements. vdr binds calcitriol (1,25-dihydroxyvitamin d3), which is the activated form of vitamin d. vitamin d, whether synthesized by skin cells as vitamin d3 or obtained from the diet as vitamin d2, is inactive and must be metabolized in the kidneys to calcitriol before binding with vdr. from the kidneys, calcitriol enters the bloodstream and ultimately ligates with and activates vdr in intestinal epithelial cells. once activated by calcitriol, vdr induces the transcription of genes required for calcium absorption by the intestinal epithelial cells. vdr dysfunction is implicated in inflammatory bowel disease (ibd) and colorectal cancer (crc). certain atypical variants of the vdr gene are associated with increased prevalence of ibd.[8] ibd patients also exhibit reduced levels of vdr in the intestinal epithelium. these findings suggest a link between vdr dysfunction and ibd. clinical studies of crc patients have shown that different variants of vdr correlate with differences in patient survival.[21] metastatic crc tumor cells exhibit reduced vdr expression compared to typical intestinal cells.[16] reduced levels of vdr expression enhance the abnormal wnt/β-catenin pathway that drives the growth of most crc tumors.[11] although the precise role of vdr deficiency in crc tumor formation is still unknown, these findings suggest that vdr inhibition drives the metastasis of crc tumors. one of the best appreciated roles of vdr is its mediation of calcium absorption in the small intestine. knockout studies investigate the function of a protein by inactivating the gene encoding the protein and observing the effect. when vdr is knocked out in both the small and large intestines, the vdrknockout mice develop calcium deficiencies and ultimately rickets, a disease characterized by severe reduction in bone density.[20] however, when vdr is knocked out in the small intestine but not in the colon, the mice are still able to absorb enough calcium to avoid developing rickets.[7] though these studies show that vdr is active in the colon, the differences in vdr-dna binding between the duodenum and colon have not yet been investigated. research objectives here, we investigate vdr-dna binding in the duodenum and colon using laboratory mice (mus musculus) as model organisms. mice and humans both express vdr, and vdr-knockout mice exhibit a phenotype analogous to severe vitamin d deficiency in humans.[20] these findings suggest that the regulatory role of vdr is similar in mice and humans. we assess vdr-dna binding in mouse duodenal villi, duodenal crypt, and colonic epithelial cells using chromatin immunoprecipitation sequencing (chip-seq). by comparing vdr-dna bindfigure 1: the gastrointestinal tract [2] the small and large intestines are the two organs of the lower gastrointestinal tract. the small intestine follows the stomach and is divided into the duodenum, the jejunum, and the ileum. the large intestine follows the small intestine and is divided into the cecum (not shown), the colon, and the rectum (not shown). aresty rutgers undergraduate research journal, volume i, issue iii ing across cell types, we identify vdr-responsive elements that are regionally-enriched—exhibiting greater vdr-binding affinity in either the duodenum or in the colon. these regionally-enriched vdr-responsive elements cannot result from differences in vdr motif presence, since genomic sequences do not differ between cell types. we hypothesize that differences in vdr-dna binding result from differences in open chromatin regions between cell types. we examine this hypothesis by assessing chromatin accessibility in each cell type using assay for transposase-accessible chromatin sequencing (atac-seq). by integrating the vdr chip-seq and the atac-seq data, we find that regionally-enriched vdr-responsive elements exhibit greater chromatin accessibility in the region of their enrichment. thus, we conclude that differential chromatin accessibility causes differential vdr-dna binding between intestinal regions. finally, we assess the transcription factor binding site motifs present in vdr-responsive elements. we find that duodenumand colon-enriched vdr-responsive elements exhibit different sets of binding site motifs for transcription factors other than vdr, suggesting that vdr may act in conjunction with different partner transcription factors in the two regions. our work is the first survey of vdr-responsive elements in the colon. we find that tissue-enriched vdr-responsive elements in duodenal and colonic epithelium cells differ in chromatin accessibility and secondary transcription factor binding motifs. these findings provide a basis for further investigations of differences in vdr-mediated gene expression between the small and large intestines. understanding the action of vdr specific to the colon may explain the role of vdr in inflammatory bowel disease and colorectal cancer. 2 methodology experimental mice four wild-type mice (c57bl/6j strain) were used as experimental animals. all animal protocols were approved by the rutgers institutional animal care and use committee. the mice were given food and water ad libitum and exposed to a 12 h light, 12 h dark cycle. the mice were euthanized when they were 4– 6 weeks old. one hour prior to euthanasia, the mice were treated with 10 ng/g body mass 1,25-dihydroxyvitamin d3 (caymon chemical: #11792), injected intraperitoneally. sample preparation tissue samples were harvested from duodenal villi, duodenal crypts, and colonic epithelium of the mice immediately following euthanasia. duodenal villi, duodenal crypts, and colonic epithelial cells were extracted from the tissue samples and pelleted by centrifugation.[4] chip-seq protocol & analysis vdr-targeted chip-seq was conducted on the duodenal villi, duodenal crypts, and colonic epithelial cells using mouse monoclonal igg2a vdr antibody d-6 (santa cruz biotechnology: #sc-13133) and rabbit polyclonal igg vdr antibody c-20 (santa cruz biotechnology: #sc-1008). the vdr binding site library was purified and amplified using qiaquick pcr purification kit #50 (qiagen: #28104) before being sequenced. sequencing adapters were removed from the vdr chip-seq read fastq files using ngmerge.[9] each pair of forward and reverse adapter-trimmed read fastq files was aligned to mouse genome assembly mm9 using bowtie2.[10,15] each alignment sam file was converted to an alignment bam file using the samtools suite.[13] a composite vdr chip-seq alignment bam file was generated for each cell type by combining each set of replicate vdr chip-seq alignment bam files using the merge utility in the samtools suite. an alignment track bigwig file was generated from each composite and replicate alignment bam file using the bamcoverage utility in the deeptools suite.[18] vdr-binding peaks were identified from each vdr chip-seq alignment bam file using the callpeak utility in macs.[22] peaks overlapping encode mm9 blacklisted regions were removed from each peak set bed file using the subtract utility in the bedtools suite.[17] the encode blacklists list genomic regions known to yield false chip-seq signals aresty rutgers undergraduate research journal, volume i, issue iii due to the inaccuracies of a particular genome assembly.[1] each peak set bed file was then shifted so that each peak would be centered on its summit–the nucleotide with the greatest chip-seq signal within the peak region, as determined by macs. the summit of each peak represents the most probable transcription factor binding site; therefore, summit-centering ensures that each peak is centered around the binding site. for each cell type composite, the vdr chipseq signal was plotted versus distance from the nearest vdr chip-seq peak using the siteprobw program included in the cistrome suite.[14] atac-seq protocol & analysis atac-seq was conducted on the duodenal villi, duodenal crypts, and colonic epithelial cells using nextera tn5 transposase (illumina: #fc-121-1030). the transposed chromatin was purified and amplified using qiaquick pcr purification kit #50 (qiagen: #28104) before being sequenced. sequencing adapters were removed from the atac-seq read fastq files using ngmerge.[9] each pair of forward and reverse adapter-trimmed read fastq files was aligned to mouse genome assembly mm9 using bowtie2.[10,15] each alignment sam file was converted to an alignment bam file using the samtools suite.[13] a composite atac-seq alignment bam file was generated for each cell type by combining each set of replicate atac-seq alignment bam files using the merge utility in the samtools suite. an alignment track bigwig file was generated from each composite and replicate alignment bam file using the bamcoverage utility in the deeptools suite.[18] the median alignment size of each atacseq alignment bam file was determined using the collectinsertsizemetrics utility included in the picard suite.[3] each alignment bam file was converted to a bed file using the bamtobed utility included in the bedtools suite.[17] peak region bed files were called from each alignment bed file using macs.[22] it was necessary to convert the alignment bam files to bed files so that macs would properly interpret the nonoverlapping forward and reverse peaks typical of atac-seq but not of chip-seq. macs was run with a shift distance of negative one-half the median alignment size and an extsize distance equal to the median alignment size for each alignment bed file. for each cell type composite, the atac-seq signal was plotted versus distance from the nearest atac-seq peak using the siteprobw program included in the cistrome suite.[14] differential binding analysis peaks exhibiting differential vdr-binding affinities between cell types were identified using diffbind.[19] the following contrasts were examined: colonic epithelium versus duodenal crypts, colonic epithelium versus duodenal villi, and duodenal crypts versus duodenal villi. each set of differentially bound peaks was filtered to only include peaks assigned a significance value less than 0.001 by diffbind. each filtered peak set was exported to a bed file. for each contrast, the vdr chip-seq signals of each contrasted cell type were plotted versus the enriched vdr chip-seq peak sets for each contrasted cell type using the siteprobw program included in the cistrome suite.[14] motif analysis transcription factor motifs were identified from the composite vdr chip-seq peak set bed files of each cell type using homer. due to limited computational resources, the size of each composite peak set was reduced by random sampling. sample peak set bed files were generated by randomly selecting one-fifth of the peaks in each composite peak set bed file. homer was used to identify motifs enriched in each sample peak set bed file. transcription factor motifs were also identified from each differential vdr chip-seq peak set bed file using homer. however, the smaller size of the differential peak sets meant that sampling was not necessary; the entirety of each differential peak set bed file was analyzed using homer. 3 results genome alignment all of the vdr chip-seq read fastq files were successfully aligned to mouse genome assembly aresty rutgers undergraduate research journal, volume i, issue iii mm9.[15] the successful genomic alignments are indicated by the high alignment rates; all samples exhibited alignment rates greater than 85%, and all but one sample exhibited alignment rates greater than 90% (table 1). peak calling the accuracy of vdr chip-seq peak calling was assessed by comparing each cell type’s composite peak set bed file to its composite track bigwig file. the resulting signal plots indicate that all of the com posite peak sets exhibit a majority of reads near the centers of peak regions (figure 2). thus, we confirm that the vdr-responsive elements identified by macs exhibit elevated vdr binding, as expected of responsive elements. differential binding analysis the correlation between vdr chip-seq peak sets was assessed using diffbind.[19] the resulting correlation matrix indicates that the colonic epithelium peak sets are all more closely correlated to each table 1: vdr chip-seq genome alignment metrics [a] composite alignment bam files were constructed using samtools merge.[14] [b] genome alignment was conducted using mouse genome assembly mm9 and bowtie2.[10,15] [c] duplicate alignments were removed using macs filterdup.[22] aresty rutgers undergraduate research journal, volume i, issue iii figure 2: vdr chip-seq signal versus distance from vdr chip-seq peaks most vdr chip-seq reads are located near vdr chip-seq peaks. vdr chip-seq signal versus distance from nearest vdr chipseq peak. figure generated using siteprobw.[14] figure 3: vdr chip-seq differential binding heatmap the difference between colonic and duodenal vdr chipseq peak sets is greater than the difference between duodenal crypt and duodenal villi vdr chip-seq peak sets. darker shading in the heatmap represents greater similarity between peak sets; greater vertical separation in the dendrogram represents greater difference between peak sets. figure generated by diffbind.[19] figure 4: differential vdr chip-seq peak set sizes the majority of vdr chip-seq peaks are common to duodenal villi, duodenal crypt, and colonic epithelial cells. a small minority of vdr chip-seq peaks differ between duodenal epithelial and colonic epithelial cells. aresty rutgers undergraduate research journal, volume i, issue iii other than to any of the duodenal crypt or duodenal villi peak sets (figure 3). the correlation matrix also indicates that the duodenal crypt and duodenal villi peak sets do not differ significantly. the majority of vdr chip-seq peaks are common to all three cell types. out of the 23,381 vdr-binding sites exhibited in either colonic epithelial or duodenal villi cells, only 1,741 sites differ be figure 5: vdr chip-seq signal versus distance from differential vdr chip-seq peaks differential vdr chip-seq peaks exhibit greater vdr binding in the tissue of their enrichment. vdr chip-seq signal versus distance from nearest differential vdr chip-seq peak. figure generated by siteprobw.[14] figure 6: atac-seq signal versus distance from differential vdr chip-seq peaks differential vdr chip-seq peaks exhibit greater chromatin accessibility in the tissue of their enrichment. atac-seq signal versus distance from nearest differential vdr chip-seq peak. figure generated by siteprobw.[14] aresty rutgers undergraduate research journal, volume i, issue iii table 2: vdr chip-seq peak calling metrics [a] composite alignment bam files were constructed using samtools merge.[14] [b] peak calling was conducted using macs.[22] [c] peaks included in the encode mm9 blacklist were removed.[1] table 3: differential vdr chip-seq peak set metrics [a] enriched peaks were determined using diffbind. [19] [b] 𝑝𝑝 < 0.001 aresty rutgers undergraduate research journal, volume i, issue iii tween the two cell types (table 3). there is a comparable difference between colonic epithelial and duodenal crypt cells, which differ in only 1,102 vdrbinding sites out of 20,751 sites total. nonetheless, these small differences are far greater than the minuscule difference between duodenal crypt and duodenal villi cells, which differ by only 85 vdr-binding sites out of 20,479 sites total (figure 4). the accuracy of the differential binding analysis performed by diffbind was assessed by comparing each cell type’s enriched peak set bed file against all of the composite track bigwig files. the resulting signal plots indicate that all of the tissueenriched peak sets exhibit greater overlap with vdrresponsive elements in the tissue of their enrichment than with vdr-binding sites in other tissues (figure 5). thus, we confirm that the tissue-enriched vdr-responsive elements identified by diffbind exhibit elevated vdr binding in the tissue of their enrichment, as expected of tissue-enriched responsive elements. comparison of vdr chip-seq & atac-seq to investigate our hypothesis that regionally-enriched vdr binding results from differential chromatin accessibility between tissues, we compared duodenumand colon-enriched vdr-binding peaks bed files against all of the composite atac-seq bigwig files. the resulting signal plots indicate that all of the regionally-enriched vdr chip-seq peak sets exhibit greater overlap with open chromatin sites in the region of their enrichment than with open chromatin sites in the other region (figure 6). thus, we conclude that differences in vdr binding result from differences in chromatin accessibility between regions. motif finding due to limited computational resources, motif finding was conducted on samples generated by randomly selecting one-fifth of the peaks in each composite vdr chip-seq peak set (table 2). vdr is the most significant motif present in any of the composite vdr chip-seq peak set samples (figure 7), indicating that the vdr chip-seq was performed correctly. besides vdr motifs, the colonenriched peak set samples also exhibit hoxb13, figure 7: vdr motif presence in vdr chip-seq peak set samples vdr motif was the top transcription factor motif identified in all of the composite vdr chip-seq peak sets. motifs were identified using homer on a random sample of onefifth of each composite peak set figure 8: vdr and secondary motif presences in differential vdr chip-seq peak sets colonand duodenum-enriched vdr chip-seq peaks exhibit different secondary transcription factor motifs other than vdr. motifs were identified using homer on each differential peak set. aresty rutgers undergraduate research journal, volume i, issue iii cdx2, and foxa2 motifs, whereas the cryptand villi-enriched peak set samples exhibit hnf4α, erra, and gata4 (figure 8). these results suggest that differences in vdr binding between the duodenal epithelium and colonic epithelium may result from differences in the helper transcription factors that facilitate vdr-dna binding. 4 discussion we found that the vdr-binding profile of colonic epithelial cells is largely similar to that of duodenal villi and duodenal crypt cells. out of about twenty thousand vdr binding sites total, only a minority of several hundred binding sites differ between colonic epithelial cells and duodenal epithelial cells. nonetheless, this difference is greater than that between duodenal villi and duodenal crypt cells, which differ in less than one hundred binding sites. by comparing vdr chip-seq and atac-seq data, we found that colonand duodenum-enriched vdr-binding sites exhibited greater chromatin accessibility in the tissue of their enrichment. we also determined that colonand duodenum-enriched vdr-binding sites exhibit distinct sets of secondary (i.e. non-vdr) transcription factor motifs. colon-enriched vdr-binding sites exhibit hoxb13, cdx2, and foxa2 motifs; duodenum-enriched vdr-binding sites exhibit hnf4α, gata4, and erra motifs. these findings concur with those of a previous investigation which found that vdr-binding sites in the duodenum exhibited hnf4α and gata4 motifs in addition to vdr motifs.[12] tissue-specific secondary transcription factors may cause differential vdr binding, either by causing the differences in open chromatin observed at differential binding sites, or by directly binding to vdr and affecting vdr-dna binding. possible binding interactions between vdr and secondary transcription factors could be examined using protein immunoprecipitation assays. additional investigations can be conducted using intestinal organoid models. by knocking out secondary transcription factors in intestinal organoids, the role of these factors in vdr-mediated regulation of gene expression could be determined. such investigations would ad vance our understanding of vdr’s role in intestinal health and diseases, including colorectal cancers, and possibly offer new treatments for those affected by these conditions. 5 data & source code access the complete data collected in this investigation are available upon request. this investigation did not involve human subjects, and these data do not include hipaa-protected health information. the source code for the analysis pipeline used in this investigation is available upon request. the following programs were used in the analysis pipeline: ngmerge v0.3,[9] bowtie2 v2.2.6,[10] samtools v0.1.19,[13] deeptools v3.3.0,[18] macs v2.1.0,[22] bedtools v2.17.0,[17] cistrome v0.6.7,[14] picard v2.18.27,[3] diffbind v1.16.3,[19] homer v4.8.3∎ 6 references [1] amemiya, h. m., kundaje, a., & boyle, a. p. (2019). the encode blacklist: identification of problematic regions of the genome. scientific reports, 9(1), 9354. [2] blausen medical communications. (2014). blausen 0432 gastrointestinalsystem.png. https://commons.wikimedia.org/wiki/file:blausen_0432_gastrointestinalsystem.png [3] broad institute. picard. https://broadinstitute.github.io/picard [4] chen, l., toke, n. h., luo, s., vasoya, r. p., fullem, r. l., parthasarathy, a., perekatt, a. o., & verzi, m. p. (2019). a reinforcing hnf4-smad4 feed-forward module stabilizes enterocyte identity. nature genetics, 51, 777–785. [5] crosnier, c., stamataki, s., & lewis, j. (2006). organizing cell renewal in the intestine: stem cells, signals and combinatorial control. nature reviews genetics, 7(5), 349–359. [6] de santa barbara, p., van den brink, g. r., & roberts, d. j. (2003). development and differentiation of the intestinal epithelium. cellular and molecular life sciences, 60(7), 1322– 1332. [7] dhawan, p., veldurthy, v., yehia, g., hsaio, c., porta, a., kim, k., patel, n., lieben, l., verlinden, l., carmeliet, g., & christakos, s. (2017). transgenic expression of the vitamin d receptor restricted to the ileum, cecum, and colon of vitamin d receptor knockout mice rescues vitamin d receptor–dependent rickets. endocrinology, 158(11), 3792–3804. [8] eloranta, j. j., wenger, c., mwinyi, j., hiller, c., gubler, c., vavricka, s. r., fried, m., kullak-ublick, g. a., & swiss ibd cohort study group. (2011). association of a common vitamin d-binding protein polymorphism with inflammatory bowel disease. pharmacogenetics and genomics, 21(9), 559–564. [9] gaspar, j. m. (2018). ngmerge: merging paired-end reads via novel empirically-derived models of sequencing errors. bmc bioinformatics, 19, 536. https://commons.wikimedia.org/wiki/file:blausen_0432_gastrointestinalsystem.png https://commons.wikimedia.org/wiki/file:blausen_0432_gastrointestinalsystem.png https://broadinstitute.github.io/picard aresty rutgers undergraduate research journal, volume i, issue iii [10] langmead, b., & salzberg, s. l. (2012). fast gapped-read alignment with bowtie 2. nature methods, 9(4), 357–359. [11] larriba, m. j., ordóñez-morán, p., chicote, i., martín-fernández, g., puig, i., muñoz, a., & pálmer, h. g. (2011). vitamin d receptor deficiency enhances wnt/β-catenin signaling and tumor burden in colon cancer. plos one, 6(8), e23524. [12] lee, s. m., riley, e. m., meyer, m. b., benkusky, n. a., plum, l. a., deluca, h. f., & pike, j. w. (2015). 1,25-dihydroxyvitamin d3 controls a cohort of vitamin d receptor target genes in the proximal intestine that is enriched for calcium-regulating components. journal of biological chemistry, 290(29), 18199–18215. [13] li, h., handsaker, b., wysoker, a., fennell, t., ruan, j., homer, n., marth, g., abecasis, g., durbin, r., & 1000 genome project data processing subgroup. (2009). the sequence alignment/map format and samtools. bioinformatics, 25(16), 2078–2079. [14] liu, t., ortiz, j. a., taing, l., meyer, c. a., lee, b., zhang, y., shin, h., wong, s. s., ma, j., lei, y., pape, u. j., poldinger, m., chen, y., yeung, k., brown, m., turpaz, y., & liu, x. s. (2011). cistrome: an integrative platform for transcriptional regulation studies. genome biology, 12(8), r83. [15] mouse genome sequencing consortium. (2002). initial sequencing and comparative analysis of the mouse genome. nature, 420(6915), 520–562. [16] pálmer, h. g., larriba, m. j., garcía, j. m., ordóñez-morán, p., peña, c., peiró, s., puig, i., rodríguez, r., de la fuente, r., bernad, a., pollán, m., bonilla, f., gamallo, c., de herreros, a. g., & muñoz, a. (2004). the transcription factor snail represses vitamin d receptor expression and responsiveness in human colon cancer. nature medicine, 10(9), 917–919. [17] quinlan, a. r., & hall, i. m. (2010). bedtools: a flexible suite of utilities for comparing genomic features. bioinformatics, 26(6), 841–842. [18] ramírez, f., ryan, d. p., grüning, b., bhardwaj, v., kilpert, f., richter, a. s., heyne, s., dündar, f., & manke, t. (2014). deeptools2: a next generation web server for deep-sequencing data analysis. nucleic acids research, 44(w1), w160–w165. [19] stark, r., & brown, d. (2011). diffbind: differential binding analysis of chip-seq peak data. https://bioconductor.org/packages/release/bioc/vignettes/diffbind/inst/doc/diffbind.pdf [20] suda, t., masuyama, r., bouillon, r., & carmeliet, g. (2015). physiological functions of vitamin d: what we have learned from global and conditional vdr knockout mouse studies. current opinion in pharmacology, 22, 87–99. [21] vaughan-shaw, p. g., o’sullivan, f., farrington, s. m., theodoratou, e., campbell, h., duntop, m. g., & zgaga, l. (2017). the impact of vitamin d pathway genetic variation and circulating 25-hydroxyvitamin d on cancer outcome: systematic review and meta-analysis. british journal of cancer, 116(8), 1092–1110. [22] zhang, y., liu, t., meyer, c. a., eeckhoute, j., johnson, d. s., bernstein, b. e., nusbaum, c., myers, r. m., brown, m., li, w., & liu, x. s. (2008). model-based analysis of chip-seq (macs). genome biology, 9(9), r137. dennis aldea is a senior undergraduate student in the genetics program at rutgers new brunswick. his interest in scientific research was sparked by his experience in the rutgers waksman student scholars program, a biotechnology outreach program for high school students. dennis currently works with dr. michael verzi at the human genetics institute of new jersey, where he conducts bioinformatics analyses to investigate the genetic basis of intestinal diseases and cancers. https://bioconductor.org/packages/release/bioc/vignettes/diffbind/inst/doc/diffbind.pdf https://bioconductor.org/packages/release/bioc/vignettes/diffbind/inst/doc/diffbind.pdf aresty rutgers undergraduate research journal, volume i, issue iii this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. characterizing an alternatively spliced variant of chemokine receptor 2 in painful diabetic neuropathy justine soltys lei yu (faculty advisor) ✵ abstract prior research efforts have demonstrated a link between neuroinflammation and the progression of painful diabetic neuropathy (pdn), a chronic cascade of nerve damage that presents as tingling, numbness, hypersensitivity to touch, or intense pain. current treatments are focused on pain management, serving to temporarily mask these symptoms without repressing or slowing nerve damage. the chemokine-receptor system has been closely studied for its role in perpetuating neuropathic pain, although its precise mechanistic involvement remains unclear due to the network’s complexity. because of its likely role in regulating neuropathic pain, targeting ccr2 may be the key to effective treatment of pdn. alternative splicing of ccr2 leads to two distinct isoforms with different c-terminus sequences, ccr2a and ccr2b. the present study was intended to differentiate between these isoforms through specific primer design, selection of optimized pairs, rtpcr, and amplicon sequencing to verify the pcr products. however, the study has revealed a third, previously unreported isoform, ccr2c, due to evidence of alternative splicing and both the absence and insertion of parts of a and b. in the long term, we predict that the relationship between ccr2’s al ternatively spliced transcript variants will lead to a distinct pattern of isoform prevalence in individuals suffering from pdn. discerning the genetic profiles of patients with pdn and healthy individuals will clarify the complex mechanism driving ccr2’s intracellular interactions and offer more effective therapeutic options. 1 introduction painful diabetic neuropathy painful diabetic neuropathy (pdn) gives rise to several types of peripheral nerve damage but most frequently affects the feet in a pattern of distalto-proximal severity. during disease progression, the extremities are more severely impacted than the midline of the body, which models the branching peripheral nervous system. those neurons comprised of the longest axons exhibit increased exposure due to their size and distribution, thus sustaining the brunt of the damage.[11] as the sensory nerves are affected, individuals may experience numbness tingling, or a loss of reactivity to sensations like temperature, pain, or touch. damage to the motor nerves causes hyperalgesia (enhanced pain sensitivity) and allodynia (hypersensitivity to neutral stimuli often felt as a stabbing pain).[20] the incidence of pdn within diabetes ranges between 10–26%, reflecting differences among sample populations.[2] in individuals with pdn, a number of related metabolic and vascular factors are implicated in the onset of nerve damage. nerve damage stimulates the infiltration of macrophages, leading to an increase in the number of pro-inflammatory cytokines circling throughout the body. clinical studies involving diabetic neuropathic (dn) subjects with and without pain demonstrate that groups experiencing pdn have higher inflammation markers and increased cytokine concentration. the neuroinflammatory and immune responses are thus heightened and contribute to the development of neuropathic pain.[1] several factors are believed to contribute to the pain mechanism of pdn; as a result, current treatment options are decidedly limited. glycemic aresty rutgers undergraduate research journal, volume i, issue iii control, anticonvulsants, antidepressants, and opiates are meant to assuage pain and other symptoms related to pdn, but fail to restore nerve function.[13] tightened glycemic control lowers the blood sugar levels in an effort to reduce the risk of hyperglycemia, a causative factor in neuropathy. while this practice has displayed some success in slowing the progression of pdn, it does not alleviate nerve pain, and other drugs vary in the consistency of their results.[6] individuals suffering from pdn continue to live with nerve pain, loss of sensation, and the risk of limb amputation. because of this, there is still a pressing need to understand how pdn manifests and to develop effective therapeutic options that treat pdn directly. the role of ccr2 in pdn the chemokine-receptor network has been widely implicated in both inflammatory pain conditions and neuropathic pain. chemokines are a subset of cytokines responsible for attracting white blood cells to sites of inflammation through induced chemotaxis. c-c motif chemokine ligand 2 (ccl2) binds to c-c motif chemokine receptor type 2 (ccr2), with a high level of affinity. their interaction at the monocyte membrane is represented locally by a ccr2/ccl2 axis, creating a gradient with increasing monocyte concentration in the direction of inflammatory sites. as a result of its influence on the migration of white blood cells, ccr2/ccl2 signaling guides a number of key protective and destructive responses of the immune system (hughes et al., 2018). ccr2 is known to have two alternatively spliced transcript variants, ccr2a and ccr2b, which differ in the lengths of their cytoplasmic tails at the c-terminus. ccr2a consists of three exons, while ccr2b has an additional exon which is unique to that isoform. alternative splicing describes the process wherein a single gene is assembled through different regulatory mechanisms to encode for multiple proteins. the precursor to mrna, pre-mrna, may contain exons that are omitted, shortened, or have higher sequence conservation as compared to other examples of mature mrna. conventional chemokine receptors, a family of chemokine-binding surface molecules, are often subject to alternative splicing. like variants of ccr2, chemokine receptors ccr9, cxcr3, and cxcr4 have been specifically associated with changes in ligand-binding or signaling properties, which supports the idea that isoforms of ccr2 maintain altered properties.[5,18] gprotein coupled receptors, such as ccr2, are dependent on the amino acid sequence of their c-terminal domain for intracellular signaling. therefore, it is likely that functional differences between ccr2a and ccr2b will lead to their distinct signaling roles, despite the presence of a shared extracellular protein motif.[18] preclinical studies have demonstrated the involvement of ccr2 and ccl2 in the pathophysiology of neuropathic pain, with these models suggesting that ccr2-antagonists can reduce painful symptoms.[1] a 2013 astrazeneca study sought to evaluate the efficacy and safety of the drug azd2423 as a novel ccr2-antagonist in subjects with pdn.[9] ccl2 plasma levels increased in a dosage-dependent manner, and mean monocyte levels declined by approximately 27%, indicating that azd2423 interacted with its intended target, ccr2. however, astrazeneca’s study failed to show that azd2423 alleviated primary and most secondary pain variables with greater efficacy than the placebo. we suggest that azd2423 was found to be an ineffectual clinical treatment because it was not designed to cater to the distinct structural components of ccr2. to improve the efficacy of medications like these, our research has focused on understanding the levels of ccr2 variants in populations that suffer from pdn. stratifying drug treatment based on the expression ratio of individuals’ ccr2 isoforms may enhance the efficacy of therapeutics in mitigating pain. from this data, we aim to decode the role of alternatively spliced ccr2 variants in the incidence of pdn and pathophysiology of neuropathic pain. we initially performed amplicon sequencing of pcr products to analyze the genetic variation in a specific region and confirm their genomic contents. unexpectedly, we found evidence of additional alternative splicing sites, which implies the existence of a third isoform, aresty rutgers undergraduate research journal, volume i, issue iii ccr2c. the present study identifies ccr2 and characterizes its presumed functional relationship with ccl2 following rt-pcr, gel electrophoresis, and analysis of amplicon sequencing results. 2 methodology blood samples were obtained from 17 donors; 11 were from individuals with diabetes and/or pdn, and 6 from individuals who were healthy. each subject provided three 10 ml collection tubes of blood. we assigned each subject a letter from a through q and marked the blood samples with their corresponding letter. because monocytes are a direct contributor to the neuroinflammatory model of dn, we were primarily interested in the buffy coat and plasma, as these components have the highest concentrations of white blood cells. we treated the collection tubes of buffy coat with rnalater, a stabilizing reagent that inactivates rnases and maintains the integrity of the rna. we extracted rna from the whole blood samples on the date of their arrival in an effort to re duce degradation. first, we distributed the 5 ml of blood that was initially set aside into 200 μl test tubes, using the standard protocol from the neb monarch total rna mini-kit. the whole blood samples were lysed upon mixing with a 2x monarch dna/rna protection reagent concentrate. the whole blood rna was then distributed into aliquots and stored at -80° c until further processing. heparinase treatment because of the scarcity of the buffy coat layer and its viscosity, we experienced challenges extracting quantifiable levels of rna, thus requiring that cdna production in reverse transcription (rt) be maximally optimized. to improve cdna yield, heparinase i, a polysaccharide lyase, was added to the rna before rt to cleave heparin, an anticoagulant often used in blood sample processing. heparin tends to be used to prevent the clotting and clumping of cells, but it also inhibits the activity of the reverse transcriptase enzyme. test tubes were prepared containing rna, heparinase i, buffer, mgcl2, figure 1: a comparison of ccr2’s previously recorded isoforms, which depicts that ccr2a lacks an additional exon. exons 1 and 2 represent a conserved region between the variants, 67 base pairs (bp) and 992 bp long, respectively. the last exon is 2 bp longer in ccr2b than in ccr2a. primers listed within parentheses are situated at junctional sites, thus spanning two exons. aresty rutgers undergraduate research journal, volume i, issue iii and rnase inhibitor (new england biolabs reagents). some of these reagents were added in halfvolumes or less, with the remainder pipetted into the test tube after it was placed in a ptc-100 machine at 25°c for a two hour-long incubation period. rt-pcr and gel electrophoresis in designing isoform-specific primer pairs, we first developed primers which could distinguish ccr2a from ccr2b by binding to the regions unique to each variant. a two-letter nomenclature was developed, with the first letter referencing the forward primer and the second corresponding to the reverse. because they are complementary to sequences on ccr2b-exon 3, reverse primers f, g, and h, in combination with a forward primer, can only successfully bind to and amplify ccr2b during pcr. however, conserved primer pairs able to bind to both ccr2a and ccr2b might still selectively amplify only a single variant, depending on the resulting size of the pcr product. any product greater than 1,000 base pairs (bp) long is generally longer than the polymerization rate of the taq polymerase enzyme can accommodate under standard pcr extension times. in each cycle of pcr, 1 minute generally allows for up to 1 kb of the amplicon to be amplified, thus eliminating larger products. for example, primer pair in was previously confirmed by the author as a reliable ccr2a primer combination, yielding a pcr product 267 bp long. in can also bind to ccr2b, but the presence of the third exon creates an unlikely pcr product that stretches 1,464 bp; there is no current experimental data to support its specificity to ccr2b. ch and ih have both worked equally well as ccr2b-specific primer sets, likely because they bind to locations sitting relatively near each other. ch and ih isolate 309 and 222 bp pcr products in ccr2b, respectively. both primer combinations pq and vr have been validated as ccr2 internal controls because they bind to locations within the conserved region shared by each of the transcript variants, exons 1 and 2, producing identical amplicons in each. following heparinase treatment, buffer, mgcl2, dntps, random hexamers, oligo dt, and reverse transcriptase were incorporated into the rna solute (supplemental table 1). the precise volumes of the reagents for pcr (supplemental table 2) tended to be more flexible because 1 to 3 primer sets can be tested against a patient sample, which slightly alters the volume of water added to bring the master mix to 20 μl. the annealing step of pcr was consistently run at 54°c for 30 seconds. to make the agar component of a gel, we added 0.35 g of agarose and 25 ml of 1x tbe to a 50 ml flask and heated repeatedly until the solution was completely clear, with no pellets of agarose visible. the solution was poured into the tray of the gel box, solidified, and run at approximately 110 volts until the bands travelled two-thirds of the way to the opposite end of the gel. after being removed from the basin, the gel was stained with ethidium bromide for 20–30 minutes and photographed. all primers were designed using amplifx and oligo 7 software and ordered through integrated dna technologies. we carried out rt-pcr and gel electrophoresis for 12 primers pairs which had previously produced bands for ccr2a and ccr2b using reference rna: cd, cj, cl, cm, cn, co, id, ij, il, im, in, and io. amplicon sequencing was performed by genewiz, using next generation sequencing technology. 3 results during our initial testing, we performed rtpcr and gel electrophoresis in order to establish reliable sets of primers for further quantification of ccr2 variants in patient samples. rt generated complementary dna (cdna), a copy of mrna used as a template for pcr. in the following step, ccr2 variant-specific primer pairs amplified ccr2a and ccr2b, creating millions of copies. these pcr products were evaluated using gel electrophoresis to confirm that their lengths matched the determined amplicon size of each primer combination. while gel electrophoresis is a reliable visual tool, amplicon sequencing was needed to positively identify each dna sequence as a product of either ccr2a or ccr2b. each lane of figure 2 represents the contents of a ccr2a-specific primer combination. clear bands can be seen at the expected amplicon lengths for all primer combinations except cf, whose attributed pcr products were eliminated from the in aresty rutgers undergraduate research journal, volume i, issue iii ventory stored for amplicon sequencing. primer dimers appeared in all forward primer c-containing gel lanes below the 100 bp mark, often seen in ethidium bromide-stained gels in the 30-50 bp range, resulting from primers attaching to each other. we also ran rt-pcr for other combinations that could be attributed to ccr2b (gel not pictured). there were a few instances of multiple bands in the case of cd, im, in, and cj, which proved unusual due to the 1-minute cycle extension time during pcr, but not impossible for ccr2b-specific pcr products of approximately 1,000 bp. in theory, all four sets could bind to both variants, ccr2a and ccr2b. for the samples that presented the expected base pair lengths, our next step was to con firm these results with amplicon sequencing. following high-throughput marker gene analysis, the amplicon sequences were read and totaled, representing a match to a predicted primer pair–generated pcr sequence. a major discrepancy was seen in the produced sequence alignment for primer set cd, which illustrated a missing piece of variant ccr2a nearly 300 base pairs long. amplicon sequencing of cd yielded four significant sequences which we referred to by their lengths in base pairs. 190, 356, and 208 were a match to other chromosomes, while 115 corresponded to an ~180 gel band after accounting for the size of the forward and reverse primers. however, 115’s alignment to ccr2a was clearly dis figure 2: gel depicting cd, cj, cf, cm, cn, co, id, ij, il, im, in, and io primer bands with plasma-extracted rna as the template. each band corresponds to its “expected length in ccr2a” (table 1), though no band appears for cf. primer pair cf may have been unsuccessful at binding to the cdna if its melting temperature was outside of the range of the other sets, or within a region of secondary structure. table 1: expected primer pair lengths relative to isoform ccr2a, along with their approximate experimental band lengths determined from the published isoform’s dna sequence. aresty rutgers undergraduate research journal, volume i, issue iii rupted, though there was no immediate evidence of alternate splicing against ccr2b (figure 3). this finding is highly irregular, considering that the primer pair must have successfully bound to the cdna for a pcr product of any length to have been amplified. primers cn, in, cd, cj cm, co, id ij, il, and im produced segments that matched their expected ccr2a lengths. however, a few amplified fragments attributed to cn, in, and id were only partial matches to the entire expected sequence. for each primer combination, we analyzed the alignment of the sequences as compared the expected amplification on either ccr2a, ccr2b, or both and generated dot matrix plots (not pictured) to better visualize the accuracy of the predicted to the actual ampli fication. at this stage, we intended to identify the major sequences amplified, determine the primer sequences which might need to be trimmed or excluded from future analysis, and differentiate between those that were exclusive to either ccr2a or ccr2b. of in’s two significant sequences, in 78 corresponded to various human and primate sequences other than ccr2, but in 364 was an imperfect match to ccr2b, displaying possible evidence of alternative splicing due to the presence of multiple gt-ag splicing sites (figure 4). in 364 exhibited other unexpected alignments, including a portion of ccr2b’s third unique exon situated within a region of ccr2a, as shown in figure 5. figure 3: the spaces highlighted in blue indicate expected base pairs absent from amplified cd 115, in relation to ccr2a. this missing sequence does not display the 5’-gt ag-3’ sequence characteristic of intron removal, as a result of alternative splicing. aresty rutgers undergraduate research journal, volume i, issue iii figure 4: this alignment of in 364 to ccr2b appears to be the result of alternate splicing, as evidenced by the yellow gtag sequences. figure 5: the base pairs highlighted in blue indicate a portion of the amplified sequence in 364 that was not part of the predicted ccr2a sequence. aresty rutgers undergraduate research journal, volume i, issue iii figure 6 identifies both an absent sequence from ccr2a and the addition of part of ccr2b, producing a drastically changed sequence, shortened from the presence of additional splicing sites and yet incorporating parts of ccr2b-exon 3. 4 discussion collectively, the sequence alignments comparing cd 115 and in 364’s amplified products to our predicted sequences provide strong evidence of a third variant of the ccr2 gene. from our preliminary results, we observed unexpected alternative splicing and absent portions of the targeted isoform, revealing the existence of ccr2c. due to the distinct inclusion of exon 3 in ccr2b and parts of the same exon in ccr2c, the amino acid sequences of their c-terminus signaling region will be changed, likely causing functional differences in intracellular signaling to arise. ccr2c may have significance in the quantification of isoform expression levels between normal and diabetic patients. we propose that individuals’ expression profiles will translate to ccr2-antagonist responsiveness for azd2423 and contribute to a genetic stratification approach for other pdn treatments. ccr2’s coding frame presents a long peptide with the signature motif of an integral membrane protein and a transmembrane domain. alternative splicing shifts the translational frame, causing each of the three variants to represent a different reading frame. despite lacking a fourth exon, ccr2a’s open reading frame is longer than ccr2b’s; thus, the variant maintains a longer carboxy terminal tail. ccr2c’s cytoplasmic tail is greatly truncated, suggesting that the open reading frame ends a few amino acids into the c terminal (figure 7). figure 6: in this region of the in 364 sequence, the gene behaved primarily like ccr2a, though the blue highlights base pairs which should be present, and the yellow highlights a portion of ccr2b unexpectedly included in the alignment. at the left is the proposed third isoform ccr2c. this alignment suggests that the only difference between ccr2c and ccr2b is the contiguous nature of exon 3 in ccr2b, and the removal of two introns from exon 3 in ccr2c. aresty rutgers undergraduate research journal, volume i, issue iii the shortness of its carboxy tail could result in an effectively nonfunctional ccr2c isoform, the equivalent of a nonsense or stop mutation. ccr2c has a transmembrane and extracellular domain which appear identical to ccr2a and ccr2b, so ccr2c may become a sink, soaking up peptides or ligands that activate normal receptors. ccr2c would be similarly able to bind to ccl2, but this interaction would not trigger an active, intracellular response, reducing signal transduction. assuming that ccr2c has the same 5’ sequence as ccr2a and ccr2b, either pq or vr should, in theory, be able to target ccr2 equally well. as a result, the amplification of sequences using primers pq and rv represents the total level of ccr2 expression, though there may be other undiscovered isoforms of the ccr2 gene, which do not possess complete exon 1 similarity. the next step with regards to primer design would be to synthesize primer sets that can reliably differentiate between ccr2b and ccr2c so that each isoform can be quantified separately. the pcr product of ccr2c should be shorter than ccr2b because of the removal of two introns, allowing us to discern between the variants. two bands may be amplified, but given a distinctive difference in their lengths, gel electrophoresis bands should be clearly attributed back to ccr2b or ccr2c. we intend to determine ccr2 expression levels in order to understand the role of its isoforms in perpetuating nerve damage, pain, and/or loss of peripheral sensation. although we began this study to identify ratio differences in ccr2a/b in patient populations, amplicon sequencing results revealed distinct differences in sequences recovered from pcr, pointing to a third variant of ccr2. alternative splicing of premrna thus establishes three isoforms: a, b, and c. as a target of interest in prominent inflammatory and neuropathic pain conditions, antagonists of ccr2, such as azd2423 have been clinically tested with some level of success.[9] however, ccr2 variant expression may offer insight into drug responsiveness, serving as a biological marker of pdn treatment efficacy∎ 5 acknowledgements i would like to thank dr. lei yu for his tireless support and guidance. it is because of his far-reaching intellectual curiosity that i have had the privilege of working on this incredibly fulfilling project and have learned so much. i would also like to extend my thanks to fellow neuroinflammation lab members anika patel and madhuri achanta, who not only contributed data, but also carried the enthusiasm to follow scientific leads like true detectives. figure 7: contrasts the extracellular protein sequences of ccr2’s variants. the size of the protein’s cytoplasmic tail is a function of the reading frame’s positioning, which determines the amino acids encoded by the gene. aresty rutgers undergraduate research journal, volume i, issue iii 6 references [1] abbadie, c., et al. (2003). impaired neuropathic pain responses in mice lacking the chemokine receptor ccr2. proceedings of the national academy of sciences of the united states of america, 100(13), 7947–7952. [2] abbott ca, malik ra, van ross er, kulkarni j, boulton aj. prevalence and characteristics of painful diabetic neuropathy in a large community-based diabetic population in the u.k. diabetes care. 2011;34:2220–2224. [3] biolabs, new england. “guidelines for rna purification from whole blood.” neb www.neb.com/tools-and-resources/usage-guidelines/guidelinesfor-rna-purification-from-whole-blood [4] callaghan, b., cheng, h., stables, c., smith, a., & feldman, e. (2011). diabetic neuropathy: clinical manifestations and current treatments. lancet neurology, 11(6), 521–534. [5] charo, i. f., myers, s. j., herman, a., franci, c., connolly, a. j., & coughlin, s. r. (1994). molecular cloning and functional expression of two monocyte chemoattractant protein 1 receptors reveals alternative splicing of the carboxyl-terminal tails. proceedings of the national academy of sciences, 91(7), 2752–2756. [6] datta, i., bhadri, n., shahani, p., majumdar, d., sowmithra, s., razdan, r., & bhonde, r.(2017). functional recovery upon human dental pulp stem cell transplantation in a diabetic neuropathy rat model. cytotherapy, 19(10), 1208–1224. [7] hughes, c. e., & nibbs, r. j. (2018). a guide to chemokines and their receptors. the febs journal, 285(16), 2944–2971. [8] johnson, m. l., et al. (2003). heparinase treatment of rna before quantitative real-time rt-pcr. biotechniques, 35(6), 1140–1144. [9] kalliomäki, j., jonzon, b., huizar, k., o’malley, m., andersson, a., & simpson, d. (2012). evaluation of a novel chemokine receptor 2 (ccr2)-antagonist in painful diabetic polyneuropathy. scandinavian journal of pain, 4(2), 77–83. [10] kaur, s., et al. (2011). painful diabetic neuropathy: an update. annals of neurosciences, 18(4), 168-75. [11] morales-vidal s, morgan c, mccoyd m, hornik a. diabetic peripheral neuropathy and the management of diabetic pe ripheral neuropathic pain. postgrad med. 2012;124:145–153. [12] nanodrop technologies. 260/280 and 260/230 ratios technical support bulletin. 260/280 and 260/230 ratios technical support bulletin, nanodrop technologies, inc, 2007. https://www.bio.davidson.edu/projects/gcat/protocols/nanodrop_tip.pdf [13] nathan, d. m., and dcct/edic research group. (2014). the diabetes control and complications trial/epidemiology of diabetes interventions and complications study at 30 years: overview. diabetes care 37(1), 9-16. [14] sloan, g., shillo, p., selvarajah, d., wu, j., wilkinson, i. d., tracey, i., … tesfaye, s. (2018). a new look at painful diabetic neuropathy. diabetes research and clinical practice, 144, 177–191. [15] totsch, s. k., and sorge, r. e. (2017). immune system involvement in specific pain conditions. molecular pain, 13. [16] tsai, richard. “dna purification using buffy coat.” inside biobanking, 8 jan. 2015 www.thermofisher.com/blog/biobanking/dna-purification-usingbuffy-coat/ [17] van acker, k., et al. (2009). prevalence and impact on quality of life of peripheral neuropathy with or without neuropathic pain in type 1 and type 2 diabetic patients attending hos pital outpatients clinics. diabetes & metabolism, 35(3), 206–213. [18] wong, l.-m., myers, s. j., tsou, c.-l., gosling, j., arai, h., & charo, i. f. (1997). organization and differential expression of the human monocyte chemoattractant protein 1 recep tor gene. journal of biological chemistry, 272(2), 1038–1045. [19] yoon, j. w., and jun, h. s. (2005). autoimmune destruction of pancreatic beta cells. american journal of therapeutics, 12(6), 580–591. [20] zhu, t., meng, q., ji, j., zhang, l., & lou, x. (2017). tlr4 and caveolin‐1 in monocytes are associated with inflammatory conditions in diabetic neuropathy. clinical and translational science, 10(3), 178–184. justine soltys is a senior in the school of arts and sciences, majoring in molecular biology and biochemistry, with a minor in history. her research sits within the intersection between inflammation and neuropathic pain, under the guidance of dr. lei yu at the center for alcohol studies. at the yu lab, she studies the isoform expression of ccr2’s alternatively spliced variants as a potential biomarker and treatment target in painful diabetic neuropathy. in the summer of 2021, she interned with the discovery toxicologynon-clinical research and development division of bristol myerssquibb to investigate calcium dysregulation of ryanodine receptor 2 (ryr2) in catecholaminergic polymorphic ventricular tachycardia (cpvt). the common thread underlying both projects is pharmacogenetics, unique drug response influenced by genetic variations. after her undergraduate years, she hopes to obtain a phd and contribute to pharmacology discovery and development. for questions, please contact her at: justine.soltys@rutgers.edu http://www.neb.com/tools-and-resources/usage-guidelines/guidelines-for-rna-purification-from-whole-blood http://www.neb.com/tools-and-resources/usage-guidelines/guidelines-for-rna-purification-from-whole-blood https://www.bio.davidson.edu/projects/gcat/protocols/nanodrop_tip.pdf https://www.bio.davidson.edu/projects/gcat/protocols/nanodrop_tip.pdf http://www.thermofisher.com/blog/biobanking/dna-purification-using-buffy-coat/ http://www.thermofisher.com/blog/biobanking/dna-purification-using-buffy-coat/ aresty rutgers undergraduate research journal, volume i, issue iii 7 supplemental tables supplemental table 1: rt reagent table, establishing the standard reagent volumes per reaction. supplemental table 2: pcr reagent table, illustrating the variability of reagent ratios. *can change relative to the volume of cdna and number of primer sets per reaction tube. aresty rutgers undergraduate research journal, volume i, issue iii supplemental table 3: ccr2-specific primers, including their assigned letter, exon location, and forward or reverse nature. supplemental table 4: amplicon sequence variant (asv) table of major sequences, with the count noting how many times the sequence appeared within the pcr product sample for each primer pair. the number of variants for each primer pair may indicate significant variability, or single polynucleotide polymorphisms (snps). for example, the second and third cn variants are listed separately despite being the same length and having near-identical sequences because of a point mutation at the 27th base pair, where c has been changed to a t. similarly, the second variant of in is a couple base pairs different from the first listed sequence. aresty rutgers undergraduate research journal, volume i, issue iii [a] in her article, she listed the two modes of accommodation in the opposite order. i reversed them to better serve the structure of this paper. this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. “reminisces of a dead world”: neoclassical impulses in stockhausen’s gesang der jünglinge andrew faulkenberry ✵ abstract in the years following world war ii, integral serialist composers declared their intent to defy all previous musical conventions and eradicate all “reminisces of a dead world” from their music. karlheinz stockhausen was no exception, asserting his desire “to avoid everything which is familiar, generally known or reminiscent of music already composed.” however, stockhausen’s gesang der jünglinge, despite its reputation for technical innovation, bears a strong connection to prior musical traditions. in this regard, stockhausen resembled the neoclassical school of composers that sought to accommodate antiquated musical materials within a modern context. to demonstrate these similarities, i apply to gesang a model of neoclassicism developed by martha m. hyde, a scholar on twentieth-century music. hyde identifies two modes by which a neoclassical piece “accommodates antiquity”: metamorphic anachronism and allegory. i argue both are present in gesang. first, stockhausen adopts elements of the sacred vocal tradition—including a child’s voice and antiphonal writing—and morphs them into something modern. second, stockhausen uses the biblical story on which gesang is based as an allegory for his own conflicted relationship with the musical past. this analysis reframes gesang’s significance and connects stockhausen’s work to seemingly unrelated trends in twentieth-century musical thought. 1 introduction karlheinz stockhausen (1928–2007) is often deemed a musical revolutionary. he claimed that he sought “to compose neither known rhythms nor melodies nor harmonic combinations nor figures.”[12] to this end, he was a major proponent of integral serialism, a compositional method pioneered in the mid-twentieth century that radically departed from previous musical idioms. consequently, analyses of his works tend to focus primarily on their pioneering qualities. this is particularly true of his 1956 work gesang der jünglinge, which is commonly considered a seminal piece in the history of electroacoustic music. however, such analyses obscure the extent to which gesang constructively engaged with prior musical traditions. far from avoiding the past, i argue that in gesang, stockhausen sought to integrate antiquated musical and textual materials within his modernist idiom, thus creating a deliberately ambiguous work that is both modern and ancient. to demonstrate this, i will employ an analytical framework developed by martha m. hyde to describe neoclassical pieces—works that intentionally reference and comment on past musical practices—and apply it to gesang. hyde, a scholar on twentieth-century music, identifies two modes through which a work might attempt to “accommodate antiquity” and hence be neoclassical: either through “metamorphic anachronism” or allegory.[9] [a] metamorphic anachronism occurs when a composer adopts saliently ana aresty rutgers undergraduate research journal, volume i, issue iii [b] while schoenberg was “conservative” relative to boulez and stockhausen, his music was controversial in its own right. schoenberg innovated the twelve-tone serial technique, a direct precursor to integral serialism. however, schoenberg’s reliance on classical models drew criticism from the later integral serialists, who felt these forms limited the possibilities afforded by the twelve-tone technique. chronistic materials and morphs them into something modern. allegory occurs when a composer uses music to dramatize their relationship to the canonical past. i argue that both modes are at play in gesang. regarding the first, i will demonstrate how gesang incorporated and transformed elements of the sacred vocal tradition. regarding the second, i will offer an interpretation of the piece that treats its narrative as an allegory for the conflict between tradition and modernity. in doing so, i aim to challenge the appraisal of gesang solely in terms of its novelty, instead opting for an interpretation in which the novel and the ancient commingle. 2 background the integral serialists asserted a common motivation: defying the past. integral serialism eschewed the conventions—including tonal harmony, symmetrical phrase structure, and strictly metered rhythmic gestures—that dominated the preceding classical and romantic periods. instead, musical decisions were governed by a strict series of mathematical proportions. moreover, serialist composers capitalized on budding technologies to create unfamiliar sound worlds. they extended experiments pioneered by pierre schaeffer (1910–1995) in musique concrète, a compositional approach founded in 1948 in which sounds recorded with new magnetic tape technology were manipulated to create novel sonic effects.[20] likewise, with the advent of synthesizers, composers began to write elektronische musik, an approach emerging in the 1950s in which composers generate unheard timbres from scratch rather than confining themselves to the timbres of traditional acoustic instruments.[6] perhaps the most vivid summary of the prevailing ethos was given by prominent serialist pierre boulez (1925–2016), who claimed that “all art of the past must be destroyed.”[16] stockhausen shared these ideological impulses. he declared his intent “to avoid everything which is familiar, generally known or reminiscent of music already composed.”[12] such inclinations are also reflected in his analytical writing. in his multivolume texte (1963–1971), which contained both theoretical writings and analyses of specific pieces, he sought to develop a new analytical language commensurate to the new musical language.[13] like boulez, stockhausen took great pains to craft a narrative of historical discontinuity; he sought to break free from, rather than extend, previous convention. in turn, existing literature on gesang der jünglinge tends to focus on its groundbreaking qualities. it has been lauded for dissolving the barriers between musique concrète (i.e., manipulating extant sounds) and elektronische musik (i.e., constructing new sounds ex nihilo), which had previously remained largely separate.[4] additionally, whereas prior serial compositions were constructed out of “fixed” values, such as discrete pitches, dynamic levels, and rhythmic values, gesang’s series was comprised of “non-stationary” shapes that evolved from one state to another over time.[5] gesang also advanced a new “statistical” approach to composition in which smaller musical events are conceived as contributing to large-scale perceptual phenomena called gestalts.[25] in many ways, stockhausen’s musical decisions mirrored his ideological assertions. such a trailblazing approach directly opposed that of the neoclassicists, more conservative composers who believed the “art of the past” must be revered rather than destroyed. one composer commonly associated with neoclassicism was arnold schoenberg (1874–1951).[b] emphasizing the connection between his music and his predecessors’, schoenberg claimed, “my teachers were primarily bach and mozart, and secondarily beethoven, brahms, and wagner …”[19] further, his string quartet no. 3 op. 30 (1927) was explicitly modelled after franz schubert’s string quartet in a minor, d 804 (1824).[17] integral serialist composers were critical of this approach. boulez, for one, derided schoenberg’s propensity to model his works on older ones, believing it led to the “decrepitude” of schoenberg’s works. rather than hearing connections to a rich his aresty rutgers undergraduate research journal, volume i, issue iii tory in schoenberg’s music, boulez heard “reminisces of a dead world.”[2] given the integral serialists’ animosity toward the neoclassicists’ approach, one would not expect many similarities between their respective works. yet i contend that gesang der jünglinge defies this expectation; it integrates prior traditions in ways strikingly analogous to the works of past-reverent composers like schoenberg. i will subsequently demonstrate this by analyzing gesang within scholar martha m. hyde’s model of neoclassicism, which identifies metamorphic anachronism and allegory as common features of neoclassical works. admittedly, orienting stockhausen within this framework may seem incongruous, or at the very least unorthodox. indeed, i intend it to seem incongruous. the fact that stockhausen and the neoclassicists appear so dissimilar, perhaps even antithetical, makes the revelation of their hidden commonalities even more striking. however, i do not mean to claim that there are no relevant differences between stockhausen and the composers who are commonly considered neoclassical. rather, i intend to demonstrate that stockhausen drew from the past in remarkably similar ways to the neoclassicists, contrary to what his rhetoric would suggest. despite his self-proclaimed desire to avoid the familiar, stockhausen too did not entirely eradicate from his music the reminisces of a dead world. 3 defining neoclassicism analyzing stockhausen within a neoclassical lens is difficult given the notorious ambiguity of the term “neoclassicism.” in its narrower definition, neoclassicism is a style most strongly associated with russian composer igor stravinsky (1882–1971) and characterized by features like “clarity, simplicity, objectivity, purity, refinement, constructive logic, concision, sobriety, and so on.”[11] such characteristics, though vague, are generally understood to refer to elements of classical and baroque styles, as contrasted with the emotional indulgence of the romantic era. by this definition, a neoclassical work would be one that seeks to reinstate the order of classical style. however, this categorization of neoclassic cism as a style creates as much confusion as it dispels. because different composers elaborated upon the work of the classical “masters” in radically different ways, the composers who eventually became associated with neoclassicism do not represent any unified style. for instance, both stravinsky’s and schoenberg’s neoclassical works are deeply influenced by elements from the classical era, yet they bear little stylistic resemblance to one another. it was this vagueness that led many to repudiate the term’s usefulness, as when composer-theorist milton babbitt (1916–2011) called it a “catch phrase… to be talked about by those who could not and should not talk about the music.”[24] hyde attempts to recoup some of the term’s value by defining it not as a style, but more broadly as an ideological orientation. she claims that “to be neoclassical” is to “striv[e] to be modern as well as ancient.”[9] that is, a piece is neoclassical if it saliently attempts to accommodate antiquity within a modern context. this conception of neoclassicism concerns itself less with the classical in its capital-c sense (i.e., referring narrowly to the classical era), and more with the colloquial (lowercase-c) definition of a “classic” work: “a past work that remains or becomes relevant and available as a model, or can be made so through various techniques of accommodation.”[9] it is in this broader sense that i will employ the term. there are two possible objections to this conception of neoclassicism. the first is that by prioritizing ideology over style, hyde’s model departs too radically from the common understanding of neoclassicism and is thus more of a re-definition than a definition. however, hyde’s model is built inductively from four musical examples commonly associated with the term. far from erasing the common sense of the term, hyde elaborates upon it, observing the characteristics that unify the term’s diverse usage. the second potential objection is that hyde’s conception is too broad to meaningfully describe any unified concept. scott messing notes that early twentieth-century european composers almost unilaterally felt “obliged to relate themselves to the history created by their forebears.”[11] as a result, he writes, “almost every major figure composing during aresty rutgers undergraduate research journal, volume i, issue iii [c] daniel 3:5 (niv). [d] daniel 3:6 (niv). the first three decades of [the twentieth] century was tied, loosely or umbilically, to this term.”[11] by implicating stockhausen in this trend, i run the risk of complicating matters even further. but, finding that a term is broadly applicable does not necessarily imply it is too broadly defined. hyde names a specific ideological value and further articulates specific methods by which composers tended to attain it (allegory and metamorphic anachronism). such elements are by no means universal, nor are they superficial: they have deep implications for the intent and impact of a work. thus, if we are surprised by the broad applicability of her model, it is not because she defines the model too loosely. rather, it is because the aesthetic values the model describes are more widespread than we originally supposed. 4 mode i: metamorphic anachronism the first mode of accommodating antiquity is “metamorphic anachronism.” though hyde frames it as a single phenomenon, it is useful to view it as a composite of two processes. on one hand, anachronism entails adopting an incongruously antiquated style or idea; on the other hand, metamorphosis entails subjecting a basic material to some form of transformation. these two processes should be understood as distinct, because either could conceivably occur without the other. for instance, if a twentieth-century piece were written in strict imitation of mozart, it would be anachronistic but not metamorphic. conversely, pierre schaeffer’s musique concrète piece etude aux chimens de fer (1948) is metamorphic in that it transformed its constituent materials, but it is not anachronistic since those constituent materials (recorded sounds of trains) were not antiquated.[14] hyde’s model of neoclassicism requires the confluence of both processes. that is, a piece exhibits metamorphic anachronism only if it first adopts an anachronistic idea and then applies some metamorphic process by which the anachronism becomes “modern.” for instance, schoenberg’s third string quartet is rightfully considered neoclassical because it adopts stylistic and formal elements of schubert’s a minor string quartet (anachronism) and transforms these elements through application of his twelvetone technique (metamorphosis). to establish that metamorphic anachronism is also present in gesang, it must be independently shown both that the piece is anachronistic and that it is metamorphic. anachronism anachronism entails juxtaposing elements from disparate eras. in the case of gesang, stockhausen juxtaposes a modernist musical aesthetic against a biblical story from the book of daniel, chapter 3. in this story, the king nebuchadnezzar constructs a giant idol out of gold and decrees that people of all nations and languages must worship it whenever they hear “the sound of the horn, flute, zither, lyre, harp, pipe and all kinds of music.”[c] he threatens that those who refuse “will immediately be thrown into a blazing furnace.”[d] when three jews—shadrach, meshach, and abednego—refuse to worship the false idol, nebuchadnezzar orders them thrown into the fire. but instead of perishing, the three young men are protected by an angel and emerge from the flames unharmed. they sing a song praising god; excerpts of this song comprise gesang’s text. such a subject matter was a stark departure from previous serial experiments in elektronische musik and musique concrète. prior avant-garde pieces had borne abstract titles, such as stockhausen’s konkrete etüde (1952) and studien i & ii (1953–1954), boulez’s études i & ii (1951–1952), or french composer olivier messiaen’s timbres-durées (1952). with such titles, composers deliberately distanced themselves from any historical references. in stark contrast, gesang’s use of biblical verse implicated a rich history of religiously inspired artwork. to employ a subject as traditional as biblical verse, then, would seemingly contradict the serialists' purported desire to distance themselves from past traditions. yet in stockhausen’s original vision for the piece, the anachronism was even more striking. he initially intended the piece as an explicitly liturgical aresty rutgers undergraduate research journal, volume i, issue iii work to be premiered in the cologne cathedral. however, the cathedral, sensing the incongruity, rejected his proposal on the grounds that loudspeakers would be inappropriate in a cathedral.[21] only after being forced to revise his concept did stockhausen decide upon the story of the youths in the furnace. the voice that carries this text also bears the mark of the ancient. the human voice—used in christian liturgical settings since at least 700 a.d.—has a well-established link to antiquity.[10] the boy soprano specifically recalls the sound of the now-archaic castrati, boys who were castrated by the church before puberty to preserve their pure vocal tone.[4] further, the use of several loudspeakers that surround the congregation recreates the spatial acoustics of a cathedral. indeed, stockhausen originally intended the speakers to be suspended above the congregation, as if coming from heaven itself.[8] similarly, by sometimes isolating a single vocal track and other times layering many vocal tracks atop one another, stockhausen mimics antiphony between a cantor and a chorus of singers.[4] and, as russell wallace chait notes, gesang’s fluctuation between clearly audible text and obscured text alludes to the ancient debate over intelligibility of text in church music, dating back to the council of trent.[4] gesang’s sonic characteristics thus establish a firm link to the christian sacred vocal tradition. metamorphosis these anachronistic materials—both the sound of the boy’s voice and the text it conveys—were then “accommodated” within modernity by subjecting them to metamorphic processes. gesang was influenced by musique concrète, which is metamorphic by its very nature. pierre henry (1927–2017), one of the genre’s pioneers, thought of musique concrète as an approach in which sound is used like a physical material, able to be “render[ed]… plastic like sculpture.”[1] accordingly, concrète composition entailed literally transforming a physical material (magnetic tape) by cutting it, reversing it, changing its playback speed, and so on. as a result, the recorded sound was also morphed, so that the new sound might bear some resemblance to the original but would acquire unique sonic properties distinct from those of its source. in gesang, the source material subjected to such manipulation is the recorded boy soprano. the tape, then, served as a physical proxy for the boy’s voice; as stockhausen spliced the tape, he transfigured the voice. since the boy soprano’s voice itself was a proxy for sacred vocal music at large, stockhausen’s tape manipulations represented a metamorphosis of the sacred vocal tradition. in fact, it is unlikely such a rigorous piece could have been realized at all if not for tape’s capacity to transform the boy’s voice, for stockhausen’s compositions were demandingly precise. pitches were expressed not as chromatic notes, but as frequencies, while durations were expressed not in beats, but in centimeters.[22] as pierre schaeffer later remarked, this musical style made traditional notation “anachronistic through a rigor so absolute that the approximations of traditional scores paled before such precision.”[18] what was truly “anachronistic,” however, was not the notation per se, but that stockhausen would ask a singer (for whom such traditional notation would typically be necessary) to execute such precise music, rather than a synthesizer. at least in part, the limitations of traditional notation reflect the limitations of performers themselves. thus josef protschka, the boy whose voice appears on gesang, could only approximate stockhausen’s vision by mimicking synthesized sine-tones played through headphones.[12] the remainder of stockhausen’s vision had to be accomplished through tape manipulation. thus, the demands of serial technique required the boy’s voice be transfigured into something more “plastic” than human physiology would typically allow. the addition of synthesized tones further morphed the voice into something altogether foreign, perhaps even inhuman. stockhausen conflated the sound of the voice and the synthesizer, so that it is unclear where one begins and the other ends. he achieved this effect by creating synthesized analogues of vocal phonemes. sine-tones (pure electronic tones) were combined so as to imitate the overtone structure of various vowel sounds; filtered white noise imitated fricative and sibilant consonants (e.g., f, th, s, sh); and electronic “impulses” (percuss aresty rutgers undergraduate research journal, volume i, issue iii sive sounds with a natural decay) imitated plosive consonants (e.g., p, b, t, d).[5] further, stockhausen’s serial design did not treat these vocal and synthesized phonemes separately. instead, he integrated them on a single continuum of sounds, so that they would seamlessly blend with one another.[5] such blending is at times made explicit, as when the final consonant of the word eis is transformed into a synthesized hiss or when the final consonant of mond is followed by an explosion of electronic sounds resembling plosive consonants.[23] stockhausen blurs the distinction between electronic and acoustic, morphing the voice into something that is both familiar and new. 5 mode ii: allegory if metamorphic anachronism concerns how antiquated materials are employed in modern contexts, then allegory concerns why: for what rhetorical purpose are these antiquated materials employed? j. peter burkholder suggests that many twentiethcentury composers used musical materials and structures as allegories for their relationship with the canonical past.[3] for instance, he argues schoenberg’s preference for continuous variation rather than direct repetition stemmed from a desire to develop upon, but not to repeat, his forebears. conversely, john cage—a composer as far from a neoclassicist as is conceivable—embraced randomness and unpredictability in his music, thus negating any deliberate relationship with the past. through their music, composers attempted to define their relationship to their forebears. the same is true of stockhausen. in fact, gesang contains even greater allegorical potential since it portrays a literal narrative. the story of the three youths lends itself particularly well to allegorical reading in that it deals with the archetypical themes of persecution and perseverance, which can be mapped onto many different situations. the boys broadly represent purity, innocence, and steadfastness. king nebuchadnezzar and the furnace represent destructive forces that threaten such purity. the ubiquity of these themes makes the story amenable to allegorical readings. however, its flexibility also raises the possi bility of several contradictory interpretations. indeed, the allegorical subtext of gesang is seemingly paradoxical. on one hand, stockhausen asserts that modernity—specifically, integral serialism—is being “persecuted” by past-obsessed critics. on the other, it is antiquity that is “persecuted” by modernity which threatens to consume it in a flurry of electronic flames. the contradiction between these allegories demonstrates stockhausen’s complex relationship with modernity and antiquity. persecution of modernity in a 1998 interview, stockhausen claimed he personally identified with the youths in the flames. regarding the composition of gesang, he said: “…i myself felt like a young man in the furnace at that time. everything i did was aggressively turned down and damned by the music journalists and musicologists of the time. there was a professor blume, chairman of the german musicological society, who in a large text wrote that stockhausen was laying the ax to the roots of music and was destroying all of occidental music. therefore, i felt so like the young men in the furnace, and i could only pray that st. michael would come and pull me out of the flames.”[15] stockhausen felt persecuted by reactionaries as the youths were persecuted by nebuchadnezzar, forced to bow to an idol in which he did not believe (i.e., the tastes of journalists and musicologists). while it may seem hyperbolic to compare musical criticism to the religious persecution the youths faced, stockhausen’s aesthetic values were so deeply intertwined with his faith that an attack on one would constitute an attack on the other. he claimed: “…the proportions in my music have always been related to everything i learn from the nature of the stars and galaxies and, on the other hand, from the atoms and molecules, and the cells. everything in my music is an extension of what i experience as creation—how creation is composed.”[15] stockhausen believed the strict mathematical proportions governing serial music mirrored the orderly, mathematically-precise processes governing the cosmos. to him, serialism was the inevitable result of striving toward a more perfect, god-lier music. thus, the severe aesthetic criticism levelled at stockhausen was, in effect, a form of religious persecution. aresty rutgers undergraduate research journal, volume i, issue iii persecution of antiquity although stockhausen saw serialism as an expression of faith, serialism also posed the risk of obscuring, or even altogether destroying, the message of faith conveyed by gesang’s text. while in a more traditional vocal setting the vocalist is made the focal point and the music is subservient to the text, integral serialism is predicated on “the calculated parity of parts” within an ensemble, vocalist included.[12] in fact, privileging the comprehensibility of the text might “jam the precise inner workings of serial schemes.” [12] so, to conform to serial ideals, the text had to be subsumed within the serial scheme, not served by it. stockhausen confirmed this conception of the relationship between music and speech in his 1958 article musik und sprache, in which he advocated for a progressive “transition from speech to music” in serial vocal music.[12] for a piece based on biblical text, this has troubling allegorical implications, especially since the voice delivering said text had such strong ties to church tradition. strict adherence to serialist principles, in which the vocalist is wholly subsumed by the series, would amount to the youths’ metaphorical destruction. in this light, modernism is the furnace into which the boys are thrown. in fact, gesang’s wild bursts of electronic impulses strongly resemble bursts of flames. some critics were inclined to hear the piece in these terms. in a piece entitled “wider die natur!” (“against nature!”), one went so far as to say that stockhausen was condemning the youths, a “gift of the divine,” to a “hellish” electronic sound world.[12] thus, we are faced with two conflicting allegories: one in which antiquity persecutes modernity and one in which the opposite is true. an incomplete struggle while these allegorical interpretations might seem irreconcilable, the apparent contradiction demonstrates stockhausen’s desire to simultaneously accommodate the old and the new. as a result, in gesang, neither persecutor prevails. the old, of course, was not able to eradicate the new; if stockhausen was thrown into the furnace by his critics, he clearly survived the flames. the remarkable success gesang has enjoyed since its premiere proves as much. nor was the new able to eradicate the old. though his serial methods sometimes partially obscured the biblical text, stockhausen did not render the text (and its attendant meanings) altogether incomprehensible. by incorporating the text at varying degrees of comprehensibility, he guaranteed that, at least sometimes, the message of the text would be heard clearly.[5] moreover, the phrase that is articulated most clearly, preiset den herrn (praise ye the lord), is also the most crucial to the text’s meaning. neither the ancient nor the modern successfully destroys the other; they coexist in an uneasy tension. the sense of open-ended conflict is reflected in the piece’s inconclusive ending. stockhausen originally intended to compose seven sections (labelled a-g), but due to time constraints, he was able only to complete the first six.[5,15] while it may seem unfair to attribute meaning to compositional decisions made for such a prosaic reason, i contend they have symbolic significance. by eliminating section g, stockhausen partially compromised his modernist ideal of mathematical perfection; the attention to detail demanded by his serial designs proved too time-consuming to be feasible. since the serial design was itself incomplete, any destruction of antiquity it might entail would also necessarily be incomplete. further, because the end of section f proceeds without transition into the closing gesture, which was originally intended to come later, the piece’s ending sounds abrupt and uncertain.[5] this suggests that the end of the piece is not where the conflict between old and new truly ends; it is ongoing, perhaps even perpetual. the ancient will continue to struggle against annihilation by the modern and the modern will continue to face resistance from an inert past. gesang demonstrates that such tension can be constructive. in fact, both antiquity and modernity are reaffirmed having been challenged by each other, much like how the biblical youths’ faith is reaffirmed after being challenged by the flames. by placing ancient traditions in a modern context, stockhau sen revitalized these traditions, demonstrating their continued relevance in the era of electronics. likewise, by incorporating familiar biblical themes, he aresty rutgers undergraduate research journal, volume i, issue iii clarified the religious subtext of his pieces, which had previously been opaque. in gesang, the tension between old and new led to creation, not destruction. 6 conclusion stockhausen outwardly presented himself as one who defied convention; he was a “past-destroyer.” yet, his relationship to his musical forebears was more complex than such rhetoric would imply. in gesang der jünglinge, rather than destroying antiquity, he sought to accommodate it within a modern idiom. the piece is based on biblical text and its musical elements are deeply informed by the sacred vocal tradition. through tape manipulation and the addition of the synthesizer, these elements are transfigured into something that is both familiar and novel. furthermore, stockhausen employed the story of the youths allegorically, using it to demonstrate his contradictory identification with both the old and the new. thus, gesang is simultaneously modern and ancient; in this broad sense, it is a “neoclassical” work. this account of gesang helps reframe the piece’s significance, both within serial composition and twentieth-century music overall. early electronic serial compositions, by virtue of their abstractness, remained divorced from the rich network of meanings afforded by tradition. gesang bridged this divide; it brought tradition into the realm of the modern, thereby revitalizing both. paradoxically, gesang was innovative because it looked backwards. rather than abandoning the “reminisces of a dead world,” stockhausen reinvented them to create a world anew∎ 7 references [1] battier, marc. “what the grm brought to music: from musique concrète to acousmatic music.” organised sound 12, no. 3 (2007): 189–202. [2] boulez, pierre. “schoenberg is dead.” in notes of an apprenticeship, 268–279. new york: a. a. knopf, 1968. [3] burkholder, j. peter. “musical time and continuity as a reflection of the historical situation of modern composers.” the journal of musicology 9, no. 4 (1991): 411-29. [4] chait, ross wallace. “gesang dream: functions of faith in stockhausen’s electric mass.” perspectives of new music 52, no. 3 (2014): 185–196. [5] decroupet, pascal and elena ungeheuer. “through the sensory looking-glass: the aesthetic and serial foundations of gesang der jünglinge.” translated by jerome kohl. perspectives of new music 36, no. 1 (1998): 97–142. [6] emmerson, simon, and denis smalley. "electro-acoustic music." grove music online. 2001. [7] holy bible: new international version. colorado springs: biblica, 2011. [8] howard, luke. “the voice(-over) of god: some thoughts on the disembodied voice in contemporary music.” open space magazine 1 (1999): 109-116. [9] hyde, martha m. “neoclassic and anachronistic impulses in twentieth-century music.” music theory spectrum 18, no. 2 (1996): 200–235. [10] kelly, thomas forrest. introduction to chant and its origins, xi-xix. edited by thomas forrest kelly. new york: routledge, 2016. [11] messing, scott. neoclassicism in music: from the genesis of the concept through the schoenberg/stravinsky polemic. ann arbor: umi research press, 1988. [12] metzer, david joel. “the paths from and to abstraction in stockhausen’s gesang der jünglinge.” modernism/modernity 11, no. 4 (2004): 695–721. [13] morgan, robert p. “stockhausen’s writings on music.” the musical quarterly 61, no. 1 (1975): 194–206. [14] palombini, carlos. “machine songs v: pierre schaeffer: from research into noises to experimental music.” computer music journal 17, no. 3 (1993): 14-19. [15] peters, günter. “‘…how creation is composed’: spirituality in the music of karlheinz stockhausen.” translated by günter peters and mark schreiber. perspectives of new music 37, no. 1 (1999): 96-131. [16] peyser, joan. to boulez and beyond. lanham: scarecrow press, 2008. [17] rosen, charles. arnold schoenberg. princeton: princeton university press, 1975. [18] schaeffer, pierre. treatise on musical objects: an essay across disciplines. translated by christine north and john dack. oakland: university of california press, 2017. [19] schoenberg, arnold. style and idea. translated by leo black. edited by leonard stein. new york: st. martin’s press, 1975. [20] smalley, denis. “groupe de recherches musicales.” grove music online. 2001; accessed 28 dec. 2020. [21] smalley, john. “gesang der jünglinge: history and analysis.” columbia university department of music. 2000. [22] stockhausen, karlheinz. gesang der jünglinge, elektronische musik: faksimile-edition 2001. kürten: stockhausen verlag, 2001. [23] stone, kurt. “karlheinz stockhausen: gesang der jünglinge (1955/56).” the musical quarterly 49, no. 4 (1963): 551-54. [24] taruskin, richard. “back to whom? neoclassicism as ideology.” 19th-century music 16, no. 3 (1993): 286-302. [25] ungeheuer, elena. “statistical gestalts – perceptible features in serial music.” in music, gestalt, and computing – studies in cognitive and systematic musicology, 103–113. berlin, heidelberg: springer-verlag, 1997. aresty rutgers undergraduate research journal, volume i, issue iii andrew faulkenberry is a senior undergraduate at rutgers studying music composition. as a composer, his works have been premiered by groups including the international contemporary ensemble, rutgers wind ensemble, and julius quartet. andrew is also a recipient of the presser foundation undergraduate scholar award. andrew’s research began during an independent study with dr. rebecca cypess and was subsequently completed independently. it was motivated by a broader interest in twentieth-century musical aesthetics and their underlying ideologies. particularly, andrew became fascinated by two contradictory narratives that pervaded twentieth-century musical thought. on one hand, composers were concerned with “forward progress” and sounding “modern”. yet, on the other, they sought to demonstrate their connection to a fixed canon of “classic” works. through his research on karlheinz stockhausen, andrew hopes to demonstrate that these contradictory desires influenced even the most radical composers. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. exploring epithelial communication to the mesenchyme and its impact on the expression of genes related to tumorigenesis jay patel, michael verzi (faculty advisor) ✵ abstract homeostasis of the epithelium is dependent on the wnt/beta-catenin pathway, which regulates the proliferation of intestinal stem cells. the gain of function mutations in the beta-catenin gene leads to rapid cell proliferation and malignant growth in the epithelium. in addition, the maintenance of these stem cells appears to be linked to mesenchymal-derived factors. although the communication between epithelial and mesenchyme cell populations remains uncharacterized, understanding this mechanism will help us further understand the various pathways involved in tumor initiation processes. our results show that the communication between the mesenchyme and epithelium during transformation is influenced by varying levels of protein-expressing genes including wnt2b, grem1, and bmp6. 1 introduction the wnt/beta-catenin pathway regulates the transcription of wnt target genes—genes which are highly deregulated in solid tumors (fevr et al., 2007). in the absence of the wnt ligand, betacatenin, a protein normally sequestered in the cytoplasm, is constantly degraded and is prevented from entering the nucleus. in the presence of the wnt ligand, wnt is able to bind to its receptor frizzled, which activates downstream signaling and prevents the destruction of beta-catenin, thus enabling its entry into the nucleus. in the nucleus, beta-catenin promotes the transcription of wnt target genes, causing rapid cell proliferation (cleavers and nusse, 2012). cells dividing at an accelerated rate can become cancerous and eventually lead to the formation of a tumor. the intestinal stem cell niche is primarily made up of several cell types, including: (1) differentiated cells responsible for the transport and absorption of nutrients, and (2) mesenchymal cells which supply signals to the epithelial cells to maintain homeostasis. increased wnt signaling can be induced by activating the beta-catenin gene, which allows for abnormal epithelial growth. due to the expression of tamoxifen inducible cre recombinase, villin-creert2 transgenic mice have the ability to delete exon 3 of the beta-catenin allele in the intestinal epithelium (marjou et al., 2004). mice expressing villin-creert2 can delete a portion of the beta-catenin gene in the presence of tamoxifen, resulting in a more transcriptionally active form of the gene. mouse models have shown that mice with villin-creert2 and injected with tamoxifen show visible malignant tissue transformation in the duodenum. but in organoid models, exposed to tamoxifen, aresty rutgers undergraduate research journal, volume i, issue iv rapid cell proliferation can be seen almost immediately. however, epithelial transformation only becomes visible around day 13 in mouse models. although organoids mimic intestinal stem cells (crypts) in vitro, they do not provide a holistic view of a biological system. for example, they lack the presence of the mesenchyme. this suggests that the mesenchyme must play some role in the suppression of mutant growth. however, the way mesenchyme and epithelial cells influence one another is unknown. yet, it is known that mesenchymal-derived factors are essential to maintain intestinal epithelial stem cells, so the reduction of these factors may be contributing to the suppression of intestinal transformation in the transgenic mouse (stzepourginski et al., 2017). areg, a ligand of egfr (epidermal growth factor receptor), can trigger signaling cascades that mediate cell survival, proliferation, and motility. thus, areg may be a candidate ligand that allows the epithelium to communicate with the mesenchyme (wang et al., 2020). transformed epithelium expresses high levels of areg, which may be sensed by mesenchyme cells. mesenchyme cells like pdgfrahi are enriched near the crypt-villus junction where stem cells reside and promote bmp ligands (promotes differentiation/blocks proliferation). a subset of pdgfralo cells can be found at the bottom of crypts and provide growth factors for stem cells (mccarthy et al., 2020). if areg is communicating with pdgfralo cells in the mesenchyme, the pdgfralo cells would express decreased levels of grem1, a protein that functions to suppress bmp and permits proliferation. subsequently, this may lead to an upregulation in expression levels of bmp. early intestinal tumorigenesis reflects the deregulation of wnt and bmp signals (davis et al; 2015). it has also been recognized that bmp signaling restricts intestinal epithelium hyperproliferation (qi et al., 2017). if we could understand how the mesenchyme is communicating with the epithelium, it would give insight to how it is able to suppress the abnormal growth, allowing for novel therapeutic interventions during initial tumor development. the goal of this project is to learn how epithelial cells are communicating with mesenchyme cells in an elevated wnt activity background. elevated levels of wnt prevents beta-catenin from being degraded, which promotes expression of wnt target genes, resulting in the mutant phenotype. through signals (e.g. areg) sent by the epithelium, the mesenchyme may detect the cellular transformation and respond in hopes of preventing the abnormal growth. if mesenchyme cells are treated with conditioned media from mutant organoids (crypts), then the genes which are altered are communicating with the mesenchyme. if epithelial cells undergo abnormal growth, the ligands produced could be sensed by pdgfralo cells in the mesenchyme compartment, which may respond by increasing bmp genes and decreasing grem1 and wnt2b production. 2 materials and methods animal and tissue processing animal experiments are conducted in accordance with rutgers university institutional animal care and use committee (iacuc).. the transgenic mice in this study are engineered to conditionally express the beta-catenin exon3 deletion. to induce recombination, tamoxifen (1mg/20g) is injected intraperitoneally for 4 days daily. day 1 is considered the beginning of treatment, and mice are sacrificed and investigated at day 10. this model is ideal for studying tumorigenesis in-vivo because of similarities in mouse and human gastrointestinal tracts. at day 10, mouse intestines (duodenum) are collected and fixed in 4% paraformaldehyde solution overnight at 4°c. tissues are then washed with phosphate-buffered saline (pbs) and dehydrated in increasing concentrations of ethanol. tissues are then transferred to xylene and paraffin mixture for one hour, then 100% paraffin for another hour. tissues are embedded for sectioning. organoid forming assay crypt-derived organoids are isolated from mouse duodenum on day 10 after 4 consecutive treatment days of tamoxifen administered at 1mg/20g mice. collected duodenum samples are washed in pbs, cut into ¼ inch pieces, and rotated in 3mm edta for 5, 10, and 20 minutes (each time aresty rutgers undergraduate research journal, volume i, issue iv replace the edta). tissues are then agitated and filtered through a 70μm filter for crypt enrichment. crypts are then washed in pbs and the pellet is resuspended in bme-r1. media for the crypts include advanced dmem supplemented with glutamax (auxiliary energy source for rapidly dividing cells), hepes (used to maintain ph), egf (epidermal growth factor: involved with cell signaling pathway controlling cell division), noggin (prevents cell differentiation), nac (n-acetylcysteine: protects against rise of internal oxidant levels), n2 (promotes in vitro differentiation of stem cells), b27 (promotes growth and proliferation, but not differentiation), rspo (positively regulates wnt/beta-catenin signaling, which induces proliferation), and pen-strep (prevents bacterial contamination). after 3 days of organoids being treated with this media, it becomes conditioned media (contains ligands excreted by epithelium). genotyping mice toes are clipped and used as a source for dna genotyping analysis. dna is extracted using the kappa buffer qiagen kit followed by pcr for amplification. genotypes are confirmed by dna electrophoresis. hematoxylin and eosin staining to understand the changes in the epithelial architecture in terms of the formation of crypt progenitor cell phenotype (cpc), we conducted h&e histological analysis. tissues are initially washed in xylene twice (5 min each). then tissues go through two consecutive 100% etoh washes (5 min each), followed by a 3 min 95% etoh wash, a 3 min 85% etoh wash, a 3 min 70% etoh wash, and a 5 min double-distilled water wash. unstained sections of mouse duodenum are stained with hematoxylin (30 seconds) and dehydrated by dipping in water, followed by increasing alcohol percentages. tissues are then counterstained with eosin for 1 minute (eosin in counterstain, which distinguishes between cytoplasm and nuclei). rna/dna preparation and qpcr mesenchyme cells are dissolved in trizol. rna is prepared according to the manufacturer’s protocols. the rna is then reverse transcribed to cdna using superscript iii first-strand synthesis system. sybr green is used to amplify the cdna through qpcr analysis. vimentin is used for normalization. mesenchyme treatment with conditioned media the main experiment consists of using mesenchyme cells and treating them with (1) conditioned media from mutant organoids treated with ccm, (2) mutant organoids treated with mesenchyme media, (3) wild type organoids treated with ccm, and (4) mesenchyme media as a control. to demonstrate the response by the mesenchyme, changes in ligand expression are examined among treatment groups. after passaging mesenchyme cells, they are plated, and are treated with mesenchyme media on day 1. pictures of the cells are taken on day 4, and the media is aspirated and replaced with conditioned media or mesenchyme media for control. the plate consists of 4 rows and 3 columns; rows 1 and 2 are made up of a confluent wild-type mesenchyme cells, and row 3 consists of sub-confluent wild type mesenchyme cells. samples in the first column serve as controls and are treated with mesenchyme media, samples in the second column are treated with conditioned ccm from mutant organoids, samples in the third column are treated with conditioned mesenchyme media from mutant organoids, and samples in the fourth column are treated with conditioned ccm from wild type organoids. the fourth column is a control because the wild type organoids do not express mutant growth, so we expect there to be no change in grem1 and wnt2b expression. on day 6, the cells are harvested with trizol for rna isolation, which is then converted to cdna. for this experiment we measure 11 protein coding genes (bmp2, bmp3, bmp4, bmp5, bmp6, grem1, wnt2b, rspo1, rspo3, wnt5a, and wnt4) using qpcr and compare their relative abundances. aresty rutgers undergraduate research journal, volume i, issue iv 3 results tissue images are taken at day 10 posttamoxifen treatment and stained using hematoxylin and eosin. figure 1 shows the histology of the control and mutant epithelium. the control and mutant epithelium for replicates 1 and 2 show little difference in terms of transformation. this demonstrates the delay stated before, suggesting that the mesenchyme is playing a role to suppress mutant epithelial growth. to understand the mesenchyme response to ligands secreted by mutant epithelium, gene expression changes are examined following treatment with conditioned media derived from organoid cultures. qpcr (figure 2a and figure 2b) analysis demonstrated that confluent and sub-confluent mesenchyme cells do not respond in the same way. for the confluent cells, grem1 and wnt2b tend to be downregulated. grem1 is a bmp antagonist; therefore, by downregulating the gene bmp can have a larger effect on the epithelium which would lead to a decrease in activity of the wnt/beta-catenin pathway. for the sub-confluent mesenchyme cells, grem1 is downregulated in all experimental groups while wnt2b is also downregulated except in cells treated with conditioned ccm from wild type organoids. again, the conditioned media from mutant organoids are causing grem1 and wnt2b to be downregulated in order to suppress mutant epithelial growth. figure 1: h&e staining of two independent replicates showing the difference between the control and mutant epithelium. figure 2a: the image shows the density of the cells; cells are confluent. 647 is the confluent mesenchyme cell line. the leftmost bars represent grem1 and wnt2b expression in the control: mesenchyme cells treated with mesenchyme media. to the right of that is mesenchyme cells treated with mesenchyme conditioned media from mutant organoids (647 mm-cm exon3f/f). to the right of that is mesenchyme cells treated with ccm condition media from wild type organoids (647 ccm-cm wt), and the rightmost graph shows mesenchyme cells treated with ccm conditioned media from mutant organoids (647 ccm-cm exon3f/f). the graph shows relative gene expression (n=3), compared to cells treated with mesenchyme media. aresty rutgers undergraduate research journal, volume i, issue iv figure 2b: image shows the density of cells; cells are sub-confluent. graph shows relative gene expression of the three samples treated with conditioned media to the mesenchyme media for sub-confluent cells. graph portrays same conditions as graph above except for sub-confluent cells. the graphs in figure 3 demonstrate the relative gene expression seen across all 11 protein coding genes that were tested. the graph focuses on the mesenchyme conditioned media from mutant organoids vs. the control. the bmp profile between the confluent and sub-confluent cells shows many similarities including bmp2 and bmp3 being downregulated, bmp4 staying constant, and bmp6 being upregulated. the upregulation in sub-confluent cells is much more profound. the wnt profile in confluent and sub-confluent cells contain some differences like rspo3 being upregulated in confluent cells while being downregulated in sub-confluent cells. while wnt5a is upregulated in both confluent and sub-confluent, wnt4 is downregulated in confluent, while profoundly upregulated in sub-confluent cells. 4 discussion the goal of the experiment is to identify possible signaling pathways that mesenchymal cells could use to communicate with epithelial cells. while the multitude of cells in the small intestine and their functions are known, like trophocytes, telocytes, and enterocytes (mccarthy et al., 2020), their ability to communicate with the microenvironment is understudied. the signaling crosstalk between the stroma cell population and the epithelium remains unknown. understanding how these two types of cells may communicate can provide novel therapeutic interventions during initial tumor development. histological review of h&e stained slides reported a lack of altered or abnormal tissue structures, suggesting that the mesenchyme is responsible for suppressing mutant epithelial growth. protein expressing genes such as wnt2b, grem1, and bmp6 may be responsible for the cross talk between these two cell types. when conditioned media from mutant organoids is introduced to mesenchyme cells grem1 and wnt2b are downregulated. the downregulation of wnt prevents the expression of wnt target genes, therefore controlling mutant transformation. grem1 in these cell lines is seen to be downregulated, which directly and indirectly (by inhibiting bmp genes) prevents epithelial transformation. what is unusual is why most of the bmp genes are either downregulated or stay constant while bmp6 is upregulated. bmp signaling can be seen as the primary suppressor of dedifferentiation in the intestinal epithelium, and better understanding of its ligands will provide important steps to how the epithelium protects against oncogenesis (perekatt et al., 2018). although the bmp gene expression is found to aresty rutgers undergraduate research journal, volume i, issue iv figure 3: graphs show relative gene expression of the aforementioned 11 genes in sub-confluent and confluent mesenchyme cells. graph compares the major experimental group to the control group. because ccm is the primary media used to grow and sustain organoids, its data would be most resembled in-vivo conditions. this group would have organoids function normally so that the conditioned media would mimic in-vivo conditions, and then the mesenchyme response could be observed. not increase, this could be because grem1 expression is downregulated, leading to the translation of more bmp proteins. the up and down regulation of bmp, grem1, and wnt2b ligands may be, in part, how the mesenchyme is influencing epithelial cells, preventing early tumorigenesis. the experiment demonstrated the differences between confluent and sub-confluent mesenchyme cells, as well as how different conditioned media affects mesenchyme cells. these differences are most profoundly seen in the genes bmp6 and wnt4. confluent cells have the ability to sense when they are running out of space, which can cause the cells to become less proliferative. sub-confluent cells are rather in a more active growth phase. this could be affecting the signals the mesenchyme releases to the aresty rutgers undergraduate research journal, volume i, issue iv epithelium, and responsible for the difference seen in response between confluent and subconfluent cells. these results come from a single experimental design and provide a basis for future experiments. bmp, grem1, and wnt2b may be signals that allow the mesenchyme to communicate with the transformed epithelium. these genes could be responsible for early prevention of initial tumorigenesis. these genes are potential ligands that suppress mutant epithelial growth. from here, geneticists could design experiments involving blocking these ligands with antibodies or treatment with these ligands to see the epithelial response from the mesenchyme∎ 5 references [1] clevers, hans, and roel nusse. “wnt/β-catenin signaling and disease.” cell, cell press, 7 june 2012, https://www.sciencedirect.com/science/article/pii/s0092867412005867. [2] mccarthy, neil. “distinct mesenchymal cell populations generate the essential intestinal bmp signaling gradient.” define_me, cell press,. [3] davis, hayley, et al. “aberrant epithelial grem1 expression initiates colonic tumorigenesis from cells outside the stem cell niche.” nature medicine, u.s. national library of medicine, jan. 2015, https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4594755/. [4] stzepourginski, igor, et al. “cd34+ mesenchymal cells are a major component of the intestinal stem cells niche at homeostasis and after injury.” pnas, national academy of sciences, 24 jan. 2017, https://www.pnas.org/content/114/4/e506. [5] perekatt, ansu o., et al. “smad4 suppresses wnt-driven dedifferentiation and oncogenesis in the differentiated gut epithelium.” cancer research, american association for cancer research, 1 sept. 2018, https://cancerres.aacrjournals.org/content/78/17/4878#. [6] el marjou, fatima. “tissue-specific and inducible cre-mediated recombination in the gut epithelium.” genesis (new york, n.y.: 2000), u.s. national library of medicine, july 2004, https://pubmed.ncbi.nlm.nih.gov/15282745/. [7] qi, zhen, et al. “bmp restricts stemness of intestinal lgr5+ stem cells by directly suppressing their signature genes.” nature news, nature publishing group, 6 jan. 2017, https://www.nature.com/articles/ncomms13824. [8] zhang, ya, and xin wang. “targeting the wnt/βcatenin signaling pathway in cancer journal of hematology & oncology.” biomed central, biomed central, 4 dec. 2020, https://jhoonline.biomedcentral.com/articles/10.1186/s13045-020-009903. [9] wang, li, et al. “areg mediates the epithelialmesenchymal transition in pancreatic cancer cells via the egfr/erk/nf-κb signalling pathway.” oncology reports, d.a. spandidos, may 2020, https://www.ncbi.nlm.nih.gov/pmc/articles/pmc7107775/. [10] fevr, tea, et al. “wnt/beta-catenin is essential for intestinal homeostasis and maintenance of intestinal stem cells.” molecular and cellular biology, american society for microbiology (asm), nov. 2007, https://www.ncbi.nlm.nih.gov/pmc/articles/pmc2169070/. [11] tissue‐specific and inducible cre‐mediated ... wiley online library. https://doi.org/10.1002/gene.20042. [12] harada n;tamai y;ishikawa t;sauer b;takaku k;oshima m;taketo mm; “intestinal polyposis in mice with a dominant stable mutation of the beta-catenin gene.” the embo journal, u.s. national library of medicine, https://pubmed.ncbi.nlm.nih.gov/10545105/. https://www.sciencedirect.com/science/article/pii/s0092867412005867 https://www.sciencedirect.com/science/article/pii/s0092867412005867 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4594755/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4594755/ https://www.pnas.org/content/114/4/e506 https://cancerres.aacrjournals.org/content/78/17/4878 https://cancerres.aacrjournals.org/content/78/17/4878 https://pubmed.ncbi.nlm.nih.gov/15282745/ https://pubmed.ncbi.nlm.nih.gov/15282745/ https://www.nature.com/articles/ncomms13824 https://www.nature.com/articles/ncomms13824 https://jhoonline.biomedcentral.com/articles/10.1186/s13045-020-00990-3 https://jhoonline.biomedcentral.com/articles/10.1186/s13045-020-00990-3 https://jhoonline.biomedcentral.com/articles/10.1186/s13045-020-00990-3 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc7107775/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc7107775/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc2169070/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc2169070/ https://doi.org/10.1002/gene.20042 https://pubmed.ncbi.nl/ https://pubmed.ncbi.nlm.nih.gov/10545105/ aresty rutgers undergraduate research journal, volume i, issue iv jay patel is a junior undergraduate student conducting research in the genetics department at rutgers new brunswick. his interest in scientific research was sparked by his experience with the rutgers waksman program, during his time in high school. he currently works with dr. verzi and his mentor, oscar, where they investigate the pathway by which the intestinal epithelium and mesenchymal cell populations are communicating. along with being a researcher, jay also enjoys riding as an emt and the highland park first aid squad, volunteering as a part of the rutgers red cross, and playing basketball with his friends. aresty rutgers undergraduate research journal, volume i, issue iv differential gene expression analysis and gene ontology in triploid and diploid pocillopora acuta deeksha misri, erin e. chille, timothy g. stephens debashish bhattacharya (faculty advisor) ✵ abstract corals are marine invertebrates that are facing life-threatening environmental stressors due to climate change. polyploidy can, in such cases, be an important source of variation and adaptation in corals and other species. polyploidy is the genomic condition wherein the cells of a normally diploid organism have more than one pair of chromosomes. pocillopora acuta, also known as the cauliflower coral, is a brooding coral that can also reproduce asexually. it is a stress-sensitive coral, which means it shows clear physiological changes in response to environmental stressors like temperature, salinity, and ph. in this study, about 60% of the stony coral pocillopora acuta samples collected from kāneʻohe bay, oahu, hi, were triploid. the aim of this study was to identify the differences in gene expression patterns between triploid cluster 1 (t1), triploid cluster 2 (t2), and diploid samples (d) of p. acuta. pairwise comparisons were carried out between all categories: t1 vs. d, t2 vs. d, and t1 vs. t2. while there were a large number of genes exhibiting similar expression patterns in both triploid clusters, many genes were differentially regulated in t1 when compared to t2. this result provides evidence suggesting that the two triploid lineages originated from separate triploidization events in kāneʻohe bay. the differentially expressed genes shared between these two triploid lineages, when compared to the diploid coral lineage, suggests changes in cellular physiology as a result of polyploidization. functional analysis of the p. acuta genes can provide deeper insight into the specific, differentially regulated molecular functions and biological processes in triploids when compared to diploid p. acuta. future studies involving comparative functional enrichment analysis with more triploid and diploid samples of p. acuta will provide more insight into events that caused triploidization and the coral’s response to environmental stressors. key terms: corals, coral bleaching, polyploidy, bioinformatics 1 introduction polyploidy is the genomic condition wherein an organism possesses multiple chromosome copies. this heritable condition can be a result of genome duplication within a species (autopolyploidy) or from hybridization of two different species (allopolyploidy). polyploidy in an organism can cause a drastic change in cell organization and genome structure as well as problems in gene expression, genome stability, cell physiology, and cell cycle processes (wertheim et al., 2021). thus, it can result in physiological changes in the organism in response to various kinds of environmental stressors like salinity stress, thermal stress, and pco2 stress. polyploidy is a relatively common occurrence in plants, and until recently, was believed to be a rare occurrence in animals (wertheim et al., 2021). however, in recent years, polyploidy has been found in all major animal phyla and occurs relatively frequently in some groups, such as fish and amphibians (wertheim et al., 2021). in some animals, triploidy may be beneficial to the organism with respect to improved growth, pathogen resistance, and creation of more fit genotypes (kang et al., 2004). coral reefs are large marine ecosystems formed by colonies of coral polyps inhabited by a diverse microbiome, including algal endosymbionts (symbiodiniaceae), bacteria, fungi, archaea, and viruses. due to climate change, coral reefs are subjected to acutely high temperatures that disrupt the coral-algal symbiosis leading to coral bleaching aresty rutgers undergraduate research journal, volume i, issue iv (williams et al., 2021). when corals bleach, they expel their symbionts, which provides most of thecarbohydrates, lipids, amino acids, and o2 that fulfill the host’s energy requirements. extended periods of bleaching, which cause compromised immunity and starvation, are a leading cause of the mass mortality of coral reefs globally (williams et al., 2021). corals need effective and efficient mechanisms to adapt and ensure their survival against environmental stressors such as prolonged heat and oxidative stress. hawai’i houses about 80 species of scleractinia corals, and the islands are the world’s most isolated archipelago, geographically isolated from other reef systems. this reef system thus provides an ideal platform for advancing coral biology and conservation using multi-omics and genetic tools. (bhattacharya et al., 2022). to study the stress response of the stony coral pocillopora acuta (also known as the cauliflower coral), 30 samples of the species were collected in 2018 from kāneʻohe bay in hawai’i and exposed to ambient temperature and ambient co2. from these samples, about 60% of the samples were triploid. while the origin of triploidy in this population is still unknown, there are currently three mechanistic explanations for the origin of triploidy within this population, including 1) self-fertilization of a p. acuta egg followed by fertilization by a foreign sperm, 2) diploidization of one of the two parental gametes, and 3) failure of the ovum to expel the second polar body (stephens et al., 2021). it is also hypothesized that triploid genotypes may have spread throughout kāneʻohe bay by the generation of asexual brooded larvae, (nakajima et al., 2018), and clonal propagation through colony fragmentation (highsmith, 1982). because this is the first known incidence of widespread triploidy in a coral population, we aimed to gain baseline information about gene expression regulation and differential gene expression between diploid and triploid p. acuta corals in the kāneʻohe bay ecosystem (stephens et al., 2021). in the study conducted by stephens et al., an unrooted phylogenetic tree built using the snps identified in p. acuta rna-seq data showed a putative split between diploid and triploid lineages. there were two triploid lineages identified, named triploid cluster 1 (t1) and triploid cluster 2 (t2). we studied the differences in gene expression between the two triploid clades (t1 and t2) and the diploid clade (d) p. acuta. gene expression in an organism can be measured by examining the mrna, the protein made by the mrna transcripts. by exploring the differences in gene expression between the diploid and triploid colonies we hope to discover if, and in what way, polyploidy provides a selective advantage. in light of the mass mortality of reef-building corals under climate change, it is essential to generate this baseline information on how triploidy affects gene expression and regulation to develop an understanding of how triploid coral populations may respond to this burgeoning stressor. 2 methodology coral sampling and collection corals were collected from six reefs in kāneʻohe bay, a 45 km² sheltered water body in oʻahu, hawaiʻi in september 2018. 30 of the 119 collected samples were cultured under ambient temperature ambient co2 (atac) for a 4-month period. molecular samples for rna-seq and bisulfite sequencing were obtained on days 1 and 2, and on weeks 1, 2, 4, 6, 8, 12, and 16. rna extraction and sequencing, read trimming, and gene count estimation were conducted as described in stephens et al. (2021). rna was extracted from a small clipping using the zymo quick-dna/rna™ miniprep plus kit (zymo research, irvine, ca, usa) following the manufacturer’s protocol for tissue samples. rna library preparation and sequencing was then performed at genewiz (south plainfield, new jersey, usa) following the truseq stranded mrna sample preparation protocol (illumina) with poly-a enrichment and hiseq sequencing workflow targeting 15 million reads per sample. quality trimming, adaptor removal, and gene count estimation were performed, and the resulting gene count matrix was used for differential expression analysis as described below. differential gene expression using r version 4.0.5 and the edger package v. 3.15 (robinson et al., 2010), pairwise aresty rutgers undergraduate research journal, volume i, issue iv comparisons of differential gene expression were conducted on the 27,254 genes that passed the low coverage filter of gene counts less than 10 (robinson et al. 2010). three pairwise gene expression comparisons were conducted: t1 vs. d, t2 vs. d and t1 vs. t2, using deseq2 (v. 3.14) (love et al. 2014). the detected differentially expressed genes were filtered based on adjusted p-values less than 0.05 and fold changes (fc) greater than 1 for upregulated genes and less than -1 for downregulated genes. the upset r package was used to represent the six sets of differentially expressed genes (degs) (upregulated and downregulated genes in t1, t2, and d samples) (conway et al. 2017). this plot represents nonempty and overlapping genes that were differentially regulated between the pairwise comparisons. gene ontology gene ontology (go) terms are a hierarchy of terms/descriptions that uses a controlled vocabulary to describe the known or inferred properties of a gene/protein. in this study, gene ontology and annotation were conducted following the go pipeline outlined in chille et al. (2021) to annotate genes by relevant protein accession ids for triploid and diploid samples of p. acuta. briefly, diamond (v. 0.0.14) was used to align protein and translated dna sequences against the refseq non-redundant protein sequence database (nr database) compiled by the ncbi (buchfink et al. 2015). to obtain gene ontology and annotation information from multiple databases, interproscan (v. 5.53-87.0) was used to search the interpro database to compile additional information about p. acuta protein sequences (jones et al. 2014). blast2go is a powerful bioinformatics platform for gene annotation and ontology. using the previously generated diamond results as input, the mapping and annotation tools on blast2go were used to map and annotate the p. acuta gene sequences. out of the 38,532 gene sequences, 5073 were mapped. then, interproscan go results were merged with the mapped and annotated sequences from blast2go. the accession ids for the gene sequences (generated by blast2go after mapping and annotation) were then run against the uniprot database. 3 results rna sequencing, quality control and clustering of samples for all 30 samples collected at atac, 38,532 genes were predicted as described in stephens et al. (2021). pre-filtering to retain genes with cpm (counts per million) over 3.33 in at least 2 samples resulted in 27,252 genes being selected for analysis. out of the 30 p. acuta samples, 10 were diploid (d), 9 belonged to triploid cluster 1 (t1), and the remaining 11 were triploid cluster 2 (t2). pairwise differential gene expression using r for the first pairwise comparison (t1 vs. d), 1354 genes were upregulated in t1, 1,090 were downregulated in t1, and there was no significant difference in expression for the remaining 24,810 genes. for the second pairwise comparison (t2 vs. d), 1,157 genes were upregulated in t2, 1,330 were downregulated in t2, and there was no significant difference in expression for the remaining 24,767 genes. for the third and last pairwise comparison between the two triploid clusters (t1 vs. t2), 1,527 genes were upregulated in t1, 985 were downregulated in t1, and there was no significant difference in expression for the remaining 24,742 genes. this information is represented in table 1. table 1: numbers of differentially expressed genes in pairwise comparisons using edger, with adjusted p-value < 0.05, and fold change (fc) > 1 for upregulated genes and fc <-1 for downregulated genes. an upset plot is a graph used to represent intersections of two or more categories that have some shared similarity (conway et al. 2017). for this study, an upset plot was used to visualize the degs similarly regulated across pairwise comparisons. pairwise comparison upregulated (log(fc)>1) downregulated (log(fc)<1) t1 vs. d 1354 1090 t2 vs. d 1157 1330 t1 vs. t2 1527 985 aresty rutgers undergraduate research journal, volume i, issue iv figure 1: upset plot of differentially expressed genes in pairwise comparisons between t1, t2, and d samples. the violet histogram on the bottom left of the figure represents the number of degs in each category of the pairwise comparisons. the black histogram represents the number of degs common between the highlighted pairwise comparisons (exact count shown on top of each black bar). for the black bars with a single dot across all sets, there is no overlap of differentially expressed genes with other categories. 954 out of the 1,330 genes downregulated in t2 (versus d) were also upregulated in t1 (versus t2). the second largest set of intersecting genes consisted of 856 genes that were upregulated in both triploid clusters against diploid samples. in contrast, only 439 genes were downregulated in both triploid clusters against diploid samples. other intersecting sets of genes are shown in the upset plot above (figure1). 4 discussion/conclusion differential gene expression analysis the aim of this study was to analyze the differential expression of genes in t1, t2, and d and to identify the physiological differences conferred by triploidy. deseq2 was used to find the differentially expressed genes and diamond, interproscan, and blast2go were used for gene ontology and annotaaresty rutgers undergraduate research journal, volume i, issue iv tion of the degs. while there were many genes exhibiting similar expression patterns for the two triploid clusters, there were also many differences. the number of differentially regulated genes for each triploid cluster is different when compared to each other and to diploid samples. the downregulated genes in t1 and t2 with log(fc) < -1 and adjusted pvalue < 0.05 exhibit interesting differences in gene expression. although both t1 and t2 clusters are triploid, the upset plot (figure 1) indicates that the genes differentially expressed in both clusters show differences in gene regulation. 954 genes downregulated in t2 in comparison to diploid samples are upregulated in t1 compared to t2 and are not differentially expressed in t1 compared to d. this indicates that these 954 genes may be downregulated only in t2 and not t1. similarly, there are 419 degs that are only downregulated in t1 and 414 degs downregulated in t2 that have no intersections with the other triploid cluster across all comparisons. however, there are 439 degs that are downregulated in both t1 and t2 against d. these genes depict no significant differential expression in the pairwise comparison between t1 and t2. hence, while there are a significant number of genes that are exclusively downregulated within each triploid cluster, some are common between the two clusters as well. this indicates that although both t1 and t2 are triploid, there are a large number of differences and similarities in gene expression patterns. this analysis is supported by the differential expression patterns for upregulated genes as well. notably, 331 genes upregulated in t2 against d are downregulated in t1 against t2. these genes may be only upregulated in t2. there are 402 upregulated genes in t2 that are not differentially expressed in t1 and 399 genes upregulated in t1 but not in t2. 595 degs are upregulated in t1 against both t2 and d but show no differential expression in t2. however, 858 genes are upregulated in both t1 and t2 against d and show no differential expression in the pairwise comparison between t1 and t2. again, since there are many upregulated genes showing differences and similarities in expression patterns in both t1 and t2, it is possible that after independent triploidization events, t1 and t2 adapted in somewhat similar ways to the same environmental conditions. while both triploid clusters have many degs regulated in the same way, there is an equally large number of degs expressed in completely different ways even between the two clusters. this may indicate that there were two separate sources of triploidization in the p. acuta reef in kāneʻohe bay, but adaptation of similar forms of triploidization in the coral can cause some genes in both clusters to be differentially expressed in similar ways. in future studies, we will compile gene ontology data obtained from diamond, interproscan, and blast2go and perform functional enrichment analysis will be performed on all 119 samples collected in the study. this will be done in order to identify the biological processes and molecular functions that are differentially regulated in triploid p. acuta across all stress conditions as described in the methodology section (atac, athc, htac, and hthc). this analysis will shed more light on the physiological and biochemical differences between triploid and diploid p. acuta. comparative functional enrichment analysis will provide insight into pathways impacted by differential expression of some genes in triploid lineages in p. acuta under environmental stress. it will also shed light on how triploidization can help the stress-sensitive p. acuta adapt to changing environmental conditions to ensure survival in the face of rampant climate change in the decades to come. 5 acknowledgements i would like to thank my mentor, erin chille, for her guidance and support. i would also like to thank dr. timothy stephens for detecting triploidy in the samples and letting us use his findings and data as the basis for our study. lastly, i would like to thank dr. debashish bhattacharya for his continued scientific guidance. 6 references [1] andrews, s. (2010). fastqc: a quality control tool for high throughput sequence data. aresty rutgers undergraduate research journal, volume i, issue iv [2] bhattacharya, d., stephens, t. g., tinoco, a. i., richmond, r. h., & cleves, p.a. (2022). life on the edge: hawaiian model for coral evolution. limnology and oceanography. [3] buchfink, b., xie, c., & huson, d. h. (2015). fast and sensitive protein alignment using diamond. nature methods, 12(1), 59-60. [4] chille, e. e., strand, e., neder, m., schmidt, v., sherman, m., putnam, h. m., & mass, t. (2021). developmental series of gene expression clarifies maternal mrna provisioning and maternal-to-zygotic transition in the reef-building coral montipora capitata. biorxiv. [5] conway, j. r., lex, a., & gehlenborg, n. (2017). upsetr: an r package for the visualization of intersecting sets and their properties. bioinformatics. [6] highsmith, r. c. (1982). reproduction by fragmentation in corals. marine ecology progress series. oldendorf, 7(2), 207-226. [7] jones, p., binns, d., chang, h. y., fraser, m., li, w., mcanulla, c., ... & hunter, s. (2014). interproscan 5: genome-scale protein function classification. bioinformatics, 30(9), 1236-1240. [8] kang, h. j., & rosenwaks, z. (2004). triploidy-the breakdown of monogamy between sperm and egg. international journal of developmental biology, 52(5-6), 449-454. [9] love, m. i., huber, w., & anders, s. (2014). moderated estimation of fold change and dispersion for rna-seq data with deseq2. genome biology, 15(12), 1-21. [10] nakajima, y., chuang, p. s., ueda, n., & mitarai, s. (2018). first evidence of asexual recruitment of pocillopora acuta in okinawa island using genotypic identification. peerj, 6, e5915. [11] robinson, m. d., mccarthy, d. j., & smyth, g. k. (2010). edger: a bioconductor package for differential expression analysis of digital gene expression data. bioinformatics, 26(1), 139-140. [12] stephens, t. g., strand, e. l., mohamed, a. r., williams, a., chiles, e. n., su, x., & putnam, h. m. (2021). ploidy variation and its implications for reproduction and population dynamics in two sympatric hawaiian coral species. biorxiv. [13] wertheim, b., beukeboom, l. w., & van de zande, l. (2013). polyploidy in animals: effects of gene expression on sex determination, evolution and ecology. cytogenetic and genome research, 140(2-4), 256269. [14] williams, a., chiles, e. n., conetta, d., pathmanathan, j. s., cleves, p. a., putnam, h. m., & bhattacharya, d. (2021). metabolomic shifts associated with heat stress in coral holobionts. science advances, 7(1), eabd4210. [15] willis, b. l., van oppen, m. j., miller, d. j., vollmer, s. v., & ayre, d. j. (2006). the role of hybridization in the evolution of reef corals. annu. rev. ecol. evol. syst., 37, 489-517. deeksha misri is a third-year student in the honors college at rutgers university new brunswick. she is majoring in genetics with a certificate in computational genetics and a minor in philosophy. over the past two years, she has been working as an undergraduate researcher at dr. debashish bhattacharya’s lab, which she first joined as an aresty summer science fellow during her freshman year. her research focuses on using bioinformatic and other computational tools to study coral genomics and multi-omics and the genetic and biomolecular mechanisms coral adopt to adapt to climate change. after graduating from rutgers, deeksha plans to continue performing research and pursue a phd in bioinformatics. outside of the lab, deeksha is also involved in the women in stem community on campus, and previously served as president of the stem ambassadors – women in stem club at rutgers. she is also an active member of the douglass residential college for women and serves as a research advisory board mentor for douglass. previously, deeksha has been involved with aresty as a summer science participant as well as a peer reviewer for rurj. for any questions, please contact deeksha at deeksha.misri@rutgers.edu. mailto:deeksha.misri@rutgers.edu aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. cystinuria: genetic aspects and novel pharmacotherapeutics diana stachula, amrik sahota (faculty advisor) ✵ abstract this review provides an overview of the genetic aspects of cystinuria, as well as the novel pharmacotherapeutics that could potentially be used to treat the disease. cystinuria is an inherited disorder characterized by the formation of painful stones in the kidneys, bladder, and other parts of the renal system. currently, mutations responsible for cystinuria have been identified in two genes (slc3a1 and slc7a9 ), and cystinuria patients are categorized based on their genotypes which versions, or alleles, of these genes they have (mutated or wildtype). regardless of genotype, however, current treatments for all cystinuria patients have significant limitations. this has led researchers to search for more promising therapeutics. one potential treatment uses cystine analogs—compounds that are structurally similar to cystine, which is the naturally occurring chemical substance from which the stones are formed. these compounds have demonstrated the ability to inhibit stone formation by stunting cystine crystallization – the process by which cystine crystals aggregate to form stones. gene therapy may also be used to treat cystinuria in the future by replacing mutated copies of slc3a1 and slc7a9 with healthy ones. technological advancements and an improvement of our understanding of how gene therapy functions in the renal system could reveal even more treatment possibilities. 1 introduction cystinuria generally arises from mutations in the slc3a1 and slc7a9 genes. there are likely more genetic factors that contribute to the disease that are yet to be identified, as 5% of cystinuria patients do not have mutations in either of the two genes. (sahota et al., 2019). slc3a1 and slc7a9 encode crucial components of the biochemical pathway responsible for the reabsorption of dibasic amino acids in the renal system. dibasic amino acids are organic compounds that form proteins and contain two basic functional groups, typically amino groups (nh2). one dibasic amino acid is cystine (figure 2), which is made of two cysteine molecules joined by a disulfide bond (s s) (figure 1). the defective reabsorption of cystine from the kidneys into the bloodstream causes its supersaturation in urine and the formation of cystine stones in the kidneys, bladder, and ureters (sahota et al., 2019). cystine stones are jagged in shape and are considered to be the hardest stones formed in the human renal system (ringdén & tiselius, 2007). most cystinuria patients that develop their first stone in adolescence are prone to recurrent stone formation throughout their lifetimes (rogers et al., 2007). in addition to abdominal pain, patients may also experience nausea, hematuria (blood in urine), recurrent urinary tract infections, and kidney failure (mattoo & goldfarb, 2008). increased fluid intake, reduced protein consumption, and the use of currently available medications have proven to be less-than-ideal treatment methods for the disease (sahota et al., 2019). potential novel treatments of cystinuria have been studied using slc3a1 and slc7a9 knockout mouse models, which are mice with mutated, nonfunctional versions of the slc3a1 and slc7a9 genes (sahota et al., 2019). cystine diesters, such as cystine dimethylester (cdme) (figure 3), and cystine diamides, figure 1: cysteine is an amino acid with a thiol side chain (r-sh). two cysteine molecules can be oxidized to form cystine (figure 2). aresty rutgers undergraduate research journal, volume i, issue iv figure 2: cystinuria patients form stones made of cystine, an organic molecule containing a disulfide bridge (s – s) and two amine groups (-nh2). figure 3: cdme is an example of a cystine diester, a type of cystine analog. like cystine, it contains a disulfide bridge (s – s) and two amine groups (-nh2). figure 4: l-cystine bis(n′-methylpiperazide) is an example of a cystine diamide, a type of cystine analog. like cystine, it contains a disulfide bridge (s – s) and two amine groups (-nh2). such as l-cystine bismorpholide and l-cystine bis(n′-methylpiperazide) (figure 4), all of which are analogs of cystine (figure 2), have demonstrated their effectiveness as potential treatments for cystinuria through their abilities to inhibit cystine crystal growth in these mouse models (yang et al., 2018). continued study of cystine stone formation inhibitors, as well as gene therapy, will likely generate promising new treatments for human cystinuria patients. 2 cystinuria: etiology and epidemiology transport defect genetic mutations in slc3a1 and slc7a9 cause the defective reabsorption of several dibasic amino acids — cystine, ornithine, lysine, and arginine (cola) — from the kidneys into the bloodstream (figure 5) (sahota et al., 2019). more specifically, these mutations disrupt the cola transporter (b0,+), which is a heterodimer, or a molecule made up of two protein components (sahota et al., 2019). slc3a1 and slc7a9 each encode one of these components (sumorok & goldfarb, 2013); slc3a1 encodes the rbat subunit, while slc7a9 encodes the b0,+ at subunit (figure 6) (sahota et al., 2019). mutations in either gene will cause a defect in the corresponding subunit, leading to the defective reabsorption of the cola amino acids (sumorok & goldfarb, 2013). since cystine is the least soluble of the cola amino acids, it has a greater ability to crystallize in the urinary tract and form stones when improperly reabsorbed (sahota et al., 2019). stone formation cystine stones are thought to form by free solution crystallization, the process by which supersaturated solutions transform into solids (coe et al., 2010). when cystine is supersaturated in urine, it crystallizes into stones that can be found freely throughout the renal system (coe et al., 2010), though they are predominantly found in the terminal collecting ducts within the kidneys (khan et al., 2016). these stones are named depending on their specific location (figure 7). their mobility within the renal system allows them to be easily removed during surgery; crystals of large size wash away when surgically exposed (coe et al., 2010). aresty rutgers undergraduate research journal, volume i, issue iv figure 5: mutations in the slc3a1 and slc7a9 genes cause the defective reabsorption of the cola amino acids from the proximal convoluted tubule into the bloodstream. these dibasic amino acids proceed through the rest of the renal system and are excreted in urine. created with biorender.com. figure 6: slc3a1 encodes the rbat subunit (green) and slc7a9 encodes the b0,+ at subunit (blue) of the cola transporter (b0,+), which is responsible for the reabsorption of the cola amino acids in the renal system. cystine is reduced to two cysteine molecules when it is reabsorbed into the bloodstream. epidemiology although cystine stones make up only approximately 1% of all kidney stones, cystinuria is still one of the most commonly inherited genetic disorders (mattoo & goldfarb, 2008). the disease has a global prevalence of approximately 1:7,000, ranging from 1:2,500 in libyan jews to 1:100,000 in swedes. in the united states, approximately 1 in 15,000 adults have cystinuria (mattoo & goldfarb, 2008). men are twice as likely as women to develop cystine stones (leslie, sajjad & nazzal, 2020). this may be due to shorter urethral length or factors that inhibit cystine crystal aggregation in females (sahota et al., 2019). patients typically first present a stone between the ages of 2 and 40, with a median onset age of 12 in males and 15 in females (rogers et al., 2007). approximately two thirds of cystinuria patients develop stones in both kidneys, while one aresty rutgers undergraduate research journal, volume i, issue iv third only form stones in a single kidney (usawachintachit et al., 2018). among patients who develop stones, over 60% experience recurrent stone formation, with males forming new stones about every 3 years and females forming new stones about every 5 years (dello strologo et al., 2002). in addition to higher recurrence rates, males also typically experience more aggressive disease symptoms that may require more surgical interventions (edvardsson et al., 2013). figure 7: cystine stones are found freely throughout the renal system. created with biorender.com. 3 genetics inheritance and genotypes cystinuria patients are classified depending on which of their genes are mutated. those with type a, or type i, cystinuria have a mutation in slc3a1 on chromosome 2. those with type b, or non-type i, cystinuria have a mutation in slc7a9 on chromosome 19 (fazaeli et al., 2017). every person has two copies of each gene. mutations in slc3a1 are inherited through an autosomal recessive pattern of inheritance (both copies of the gene must be mutated for disease presentation) (martell et al., 2017). meanwhile, mutations in slc7a9 follow an autosomal dominant pattern of inheritance with incomplete penetrance; typically (only one mutated copy of the gene needs to be present to allow for the formation of cystine stones) (martell et al., 2017). rarely, patients have type ab cystinuria; people who fall under this category have two mutated copies of one of the genes as well as one mutated copy of the other (sumorok & goldfarb, 2013). depending on which gene has two mutated copies and which has one mutated copy, patients can be designated as either type aab or type abb (sumorok & goldfarb, 2013). as aforementioned, both copies of slc3a1 must be mutated for disease presentation, so slc3a1 heterozygotes, who only have one mutated copy, should not present stones or any characteristics of cystinuria. slc7a9 heterozygotes, however, may present cystinuria symptoms such as variable urinary levels of cola (edvardsson et al., 2013). slc7a9 heterozygotes are unlikely to develop stones unless urine volumes are low or protein intake is significantly elevated (sahota et al., 2019). mutations over 400 total mutations have been identified in slc3a1 and slc7a9 (stenson et al., 2003), including missense, nonsense, splicing, regulatory, deletion, insertion, indel, duplication, and rearrangement mutations (stenson et al., 2003). each of these mutation types alters the dna sequences of slc3a1 and slc7a9, resulting in the formation of altered or truncated proteins (the subunits of the cola transporter). missense mutations are the largest group of mutations that result in cystinuria. such mutations change a single amino acid in the protein being encoded, which can have a range of effects on the protein — protein function may be unimpacted, impacted to some degree, or lost completely (martell et al., 2017). currently, the impact of missense mutations in slc3a1 and slc7a9 on protein function and disease presentation is unclear. (martell et al., 2017). table 1 presents the mutation type and number of mutations found in slc3a1. of the 261 mutations identified, data on 210 mutations has been made publicly available by the human gene mutation database (hgmd) from the institute of medical genetics in cardiff. table 2 presents the mutation type and number of mutations found in slc7a9. of the 170 mutations identified, data on 143 mutations has been made publicly available by hgmd (stenson et al., 2003). aresty rutgers undergraduate research journal, volume i, issue iv table 1: slc3a1 mutations listed in the hgmd database. table 2: slc7a9 mutations listed in the hgmd database. mutation type number of mutations missense/nonsense 128 splicing 13 regulatory 1 small deletions 19 small insertions 11 small indels (insertions + deletions) 2 gross deletions 30 gross insertions/duplications 5 complex rearrangements 1 repeat variations 0 public total (hgmd professional 2021.4 total) 210 (261) mutation type number of mutations missense/nonsense 75 splicing 18 regulatory 0 small deletions 29 small insertions 10 small indels (insertions + deletions) 1 gross deletions 9 gross insertions/duplications 1 complex rearrangements 0 repeat variations 0 public total (hgmd professional 2021.4 total) 143 (170) aresty rutgers undergraduate research journal, volume i, issue iv 4 mouse models knockout mouse models several mouse models have been generated to observe the traits associated with types a, b, and ab cystinuria (sahota et al., 2019). among these is a knockout slc3a1 mouse model, slc3a1-/-, in which both copies of the slc3a1 gene were mutated to become nonfunctional, or “knocked out” (sahota et al., 2019). urine analyses have revealed the presence of supersaturated cystine crystals in the slc3a1-/mice (figure 8). computed tomography (ct) scanning was also used to view the cystine stones found in these knockouts (figure 9). a slc7a9/knockout mouse model with deletion mutations in both copies of slc7a9 was also created (fontllitjós et al., 2007). both type a slc3a1-/and type b slc7a9-/mice presented higher urinary levels of cystine in comparison to wild-type (non-mutated) mice (beckermann et al., 2020; font-llitjós et al., 2007). a mouse model of type ab cystinuria (slc3a1+/−, slc7a9+/−) was generated by crossing type a and type b mice (sahota et al., 2019). these type ab mice also had cola hyperexcretion; however, they presented more severe stone formation than type a or type b mice (espino et al., 2015). figure 8: the hexagonal cystine crystals observed in an slc3a1-/mouse. image provided by amrik sahota, ph.d. figure 9: cystine stones in an slc3a1 knockout mouse (top), shown to scale (bottom). figure provided by amrik sahota, ph.d. gender differences males with cystinuria experience more aggressive disease symptoms than females, a characteristic reflected by slc3a1-/mice (sahota et al., 2019). knockout males and females presented cystine crystals of similar size and distribution; however, bladder stones only formed in a few female mice and with a later onset (>18 months) than their male counterparts (sahota et al., 2019). sex differences were not observed in the knockout slc7a9-/mice, as both males and females formed stones in a 1:1 ratio with an onset age of one month (feliubadaló et al., 2003). 5 current treatments and limitations individuals with cystinuria will experience recurrent cystine stone formation throughout their lifetimes, so behavioral management and pharmacological therapies are often necessary to increase quality of life (siener et al., 2021). treatment methods for cystinuria have remained largely unaltered for the past few decades. currently, most cystinuria patients are advised to increase their fluid intake and reduce their protein and sodium consumption https://orcid.org/0000-0002-3603-673x https://orcid.org/0000-0002-3603-673x https://orcid.org/0000-0002-3603-673x aresty rutgers undergraduate research journal, volume i, issue iv (siener et al., 2021). in addition to behavioral modifications, urinary alkalinization (increasing urine ph) is considered a primary treatment because cystine is more soluble at higher ph values (pearle et al. 2014). afflicted individuals may take potassium citrate to achieve a urine ph of 7.0-7.5 (the normal average urine ph is 6.0) (pearle et al. 2014). in more severe cases, patients may be prescribed thiol drugs, which contain a thiol functional group (-sh) that binds to cystine (pearle et al. 2014). the previously mentioned treatment methods all have limitations. many cystinuria patients have trouble adhering to behavioral modifications, especially young children who may find it difficult to consume large amounts of water (sahota et al. 2019). excess potassium citrate can lead to the formation of calcium phosphate stones (another type of kidney stone), and thiol drugs have several dosedependent adverse effects (pereira, schoolwerth & pais, 2015). such side effects include, but are not limited to, skin diseases, liver abnormalities, and blood disorders (deberardinis et al., 2008). therefore, there is a clear need for more tolerable, preventative treatment options. due to the limitations of current treatments, most cystinuria patients require multiple surgical interventions throughout their lifetimes. non-invasive stone-removing procedures include extracorporeal shockwave lithotripsy (eswl), which directs a shock wave at the stone (wood et al., 2011). however, cystine stones are somewhat resistant to eswl, so multiple rounds of treatment are necessary (wood et al., 2011). furthermore, only 37.5% of cystinuria patients remain stone-free for three months after undergoing eswl (landau et al., 2009). several concurrent eswl treatments increase the risk of kidney damage. renal injuries as the result of eswl include, but are not limited to, hemorrhages, rupturing of small veins and capillaries, necrosis (premature cell death), hematomas (severe bruises), and complete loss of kidney function (mcateer & evan, 2008). 6 cystine stone inhibitors cystine analogs cystine crystallization is a critical step in stone formation; therefore, potential treatments for cystinuria have been evaluated for their ability to inhibit cystine crystallization (yang et al., 2018). atomic force microscopy (afm), a high-resolution microscopy technique, was used to visualize growth on the surface of cystine crystals in the presence of 31 prospective crystal inhibitors (poloni et al., 2017). the data showed that the most effective inhibitors of cystine crystal growth were cystine analogs, also known as “molecular imposters.” cystine diesters and cystine diamides, two types of cystine analogs, demonstrated the greatest inhibitory effects on cystine crystal growth (poloni et al., 2017). in the presence of these inhibitors, cystine crystals were smaller and changed shape from hexagonal to tetragonal, making them more soluble (poloni et al., 2017). maintaining higher levels of cystine in solution is crucial to inhibiting cystine crystallization (hu et al., 2016). cystine diamides a series of cystine diamides were designed, synthesized, and then evaluated for their ability to inhibit cystine crystallization (yang et al., 2018). of the synthesized cystine diamides, l-cystine bismorpholide and l-cystine bis (n′-methylpiperazide) were the greatest crystallization inhibitors; they were 7 and 24 times more potent, respectively, as well as more stable than a previously studied cystine diester, l-cystine dimethylester (cdme) (yang et al., 2018). additionally, l-cystine bis (n′-methylpiperazide) has been able to successfully inhibit stone formation in an slc3a1 knockout mouse model, indicating that cystine diamides could potentially be used to prevent the formation of cystine stones in human cystinuria patients (yang et al., 2018). however, because the knockout mice form bladder stones rather than kidney stones (woodard et al., 2019), the direct application of these results to human patients may have some limitations. since cystine diamides have greater chemical stability than cystine diesters (e.g., cdme), they are likely more resistant to proteolytic degradation aresty rutgers undergraduate research journal, volume i, issue iv (the breakdown of peptide bonds in an amino acid) (hu et al., 2016). l-cystine bismorpholide and l-cystine bis (n′-methylpiperazide) are more promising treatments than cdme not only because of their increased chemical stability, but also because they are orally bioavailable (they can be easily ingested by mouth and absorbed by the body) (hu et al., 2016). while cdme has been effective in decreasing cystine stone size and mass, its efficacy post oral administration may be reduced due to esterase-mediated hydrolysis, a process causing the degradation of diesters (hu et al., 2016). furthermore, afm has revealed that cystine diamides are better than cdme and other cystine diesters at maintaining higher levels of cystine in solution, an important factor in preventing cystine crystals from aggregating into stones. (hu et al., 2016). 7 future direction as our knowledge and understanding of the pathophysiology of cystinuria expands, new therapies and treatments will continue to emerge. advancements in imaging technology and its interpretation will progress the current treatment management systems toward more effective methods. the application of afm in identifying crystal growth inhibitors and the continued use of mouse models will provide greater insight into these alternative therapies (pereira et al., 2015). aside from cystine analogs, gene therapy appears to be a promising treatment for cystinuria as well. gene therapy is a disease treatment technique in which a diseased gene copy is replaced with a healthy gene copy in a living organism. crispr/cas9 precision gene editing was recently utilized to create a slc7a9-/knockout mouse model of cystinuria (bai et al., 2019). research groups are attempting to use gene therapy to repair the slc7a9 deletions in these knockouts. however, multiple obstacles have presented themselves, including immune responses against vectors, which are organisms, usually bacteria, that deliver foreign dna to recipient cells (bai et al., 2019). the location and anatomy of the kidney could be causing difficulties in vector delivery (bai et al., 2019). more must be learned about applying gene therapy technologies to the renal system to proceed (bai et al., 2019). if research efforts are successful, gene therapy could become an ideal, one-time treatment for cystinuria as it has been for other rare genetic disorders (e.g., spinal muscular atrophy) (mendell et al., 2017). even though there is great diversity in the mutations that can lead to cystinuria (stenson et al., 2003), gene therapy is versatile in the mutations it can correct with a healthy gene copy (luther et al., 2018). 8 conclusions cystinuria is a genetic disorder that causes the formation of cystine stones in the renal system as a result of mutations in the slc3a1 and slc7a9 genes. distributions of the disease vary by population, although males are more likely than females to have severe disease presentation. because cystine stones have a high recurrence rate, cystinuria patients frequently require several surgical interventions throughout their lifetimes. current treatments, such as increased fluid intake, urine alkalinization, and thiol drugs, are aimed at delaying, but not necessarily eliminating, the need for surgical interventions. this renders them nonoptimal, especially because they may cause severe side effects. cystine analogs, particularly cystine diesters (e.g. cdme) and cystine diamides (e.g., l-cystine bismorpholide and l-cystine bis (n′-methylpiperazide)), have demonstrated their ability to effectively inhibit cystine crystal growth and, in some cases, stone formation. this qualifies them as potential alternatives to the cystinuria treatments currently in use. gene therapy has also been considered as a potential treatment; however, not enough information about its use in the renal system is presently known. with continued research, a new treatment that will improve the quality of life of human cystinuria patients could be made available in the near future∎ 9 acknowledgements amrik sahota, ph.d. (human genetics institute of new jersey) provided the images used in figures 6 and 7 of this literature review. i’d like to thank him for his contribution and his guidance as a research mentor. aresty rutgers undergraduate research journal, volume i, issue iv 10 references [1] bai, y., tang, y., han, p. et al. gene therapy for cystinuria. urolithiasis 47, 309–310 (2019). https://doi.org/10.1007/s00240-019-01111-7 [2] beckermann, t. m., welch, r. c., williams, f. m., mortlock, d. p., sha, f., ikizler, t. a., woodard, l. e., & wilson, m. h. (2020). crispr/cas9 engineering of albino cystinuria type a mice. genesis (new york, n.y.: 2000), 58(5), e23357. https://doi.org/10.1002/dvg.23357 [3] coe, f. l., evan, a. p., worcester, e. m., & lingeman, j. e. (2010). three pathways for human kidney stone formation. urological research, 38(3), 147–160. https://doi.org/10.1007/s00240-010-0271-8 [4] deberardinis, r. j., coughlin, c. r., 2nd, & kaplan, p. (2008). penicillamine therapy for pediatric cystinuria: experience from a cohort of american children. the journal of urology, 180(6), 2620–2623. https://doi.org/10.1016/j.juro.2008.08.057 [5] dello strologo, l., pras, e., pontesilli, c., beccia, e., ricci-barbini, v., de sanctis, l., ponzone, a., gallucci, m., bisceglia, l., zelante, l., jimenez-vidal, m., font, m., zorzano, a., rousaud, f., nunes, v., gasparini, p., palacín, m., & rizzoni, g. (2002). comparison between slc3a1 and slc7a9 cystinuria patients and carriers: a need for a new classification. journal of the american society of nephrology: jasn, 13(10), 2547– 2553. https://doi.org/10.1097/01.asn.0000029586.17680.e5 [6] edvardsson, v. o., goldfarb, d. s., lieske, j. c., beara-lasic, l., anglani, f., milliner, d. s., & palsson, r. (2013). hereditary causes of kidney stones and chronic kidney disease. pediatric nephrology (berlin, germany), 28(10), 1923–1942. https://doi.org/10.1007/s00467-012-2329-z [7] espino, m., font-llitjós, m., vilches, c., salido, e., prat, e., lópez de heredia, m., palacín, m., & nunes, v. (2015). digenic inheritance in cystinuria mouse model. plos one, 10(9), e0137277. https://doi.org/10.1371/journal.pone.0137277 [8] fazaeli, s., ashouri, s., kheirollahi, m., mohammadi, m., & fazilati, m. (2017). a novel mutation in slc7a9 gene in cystinuria. iranian journal of kidney diseases, 11(2), 138–141. [9] feliubadaló, l., arbonés, m. l., mañas, s., chillarón, j., visa, j., rodés, m., rousaud, f., zorzano, a., palacín, m., & nunes, v. (2003). slc7a9-deficient mice develop cystinuria non-i and cystine urolithiasis. human molecular genetics, 12(17), 2097–2108. https://doi.org/10.1093/hmg/ddg228 [10] font-llitjós, m., feliubadaló, l., espino, m., clèries, r., mañas, s., frey, i. m., puertas, s., colell, g., palomo, s., aranda, j., visa, j., palacín, m., & nunes, v. (2007). slc7a9 knockout mouse is a good cystinuria model for antilithiasic pharmacological studies. american journal of physiology. renal physiology, 293(3), f732–f740. https://doi.org/10.1152/ajprenal.00121.2007 [11] hu, l., yang, y., aloysius, h., albanyan, h., yang, m., liang, j. j., yu, a., shtukenberg, a., poloni, l. n., kholodovych, v., tischfield, j. a., goldfarb, d. s., ward, m. d., & sahota, a. (2016). l-cystine diamides as lcystine crystallization inhibitors for cystinuria. journal of medicinal chemistry, 59(15), 7293–7298. https://doi.org/10.1021/acs.jmedchem.6b00647 [12] khan, s. r., pearle, m. s., robertson, w. g., gambaro, g., canales, b. k., doizi, s., traxer, o., & tiselius, h. g. (2016). kidney stones. nature reviews. disease primers, 2, 16008. https://doi.org/10.1038/nrdp.2016.8 [13] mattoo a, goldfarb ds. (2008) cystinuria. semin nephrol. 28(2):181-91. doi: 10.1016/j.semnephrol.2008.01.011. pmid: 18359399. [14] landau, e. h., shenfeld, o. z., pode, d., shapiro, a., meretyk, s., katz, g., katz, r., duvdevani, m., hardak, b., cipele, h., hidas, g., yutkin, v., & gofrit, o. n. (2009). extracorporeal shock wave lithotripsy in prepubertal children: 22-year experience at a single institution with a single lithotriptor. the journal of urology, 182(4 suppl), 1835–1839. https://doi.org/10.1016/j.juro.2009.04.084 [15] leslie, s. w., sajjad, h., & nazzal, l. (2020). cystinuria. in statpearls. statpearls publishing. [16] luther, d. c., lee, y. w., nagaraj, h., scaletti, f., & rotello, v. m. (2018). delivery approaches for crispr/cas9 therapeutics in vivo: advances and challenges. expert opinion on drug delivery, 15(9), 905– 913. https://doi.org/10.1080/17425247.2018.1517746 [17] martell, h. j., wong, k. a., martin, j. f., kassam, z., thomas, k., & wass, m. n. (2017). associating mutations causing cystinuria with disease severity with the aim of providing precision medicine. bmc genomics, 18(suppl 5), 550. https://doi.org/10.1186/s12864-017-3913-1 [18] mendell, j. r., al-zaidy, s., shell, r., arnold, w. d., rodino-klapac, l. r., prior, t. w., lowes, l., alfano, l., berry, k., church, k., kissel, j. t., nagendran, s., l'italien, j., sproule, d. m., wells, c., cardenas, j. a., heitzer, m. d., kaspar, a., corcoran, s., braun, l., … kaspar, b. k. (2017). single-dose gene-replacement therapy for spinal muscular atrophy. the new england journal of medicine, 377(18), 1713–1722. https://doi.org/10.1056/nejmoa1706198 https://doi.org/10.1007/s00240-019-01111-7 https://doi.org/10.1002/dvg.23357 https://doi.org/10.1007/s00240-010-0271-8 https://doi.org/10.1016/j.juro.2008.08.057 https://doi.org/10.1097/01.asn.0000029586.17680.e5 https://doi.org/10.1007/s00467-012-2329-z https://doi.org/10.1371/journal.pone.0137277 https://doi.org/10.1093/hmg/ddg228 https://doi.org/10.1152/ajprenal.00121.2007 https://doi.org/10.1021/acs.jmedchem.6b00647 https://doi.org/10.1038/nrdp.2016.8 https://doi.org/10.1016/j.juro.2009.04.084 https://doi.org/10.1080/17425247.2018.1517746 https://doi.org/10.1186/s12864-017-3913-1 https://doi.org/10.1056/nejmoa1706198 aresty rutgers undergraduate research journal, volume i, issue iv [19] mcateer, j. a., & evan, a. p. (2008). the acute and long-term adverse effects of shock wave lithotripsy. seminars in nephrology, 28(2), 200–213. https://doi.org/10.1016/j.semnephrol.2008.01.003 [20] pearle, m. s., goldfarb, d. s., assimos, d. g., curhan, g., denu-ciocca, c. j., matlaga, b. r., monga, m., penniston, k. l., preminger, g. m., turk, t. m., white, j. r., & american urological assocation (2014). medical management of kidney stones: aua guideline. the journal of urology, 192(2), 316–324. https://doi.org/10.1016/j.juro.2014.05.006 [21] pereira, d. j. c., schoolwerth, a. c. & pais. v. m. (2015). cystinuria: current concepts and future directions. clinical nephrology. 83, 138-146. doi: 10.5414/cn108514 [22] poloni, l. n., zhu, z., garcia-vázquez, n., yu, a. c., connors, d. m., hu, l., sahota, a., ward, m. d., & shtukenberg, a. g. (2017). role of molecular recognition in l-cystine crystal growth inhibition. crystal growth & design, 17(5), 2767–2781. https://doi.org/10.1021/acs.cgd.7b00236 [23] ringdén, i., & tiselius, h. g. (2007). composition and clinically determined hardness of urinary tract stones. scandinavian journal of urology and nephrology, 41(4), 316–323. https://doi.org/10.1080/00365590601154551 [24] rogers, a., kalakish, s., desai, r. a., & assimos, d. g. (2007). management of cystinuria. the urologic clinics of north america, 34(3), 347–362. https://doi.org/10.1016/j.ucl.2007.04.006 [25] sahota, a., tischfield, j. a., goldfarb, d. s., ward, m. d., & hu, l. (2019). cystinuria: genetic aspects, mouse models, and a new approach to therapy. urolithiasis, 47(1), 57–66. https://doi.org/10.1007/s00240-018-1101-7 [26] siener, r., bitterlich, n., birwé, h., & hesse, a. (2021). the impact of diet on urinary risk factors for cystine stone formation. nutrients, 13(2), 528. https://doi.org/10.3390/nu13020528 [27] stenson et al. (2003), the human gene mutation database (hgmd®) [28] sumorok, n., & goldfarb, d. s. (2013). update on cystinuria. current opinion in nephrology and hypertension, 22(4), 427–431. https://doi.org/10.1097/mnh.0b013e3283621c5d [29] usawachintachit, m., sherer, b., hudnall, m., tzou, d. t., taguchi, k., hsi, r. s., stoller, m., & chi, t. (2018). clinical outcomes for cystinuria patients with unilateral versus bilateral cystine stone disease. journal of endourology, 32(2), 148–153. https://doi.org/10.1089/end.2017.0335 [30] wood, k., keys, t., mufarrij, p., & assimos, d. g. (2011). impact of stone removal on renal function: a review. reviews in urology, 13(2), 73–89. [31] woodard, l. e., welch, r. c., veach, r. a., beckermann, t. m., sha, f., weinman, e. j., ikizler, t. a., tischfield, j. a., sahota, a., & wilson, m. h. (2019). metabolic consequences of cystinuria. bmc nephrology, 20(1), 227. https://doi.org/10.1186/s12882-019-1417-8 [32] yang, y., albanyan, h., lee, s., aloysius, h., liang, j. j., kholodovych, v., sahota, a., & hu, l. (2018). design, synthesis, and evaluation of l-cystine diamides as l-cystine crystallization inhibitors for cystinuria. bioorganic & medicinal chemistry letters, 28(8), 1303– 1308. https://doi.org/10.1016/j.bmcl.2018.03.024 https://doi.org/10.1016/j.semnephrol.2008.01.003 https://doi.org/10.1016/j.juro.2014.05.006 https://doi.org/10.1021/acs.cgd.7b00236 https://doi.org/10.1080/00365590601154551 https://doi.org/10.1016/j.ucl.2007.04.006 https://doi.org/10.1007/s00240-018-1101-7 https://doi.org/10.3390/nu13020528 https://doi.org/10.1097/mnh.0b013e3283621c5d https://doi.org/10.1089/end.2017.0335 https://doi.org/10.1186/s12882-019-1417-8 https://doi.org/10.1016/j.bmcl.2018.03.024 aresty rutgers undergraduate research journal, volume i, issue iv diana stachula is a rutgers presidential scholar who graduated summa cum laude from the honors college at rutgers university new brunswick in may 2022, having completed a b.a. in genetics and a minor in psychology. from 2020-2022, diana worked as a research assistant for dr. amrik sahota at the human genetics institute of new jersey (hginj), where she studied cystinuria a rare genetic disorder that causes the formation of kidney stones. more specifically, she researched novel pharmacological agents as potential crystal growth and stone formation inhibitors. as a lab member, diana performed gel electrophoresis, polymerase chain reactions (pcr), quantitative pcr (qpcr), nanopore sequencing, rna/dna extractions from tissue, and various other genetic techniques. she has presented her research at several research symposiums, including the university-wide aresty undergraduate research symposium. diana also worked as a peer instructor at the aresty research center, mentoring new research assistants on professional development and communicating their research findings. currently, diana is working as an ophthalmic technician and medical assistant in order to gain more experience prior to applying to medical school. as a future physician, diana hopes that she can blend her interests in science and medicine by performing clinical research. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. rilke in translation: uncovering the panther in the english language serena luckhoff ✵ abstract walter benjamin’s influential essay, the task of the translator, reflects on how translation can be used to create new works of art and literature instead of attempting to create exact replicas of original works. rather than translating by searching for words in the target language that are most equivalent to words in the original language, benjamin encourages the practice of translation as a process which lets two languages influence and change one another. using benjamin as a guide, i attempt to create an english translation of rainer maria rilke’s german language poem “der panther” that distributes the meaning of the original work throughout the entire translation. to do this, i shift my focus away from translating words and phrases in isolation to reflect on how the effects of seemingly small translation choices reverberate through the work as a whole. i will also compare my translation to the historically important translations of “der panther” by stephen mitchell, robert bly, and c. f. macintyre, and argue how mine allows for analyses of themes such as stillness, the panther’s mental state, and the panther’s mind-body relation which more closely emulate rilke’s original work. key terms: translation, rilke, benjamin, kunst-ding, german 1 introduction when translating a work, translators must engage with structural differences between the original language and the target language, examining how the syntax and semantics of the two languages affect the integrity of the translation. syntax, or grammar, can affect a translation by altering the order in which words appear in a sentence. semantics, which concerns the meanings of words, must also be considered because even when words appear to have equivalents in other languages, slight differences in meaning can affect the translation’s quality. in his influential essay “the task of the translator,” walter benjamin discusses how these differences should not be dealt with by practicing one-to-one translation, forcing the target language into the restraints of the original language, but rather by allowing the translation the flexibility to create something new. guided in part by benjamin’s essay, i will present my own english translation of rainer maria rilke’s german language poem “der panther,” which does not look to translate words and phrases in isolation but distributes themes found in rilke’s german throughout the translation. i will discuss how my translation, which takes influence from the structure of the german language, better preserves rilke’s portrayal of the panther than certain historically important english translations of the poem. in particular, i will focus on semantic challenges surrounding rilke’s characterization of the panther as a powerful being reduced to listlessness and syntactic challenges related to how english and german word orders are parsed differently. i will also discuss how rilke’s language creates a kunst-ding (art-object) by portraying ambiguity between physical and non-physical worlds and how this ambiguity can be preserved in english translation. below, i present rilke’s original poem, my translation, and the translations of stephen mitchell, robert bly, and c. f. macintyre for comparison. aresty rutgers undergraduate research journal, volume i, issue iv der panther im jardin des plantes, paris 1 sein blick ist von vorübergehen der stäbe 2 so müd geworden, daß er nichts mehr hält. 3 ihm ist, als ob es tausend stäbe gäbe 4 und hinter tausend stäben keine welt. 5 der weiche gang geschmeidig starker schritte, 6 der sich im allerkleinsten kreise dreht, 7 ist wie ein tanz von kraft um eine mitte, 8 in der betäubt ein großer wille steht. 9 nur manchmal schiebt der vorhang der pupille 10 sich lautlos auf—. dann geht ein bild hinein, 11 geht durch der glieder angespannte stille — 12 und hört im herzen auf zu sein. rainer maria rilke the panther in the jardin des plantes, paris 1 his gaze has become so weary from the passing of the bars, 2 that it does not hold anything anymore. 3 for him, it is as though there were a thousand bars, 4 and behind a thousand bars no world. 5 the soft stride, the powerful footsteps, 6 which pace in the smallest of circles, 7 are like a dance of power around a center 8 where a great will, reduced to languor, stands. 9 only sometimes does the curtain of the pupil 10 push noiselessly open—. then an image enters, 11 passes through the tense limbs of stillness— 12 and ceases, in the heart, to be. my translation translated by stephen mitchell: in the jardin des plantes, paris 1 his vision, from the constantly passing bars, 2 has grown so weary that it cannot hold 3 anything else. it seems to him there are 4 a thousand bars, and behind the bars, no world. 5 as he paces in cramped circles, over and over, 6 the movement of his powerful soft strides 7 is like a ritual dance around a center 8 in which a mighty will stands paralyzed. 9 only at times, the curtain of the pupils 10 lifts, quietly—. an image enters in, 11 rushes down through the tensed, arrested muscles, 12 plunges into the heart and is gone. translated by robert bly in the jardin des plantes, paris 1 from seeing the bars, his seeing is so exhausted 2 that it no longer holds anything anymore. 3 to him the world is bars, a hundred thousand 4 bars, and behind the bars, nothing. 5 the lithe swinging of that rhythmical easy stride 6 which circles down to the tiniest hub 7 is like a dance of energy about a point 8 in which a great will stands stunned and numb. 9 only at times the curtains of the pupil rise 10 without a sound . . . then a shape enters, 11 slips through the tightened silence of the shoulders, 12 reaches the heart, and dies. aresty rutgers undergraduate research journal, volume i, issue iv translated by c. f. macintyre in the jardin des plantes, paris 1 his sight from ever gazing through the bars 2 has grown so blunt that it sees nothing more. 3 it seems to him that thousands of bars are 4 before him, and behind him nothing merely. 5 the easy motion of his supple stride, 6 which turns about the very smallest circle, 7 is like a dance of strength about a center 8 in which a will stands stupefied. 9 only sometimes when the pupil’s film 10 soundlessly opens . . . then one image fills 11 and glides through the quiet tension of the limbs 12 into the heart and ceases and is still. 2 the linguistic relationship between noiselessness, motionlessness, and tranquility in german instead of committing to a literal translation which will inevitably fail to transmit some aspects of the original, benjamin, quoting rudolf pannwitz, suggests that translators should instead allow “[their] language to be powerfully affected by the foreign tongue. [...] our translations, even the best ones, proceed from a mistaken premise. they want to turn hindi, greek, english into german instead of turning german into hindi, greek, english” (benjamin, 1996, pp. 261-262). according to benjamin, a translator can best highlight the original work when they figure out how to convey in their own language what is unique to the foreign language. by integrating elements of the foreign language, the translator’s language becomes enriched. drawing on this concept, in my translation, i attempt to reproduce rilke’s portrayal of the panther not by finding english words that replicate german words, but by letting the semantics of german words be distributed throughout my entire translation. in rilke’s poem, the connotations of noiselessness and motionlessness in the words “stille” (stillness, silence), “lautlos” (noiseless), and “betäubt” (numb), are central to his portrayal of the panther as a powerful spirit whose force is suppressed in captivity. as the characterization of the panther is grounded in these words, i will examine how these words interact with one another to immerse the poem with a tone of stillness and tranquility and explain how i attempt to create a similar effect in my translation. i will begin by discussing the word “stille” (rilke, 1907, ll) to explain how the notions of noiselessness and motionlessness are related to one another in german. stille means both “silence” and “stillness” in english in addition to connoting tranquility or serenity. it is clear that there is not one singular word in english that could capture its entire essence. it is not coincidental that the concepts of silence and stillness are represented by the same word in german; other german words also convey an underlying relationship between silence, stillness, and serenity. the word ruhe, for example, can mean peace and silence, and its adjective form ruhig means quiet or calm. furthermore, the word taub embodies noiselessness in its meaning of deafness as well as connoting physical immobility in its meaning of numbness. through the semantic information encoded into these words, german conveys that motionlessness and noiselessness are qualities of the peace and serenity that stille embodies. the word “lautlos” (rilke, 1907, 10), which i translate as “noiselessly,” already strongly communicates the notion of noiselessness or silence, and therefore when translating stille, i prioritized its connotation of motionlessness. thus, i opted to translate stille as “stillness” rather than “silence.” the word taub mentioned above, which means “deaf” or “numb,” also makes an appearance in rilke’s poem as the root of the word betäubt. betäubt appears in rilke’s poem in a sentence which could most literally be translated to “where a great will stands numb” (rilke, 1907, 8). however, the nature of the numbness depicted in rilke’s poem could benefit from a less literal translation than simply settling for the word “numb” or even “benumbed.” when the prefix beis applied to a word in german, as in the word be-täubt, it means that that thing is being inflicted upon something or someone. therefore, when reading the german aresty rutgers undergraduate research journal, volume i, issue iv poem, an image comes to mind of the panther’s will becoming numb, as opposed to already existing in a state of numbness. the figurative motionlessness of the state of numbness and the literal motionlessness of stehen (literally translated as “standing,” and thus having a similar connotation of motionlessness in english), in conjunction with the panther’s enduring might, also contribute to a regal calmness to characterize the panther’s will which further resonates with stille. “numb” and “benumbed” are too lifeless to describe the complexity of the panther’s state of mind, while translations of the word betäubt from other translators such as “stunned” (bly, 1981, 8), “paralyzed” (mitchell, 1989, 8), or “stupefied” (macintyre, 1971, 8) are too violent and connote too much permanence for the calm tone of the original poem. if i were to settle for “stillness” as a translation of stille and “numb” as a translation of betäubt, i would be covering the original intent of the poem by eliminating meaning through literal translation. instead, i stray from literalness and return to the relationship between stille and taub and the delicate relationships between motionlessness, noiselessness, and tranquility. instead of attempting a direct translation of betäubt, i opt for a word that also incorporates aspects of the essence of stille, which i could not convey through the single word “stillness.” i chose “languor” in place of betäubt—not because they mean equivalent things but because “languor” captures many aspects of the panther’s body and mind throughout the poem—in addition to resonating with stille and betäubt. “languor” embodies the lack of motion and decreased mental activity implied by the word betäubt yet also contributes weariness, apathy, and lack of vitality. these effects are reflective of the panther’s body and mind in his captivity: his gaze is weary, pacing repetitive and passionless, and his power restricted. “languor” also resonates with the tranquility of stille in its connotations of relaxation and motionlessness in its connotation of a stillness or suspension in the air. furthermore, to preserve a changing state implied by the word betäubt, i write that the will is “reduced to” languor. this phrasing also reflects the state of being restricted, and strongly resonates with the physical and mental detriment the panther’s imprisonment inflicts on him. instead of losing meaningful parts of stille and betäubt through one-to-one translations, i attempt to distribute aspects of their meanings throughout my translation. in doing so, i attempt to convey the german language’s relationship between motionlessness, noiselessness, and tranquility in english. in “the task of the translator,” walter benjamin explains why translations should not aim to be literal: “fidelity in the translation of individual words can almost never fully reproduce the sense they have in the original” (benjamin, 1996, p. 259). even when a direct translation of a word appears to be possible, there will always be subtle differences in semantic information, emotional connotation, or relationships the word in question has with other words. this diminishes the poetic quality of the translation. therefore, in translating words related to the panther’s apathetic state, i do not simply evaluate the local meanings of words such as stille and betäubt, but also take into account how their meanings relate to one another and to the panther’s overall characterization. for the word betäubt in particular, my translation—“reduced to languor”—does not emphasize the local connotation of numbness as much as it prioritizes the motionlessness and lack of vitality characteristic to the panther throughout the entirety of the poem. thus, i ensure that the way rilke characterizes the panther with the themes of motionlessness and noiselessness is distributed throughout my translation, instead of just being present in individual words. prioritizing the selection of words i feel best represent the notions of noiselessness, motionlessness, and tranquility as represented in rilke’s poem, comes at the cost of replicating aesthetic components such as its meter and sound. my translation does not replicate rilke’s abab rhyme scheme, as mitchell does, or his loose iambic pentameter, as macintyre does. to preserve these components of rilke’s work, mitchell and macintyre are prone to choosing words that over translate the original. their choices for the word betäubt— “paralyzed” and “stupefied” respectively—are both words which igaresty rutgers undergraduate research journal, volume i, issue iv nore the quietness characteristic to betäubt. another example is the word “geht” (rilke, 1907, 11), translated as “rush” (mitchell, 1989, 11), or “glide” (macintyre, 1971, 11). these words overdramatize rilke’s tone; geht, which can mean “to walk,” is a very ordinary, unremarkable word which does not describe specific, complex, or rushed movement. therefore, i opted to translate geht as “passes through” in line 11, and as “enters” in line 10. these words convey a displacement from one point to another without inserting extra emotion or unique movement qualities not already present in rilke’s work. 3 differences in syntax english language is not as syntactically equipped to portray rilke’s easy flowing prose as the german language. a major difference between german and english syntactic structure is that while english is a svo (subject-verb-object) language, german is a sov (subject-object-verb) language. while both word orders can attempt to convey the same concepts, the way they are parsed in poetry differs. by having the subject and the verb—the parts of the sentence most essential to the overall meaning of the phrase—directly next to each other, a certain segmentation is imposed upon my english translation of the poem which does not occur in german. one example is the opening line, “sein blick ist von vorübergehn der stäbe /so müd geworden” (“his gaze has by passing of bars so tired become”) (rilke, 1907, 1-2). in german, the reader cannot stop reading anywhere along this phrase and still grasp the intended meaning of the sentence. the verb phrase “müd geworden” or “become tired” lies at the very end of the sentence, yet is essential for understanding the action. if all the supplementary information is removed from the sentence, “sein blick ist müd geworden” remains. when the supplementary information, “von vorübergehen der stäbe,” is added, the verb geworden remains at the end of the sentence. because the supplemental information is integrated between the subject and the verb, it is integrated seamlessly into the sentence. in my english translation, however, all the essential information, “his gaze has become so weary,” appears at the beginning of the sentence, while “from the passing of the bars” appears after it, as if an afterthought. since the reader already has all the necessary information from the first half of the sentence, they are not required to fully attend to the meaning of “from the passing of the bars.” there is also a clear separation between the phrases “his gaze has become so weary” and “from the passing of the bars” which disrupts the unity of the phrase and imposes some choppiness. in paul de man’s response to benjamin’s “the task of the translator,” he emphasizes that syntax or grammar influence meaning. de man explains that when syntax is translated literally, meaning is lost because the way grammar yields meaning differs between languages (de man, 1985, 41). this idea applies to the syntactic differences between german and english i described before in that by translating “der panther” into a language which places verbs after the subject rather than after the object, the way information is emphasized in certain lines of the poem changes. despite this, i opted to mainly stay with colloquial english word orders rather than attempt to rearrange my word orders to better imitate the german word order. this is because rilke’s poem reads quite colloquially, with his word orders being those which would be used in common, everyday german. the choice to maintain a colloquial english word order was made to preserve the causal way rilke’s poem reads in german. this way, i incorporated some aspects of his style in my translation even while choosing not to replicate his rhyme scheme or meter. while i believe that translating syntax less literally has the benefit of replicating rilke’s colloquial language in english, and that translating semantics less literally works for words such as stille and betäubt, these remain situation-specific choices. i believe the last line of the poem warrants a more literal translation in regards to both syntax and word choice. in the last stanza of the poem, rilke describes an image which comes through the panther’s pupil, travels to his heart, and ceases to be. the last line of the poem, describing the image’s final moments, reads: “und hört im herzen auf zu aresty rutgers undergraduate research journal, volume i, issue iv sein” (rilke, 1907, 12). significantly, the last word of the poem “sein,” meaning “to be,” is the same as the first word of the poem “sein,” which in that case is the possessive pronoun “his.” that the first and last words of the poem are the same is important because the transition from sein as a possessive pronoun to sein as a description of the cessation of existence describes a deterioration of the panther’s being: at the beginning of the poem the panther is a being capable of possessing, but by the end he fails to hold the image in his body. my translation of the last line, “and ceases, in the heart, to be,” keeps the infinitive “to be” (in german, sein) as the last word of the poem while also replicating the way rilke shortens this line compared to the rest of the poem and imitating the way his german inserts information between the two segments of the separable verb aufhören. similar to german, my separable verbs nestle information between subjects and verbs on a smaller scale. in the last line, “und hört im herzen auf zu sein,” the verb aufhören is broken up into the segments auf and hört. verbs that appear in sentences as two separate particles do not occur in english. however, by representing the phrase like this: “and [ceases] in the heart [ceases] to be,” it can be communicated that “ceases” occupies not one but two positions in the phrase. in german, the supplementary information, the prepositional phrase “im herzen” or “in the heart” is inserted between the verb’s two segments, creating the impression of a single phrase. this allows for all information to seem equally important to the meaning of the sentence because all information is encountered before the end of a phrase. similar to how in rilke’s german, the verb aufhören is broken up into “hört” and “auf” with the phrase “im herzen” appearing between them, my translation thus reads, “and ceases, in the heart, to be.” the phrase “ceases to be” is broken into the phrases, “ceases” and “to be,” with “in the heart” appearing between them. while this isn’t the most colloquial way to state this in english, breaking up the phrase “ceases to be” with a pause after the word “ceases” emphasizes the information that is to come: the image’s discontinued existence and the invocation of the infinitive “to be” which is so central to the german poem. this more literal translation also has the benefit that it does not over translate. unlike mitchell’s colorful language, which describes the image as “[plunging]” (mitchell, 1989, 12), or bly, who somewhat misleadingly suggests that the image “dies” (bly, 1981, 12), a literal translation keeps my language free of action words or words that evoke visceral emotions while also staying close to the discussion of existence central to the word sein. 4 ambiguity between physical and non-physical worlds rilke’s german portrays an ambiguity as to whether the panther resides in physical or nonphysical realms. poetic subjects which do not conform to characterization as strictly physical objects are common to rilke’s works. in her essay “how the panther stole the poem: the search for alterity in rilke’s dinggedichte,” claire y. van den broek discusses rilke’s quest to create a kunst-ding (art-object) which can exist autonomously, unmovable by the reader. rilke claims that in order for an object to be eternal, it must be freed from time and space (van den broek, 2013). in “der panther,” rilke does this by creating ambiguity between the physical and the non-physical realm. most translations, however, fail to transmit this. consider, for example, the lines “geht durch der glieder angespannte stille-/ und hört im herzen auf zu sein” (rilke, 1907, ll-12), or according to my translation, “passes through the tense limbs of stillness / and ceases, in the heart, to be,” where rilke describes an image, or bild, passing through glieder, or “limbs,” of stillness, to land in the panther’s heart. most translators interpret the word glieder as the limbs of the panther’s body. i, however, interpret this as rilke trying to give a physical form to stillness. this is not the only time in the poem that rilke conveys a separation between the body of the panther and the bodies of abstract concepts: in the second stanza, there is a clear physical distance between the panther’s will, which stands in the center of the cage, and the panther’s body, which paces around it. by assigning the will, an abstract concept, to a location in space, rilke creates aresty rutgers undergraduate research journal, volume i, issue iv ambiguity between what belongs to the physical realm and what belongs to the non-physical realm. simultaneously, he creates ambiguity about the scope of the panther’s body and mind, as one would typically assume that a person’s will resides inside their body. again, in the phrase glieder angespannte stille, or “limbs [of] tense silence,” there is ambiguity as to what is tangible and what is not, as well as what belongs to the panther’s body and what does not. rilke assigns stillness, which generally does not have a physical form, the physical characteristics of having limbs and being tense. the image which comes through the panther’s pupil to pass through the limbs also does not necessarily have a physical form. it can pass through the stillness suggesting that like the will, stillness has a location in space. the notion that the image passes through stillness in order to enter the panther’s heart also begs the question of whether stillness itself is a part of the panther’s body. the ambiguities for which abstract concepts can take physical forms, or of what the panther’s body consists of are lost when this line is translated to show that it is the panther’s limbs which the image passes through: e.g., “the tensed arrested muscles” (mitchell, 1989, 11) or “tightened silence of the shoulders” (bly, 1981, 11). by asserting that the limbs belong to the panther, mitchell and bly transform the pathway of the image into one which occurs wholly in the panther's physical body. other translators forcing the interpretation that the limbs refer to the panther’s limbs is what werner hamacher would call an “attack.” on the topic of translation, he states: “only through separation from meaning does an idea take the place of an attack” (hamacher, 2019, xliii). it is when translators are preoccupied with transmitting a certain interpretation that translations attack the original by imposing something that was not previously there. instead of interpreting a specific meaning, my translation leaves open the possibility that the poem is set in a nonphysical realm. my translation reads: “passes through the tense limbs of stillness.” this phrasing does not force the reader to assume that the limbs belong to the panther. it can be understood in multiple ways: that the panther’s limbs are comprised of stillness, or that the limbs belong to a physical manifestation of stillness. the reader is not forced to assume that the limbs are the panther’s physical limbs, and they are also free to imagine stillness having a physical form. thus, it contributes to rilke’s interplay between physical and the nonphysical worlds. furthermore, the term angespannte stille connotes suspension of time as well as physical immobility. by opting for a more literal translation of angespannt (tense) and stille (stillness), i preserve some of the temporal aspects of this phrase. by creating ambiguity as to whether the tenseness refers to the limbs or the stillness, or whether the limbs belong to the panther or to stillness itself, the path of the image to the panther’s heart is neither bound to the physical world nor is it governed by the passage of time. the pathway of the image is bound only to the panther himself, whatever his body may consist of. this freedom of interpretation contributes to a panther who is— characteristically to kunst-dings—freed from time and space. 5 minor points regarding deviations from other translations one word which i did not find a suitable translation for is blick (rilke, 1907, 1), which refers to the panther’s gaze as he is constrained in his cage. i opted to use the word “gaze,” while other translators used “vision” (mitchell, 1989, 1), “seeing” (bly, 1981, 1), and “sight” (macintyre, 1971, 1). “vision,” and “sight” are words which emphasize the capacity for seeing, whereas “gaze” and “seeing” are words which emphasize the action of seeing. while i believe the word blick refers more to the action of seeing than the capacity for it, my translation, “gaze,” denotes a more sustained and prolonged look even though blick represents the shorter duration of a glimpse or a glance. unfortunately, the words “glimpse” and “glance” in english are more colloquially used in phrases such as “stole a glance” or “caught a glimpse.” using the word “glimpse” apart from such a phrase and translating rilke with “his glimpse has become so weary” would be awkward. one place where nearly all other translations i consulted deviated from rilke’s syntax is “der aresty rutgers undergraduate research journal, volume i, issue iv weiche gang geschmeidig starker schritte” (rilke, 1907, 5), which describes the panther’s footsteps as he paces around the cage. rilke’s phrase lacks a preposition and could be translated literally as, “the soft walk smooth strong steps.” all the translations i consulted, however, inserted a preposition into the phrase: “the easy motion of his supple stride” (macintyre, 1971, 5), “the lithe swinging of that rhythmical easy stride” (bly, 1981, 5), “as he paces in cramped circles” (mitchell, 1989, 5) (emphasizes mine). keeping with rilke’s syntax, my translation, “the soft stride, the powerful footsteps,” does not insert a preposition into this line. 6 conclusion observing the way rilke’s german conveys themes such as the noiselessness, motionlessness, and listlessness in the panther’s physical body and mental state, i attempt to provide a translation of “der panther” which preserves the way these themes are represented in the poem. by first analyzing the poem’s themes and then examining how local translation choices affect the poem in its entirety, my translation would allow for an analysis of the poem which remains closer to the original than analyses of other translations. one such example is that my translation does not restrict the scope of the panther’s mind and body to the realm of linear time and space, allowing for the panther to be analyzed as one of the kunst-dings characteristic to rilke’s writing. thus, my translation resonates with rilke’s other works. furthermore, by staying consistent with benjamin’s idea that translators should let the target language be affected by the original language, my translation takes influence from how the structure of german portrays meaning, as well as how it allows itself to be read in poetry. 7 acknowledgements i would like to thank professor nicola behrmann for her helpful comments on the paper. 8 references [1] benjamin, walter. “the task of the translator.” translated by harry zohn. in walter benjamin: selected writings volume 1 1913-1926, edited by marcus bullock and michael w. jennings, 253-263. harvard university press, 1996. [2] bly, robert, and rainer maria rilke. 1981. selected poems of rainer maria rilke. translated by robert bly. new york: harper perennial. [3] broek, claire y. van den. 2013. “how the panther stole the poem: the search for alterity in rilke’s ‘dinggedichte.’” monatshefte (madison, wisconsin : 1946) 105 (2): 225–46. https://doi.org/10.1353/mon.2013.0046. [4] de man, paul. 1985. “‘conclusions’ walter benjamin’s ‘the task of the translator’ messenger lecture, cornell university, march 4, 1983. yale french studies 65: 25-46. [5] hamacher, werner. 2019. “95 theses on philology.“ in give the word: responses to werner hamacher’s 95 theses on philology, edited by gerhard richter and ann smock. lincoln: university of nebraska press. [6] rilke, rainer maria. 1907. “der panther.” in neue gedichte und der neuen gedichte anderer teil, 3334. frankfurt am main: insel taschenbuch verlag. [7] rilke, rainer maria. 1971. “the panther.” translated by c. f. macintyre. in rilke: selected poems, edited by c. f. macintyre. berkeley, los angeles, london: university of california press. [8] rilke, rainer maria. 1989. “the panther.” translated by stephen mitchell. in the selected poetry of rainer maria rilke: bilingual edition, edited by stephen mitchell. new york: vintage internationa https://doi.org/10.1353/mon.2013.0046 aresty rutgers undergraduate research journal, volume i, issue iv serena graduated from rutgers in may 2022 with majors in german and cognitive science. she became interested in translation as a means to combine her interest in the philosophy of language with her interest in practicing german. during her time at rutgers, she was involved in various theoretical and empirical language based research projects, including serving as a research assistant for the comparative and experimental linguistics lab, conducting a project about heritage languages and emotions as an interdisciplinary research teams fellow, and working on her senior thesis in german, titled praying for language: hamacher’s bogengebet and the possibility of address. this year, serena will be in graz, austria as a fulbright student to combine her interests in science, philosophy, and literature by conducting a research project about the epistemological frameworks underlying quantum theory. aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. observing winter carbonate chemistry dynamics throughout the mid-atlantic bight shelf using novel glider technology marissa guzik, grace saba (faculty advisor), elizabeth wright-fairbanks phd ✵ abstract increased atmospheric carbon dioxide (co2) has led to global climate change and ocean acidification (oa) via the absorption of atmospheric co2 by the ocean. coastal shelves are also affected by various processes that influence the acidity of seawater, causing acidity to vary over time and space. these variations in ocean acidity can negatively impact marine species, especially calcifying organisms such as surfclams and sea scallops. in the mid-atlantic bight (mab), a subsection of the u.s. northeast shelf (nes), this variation in acidity generates ecological and economic concerns as the mab is home to some of the nation’s most productive and profitable estuaries and fisheries. in this study, rutgers university (southern mab) and stony brook university (northern mab, hudson canyon) deployed two gliders equipped with sensors measuring depth, temperature, salinity, ph, dissolved oxygen, and chlorophyll to monitor winter 2021 carbonate chemistry conditions on the shelf as well as in slope waters of the mab. for both deployments, measured ph and calculated aragonite saturation state (ωarag) showed opposing patterns, with high ph and low ωarag in shelf/nearshore and low ph and high ωarag in slope waters. these trends were attributed to different driving factors whereas ph was more influenced by biological processes (i.e. photosynthesis) and ωarag was influenced mostly by thermodynamics and chemical factors (i.e. temperature, total alkalinity). the results of this study underscore the importance of monitoring coastal acidity to understand potential impacts on important species. key terms: ocean acidification, carbonate chemistry, glider, aragonite saturation state, ph, total alkalinity, mid-atlantic bight, vertical mixing, gulf stream, labrador current common terms and abbreviations mid-atlantic bight (mab): ocean region spanning just offshore from cape hatteras, nc to martha's vineyard, ma. cold pool: mass of dense, cold bottom-water trapped due to stratification between surface and deep water during the late spring through summer months. aragonite saturation state (ωarag): calcium carbonate saturation state with respect to aragonite, measure for the potential of carbonate ions to form or dissolve; ω < 1: dissolve, ω > 1: form. total alkalinity (ta): measure of the water’s capacity to neutralize acids (h+ ions), measures the concentration of all alkaline ions (including carbonate and bicarbonate). ocean acidification (oa): the increasing acidity levels of the ocean due to increased atmospheric co2 absorption from anthropogenic co2 emissions. partial pressure co2 (pco2): the pressure from carbon dioxide gas molecules; increasing pco2 increases the solubility of carbon dioxide. shelf: region of shallower ocean water over the continental shelf (refer to figure 1). in-slope: deeper water just around shelf-break (the edge of shelf). on shelf: water on the continental shelf. nearshore: water sampled closest to the shoreline, shallower water. u.s. northeast shelf (nes): a nearshore system with an extended shallow shelf. dissolved inorganic carbon (dic): total sum of inorganic carbon (co2, hco3-, co3-2) in seawater. aresty rutgers undergraduate research journal, volume i, issue iv water-mixing: the mixing of surface and deeper water from wind-driven forces (waves). glider: autonomous underwater robots that sample continuously as they travel up and down through the water column along a programmed route (deployment refers to the initial release of a glider and recovery refers to the retrieval of the glider from the ocean). coastal downwelling: surface water that is forced downward as wind moves it toward the coastal boundary which then becomes deep water that travels away from the coast. coastal upwelling: deep water that is dragged up along the coast to replace surface water that was moved offshore by the wind. stratification: the separation of water column layers due to differing water densities and properties, typically observed in seasons with less ocean-mixing. thermocline: a thin transition layer of water separating the surface and bottom water layers; due to the strong temperature differences between these layers, a rapid change in temperature is measured as one moves through the thermocline. eddies: counter-current flow of water generating small whirlpools and circular movement of water. buffering capacity: the ability of seawater to maintain a more neutral ph after being introduced to more acidic solutions, due to the concentration of salts and ions (such as co3-2) in seawater absorbing free-floating protons (h+). 1 introduction anthropogenic activity, including the increased burning of fossil fuels and deforestation, has increased atmospheric carbon dioxide (co2) and altered global climate and ocean conditions (ipcc, 2019). the increased rate of atmospheric co2 absorption in the oceans, which results in complex chemical reactions that decrease seawater ph, has generated global concern over the acidification of ocean waters (wanninkhof et al., 2015). ocean acidification (oa) alters the balance of the carbonate system, including decreasing calcium carbonate saturation state in addition to the aforementioned decrease in ph. this presents a challenge for marine species by increasing physiological stress, as they must increase energy expenditure on maintaining body structures or processes that depend on calcium carbonate stability. this specifically affects calcifying organisms such as sea scallops, which use carbonate to form their protective carbonate shells (kroeker et al., 2013; saba et al., 2019a). seawater ph is projected to continue to globally decline over the next century, decreasing from the global average of 8.1 by 0.1-0.4 ph units (ipcc, 2019). this is expected to cause significant impacts on the health of ecosystems and play a role in the future distribution and range of species based on a species’ tolerance to more acidic conditions. coastal processes influencing nearshore carbonate chemistry while ocean acidification is increasing globally at a relatively consistent rate, carbonate chemistry conditions of the seawater vary greatly across different spatial and temporal scales in nearshore coastal systems (goldsmith et al. 2019). nearshore waters experience these shifts due to various freshwater and terrestrial inputs, as well as mixing processes, which can alter the chemistry of the water (wanninkhof et al., 2015; goldsmith et al., 2019). for example, freshwater is naturally more acidic due to lower salinity and total alkalinity (ta) (kwiatkowski & orr, 2018). additionally, photosynthesis and respiration can biologically alter carbonate chemistry, which increases and decreases ph, respectively. respiration will dominate and increase partial pressure of co2 (pco2) at deeper depths (cai et al., 2011). increased organic carbon and nutrients from rivers can lead to eutrophication of coastal systems and exacerbate this biological control on acidification (anderson et al., 2002; cai et al., 2011; xu et al., 2020). upwelling of deep, co2 and nutrient rich water can support algal blooms and has previously been shown to result in increased acidity in the surface water (anderson et al., 2002; degrandpre et al., 2002) seen along the u.s. west coast and seasonally off the coast of new jersey. water mass mixing, or a lack thereof, plays a significant role in controlling carbonate chemistry in coastal regions (cai et al.,2020; wright-fairbanks et al., 2020). during aresty rutgers undergraduate research journal, volume i, issue iv periods of strong stratification, a lack of mixing generates pco2 rich bottom water and lower pco2 surface water, creating vertical variation in carbonate chemistry. in contrast, well-mixed water columns have a relatively uniform pco2, meaning variation of carbonate chemistry is more minimal throughout the water column. in the northeast u.s., strong stratification begins in the spring and persists through summer when higher temperatures strengthen thermocline and prevent mixing, while high mixing occurs throughout the fall during seasonal overturn and persists into winter (degrandpre et al., 2002; wright-fairbanks et al., 2020). influence of currents and seasonality on carbonate chemistry the two major currents that influence the chemistry of the u.s. northeast shelf (nes) and slope water are the labrador current and gulf stream (figure 1). the labrador current is a cold, fresher current that transports water southward from the nova scotia region to the mid-atlantic bight (mab). this colder, less alkaline water allows for more co2 absorption, leading to lower calcium carbonate saturation state with respect to aragonite (ωarag). the gulf stream is a warm, salty current that transports water northward from the gulf of mexico along the coastline, diverging offshore at the edge of the mab near cape hatteras, nc. compared to the labrador current, it is characterized by a high ωarag and ta. overall, these two opposing currents influence the carbonate chemistry of the northeast shelf through mixing of waters with different co2 buffering capacities (wanninkhof et al., 2015; xu et al., 2020). along with the influence of these currents, the aforementioned mixing processes and coastal influences drive spatial and temporal variations in nes seawater carbonate chemistry. high summer temperatures generate a strengthened thermocline leading to a highly stratified water column. this warmer, less dense surface water lays on top of a bottom layer of colder fresher water — most likely originating from the labrador current — which becomes confined underneath by the density difference. this trapped figure 1: (a) map of the nes, white box highlighting the mab. (b) map focused in on the mab area, illustrating concept of nearshore, on-shelf, and in-slope water. a b aresty rutgers undergraduate research journal, volume i, issue iv cold, fresher water mass creates a unique shelf feature specific to the mab referred to as the “cold pool,” and its presence or absence (based on the degree of mixing) can impact the carbonate chemistry throughout the water column. with the formation of the cold pool during the summer months, a pocket of high pco2 is contained at the shelf floor, creating a stark difference in carbonate chemistry between the cold pool water and the lower pco2 water of the rest of the water column (wright-fairbanks et al., 2020). economic and ecological significance it is important to note that these coastal processes and their seasonal cycles are not well studied on a regional scale, and therefore the effects of seasonal variation regarding oa and the impacts on ecosystem health along the nes are not well understood either (xu et al., 2020). however, studies on oa and understanding coastal variations are important, especially for coastal fisheries. the nes supports many of the nation’s most productive and valued commercial fisheries (national marine fisheries, 2021). atlantic sea scallops (placopecten magellanicus) and atlantic surfclams (spisula solidissima) are both calcifying organisms that rely on carbonate molecules to develop their shells; these important atlantic fisheries generate $569.9 million and $30.7 million, respectively (national marine fisheries, 2021). shellfish like scallops and surfclams that inhabit low carbonate environments (hart et al., 2004; wanninkhof et al., 2015; pousse et al., 2020) can be at increased risk of mortality from predation with acidification, causing shell deformities and weaker shells (reviewed in saba et al., 2019a). gliders due to the dynamic nature of coastal systems and their importance in economically productive fisheries, it is critical to monitor and understand oa and carbonate chemistry. this is especially urgent as climate change is predicted to generate a global trend of increased acidity. high-resolution data collected over short-term, or seasonal, time periods can help to construct a better understanding of the effects of the different coastal processes on acidity trends along the northeast shelf (xu et al., 2020; xu et al., 2017; wanninkhof et al., 2015; wright-fairbanks et al., 2020). gliders are autonomous underwater vehicles that can be equipped with various sensors, including the recently developed ph sensor, enabling continuous ph measurements to be taken throughout their deployments. gliders travel off shelf and back to shore, continuously taking measurements while climbing up and down through the water column, providing highresolution observations of the shelf. glider science sensors sample at a rate of 0.5 hz, resulting in interval measurements of every 20-30cm vertically producing these high-resolution datasets (saba et al., 2019b; wright-fairbanks et al., 2020). this ability to provide high-resolution data and cross-sectional perspectives of the water column makes gliders ideal for monitoring oa and carbonate saturation along the northeast shelf (saba et al., 2019b). rutgers university continues to use ph gliders to conduct seasonal surveys of the shelf to investigate carbonate chemistry dynamics as well as to map potential hot spots and time periods of acidification in important fisheries’ habitats. the purpose of my research project is to examine shelf-wide winter carbonate chemistry dynamics using data collected from two ph glider deployments conducted in the northern and southern mab. 2 materials and methods data collection and analysis methods followed those used in seasonal surveys described in wright-fairbanks et al. (2020). the winter 2021 surveys were conducted as a collaborative effort between rutgers university (southern mab deployment) and stony brook university (northern mab deployment). teledyne-webb slocum g2 gliders were deployed in each mab site during similar timeframes. both gliders were deployed on february 26, 2021. the rutgers university glider was recovered on march 21, 2021, and the stony brook university glider was recovered on march 23, 2021, with missions lasting about 24 and 25 days respectively (figure 2a). the rutgers university glider was deployed and recovered out of tuckerton, nj (figure 2a – route r), and the stony brook university b aresty rutgers undergraduate research journal, volume i, issue iv glider was deployed out of shinnecock inlet, ny and recovered out of manasquan, nj (figure 2a – route s). each glider was equipped with a recently developed deep-isfet ph sensor (saba et al., 2019b) as well as ctd (conductivity, temperature, and pressure/depth), an aanderaa optode measuring dissolved oxygen (do), and an optics puck measuring chlorophyll fluorescence and spectral backscatter. further details on the specific procedures for preparation, deployment, and recovery of gliders is outlined in wright-fairbanks et al. (2020). during deployment and recovery, discrete water samples were collected and sent to a lab at the university of new hampshire for chemical analysis for the purpose of ground-truthing glider ph and carbonate parameter data (wright-fairbanks et al., 2020). this validation sample analysis is still ongoing. data collected by the gliders were converted using slocum power tools (kerfoot, 2014) into a format for data analysis using matlab software (version r2021a). analysis techniques described in wright-fairbanks et al. (2020) were applied to the data collected from these winter 2021 deployments and included sensor response time shift calculations and application to the full dataset, data quality assurance and quality control (qa/qc), estimating ta-salinity regression relationships to calculate ta, calculation of the full suite of carbonate chemistry parameters (ph, ta:dic, ωarag) (using co2sys v3.0 in matlab), binning parameter data by depth and distance/time, plotting parameters, and running statistical analyses/tests on the full quality-controlled dataset. methods for this stepwise analysis are described in detail below. sensor time lag can occur because temperature measurements are taken external to the ctd conductivity cell while conductivity measurements, used for calculating other variables including salinity and ta, are taken within the cell. this separation can cause a “thermal lag” where temperature measurements misalign and generate offset measurements (saba et al., 2019b). to account and correct this thermal lag, different potential time shifts from 0 to 60 seconds (at 1 second intervals) were run on paired upcast/downcast data, and average time shift for the entire deployment was applied to dataset (wright-fairbanks et al., 2020). once time shift was applied, qa/qc tests (gap test, syntax test, location test, gross range test, and spike test) were run on the glider dataset to assess the data for any “bad” data points (outlying, incredibly deviated points) created from sensor failure/malfunction and remove if present (further detailed in ioos, 2019). because discrete sample analysis is ongoing, we were unable to construct a deployment-specific linear regression model of ta as a function of salinity. instead, a previous winter deployment (2019) ta-salinity relationship was used to derive regression coefficients for calculating ta values, meaning slight variations in ta values between true calculations (if values were calculated using 2021 discrete samples) and calculated are possible. the previous winter deployment used to calculate ta values lasted a duration of 19 days, from february the 1st to the 19th, 2019, and was deployed offshore sandy hook, nj maintaining a course in the mab in between the glider tracks of the winter 2021 deployments (wright-fairbanks et al., 2020). this refined data set was then used to calculate carbonate chemistry parameters using co2sys for matlab (v3.0) (lewis & wallace, 1998; sharp et al., 2020; van heuven et al., 2011; wright-fairbanks et al., 2020). following this, the data were then binned by time (1-hour bins) to generate a series of profiles of the measured or calculated parameters over depth (m) and time (hours). these parameters included: temperature (°c), salinity (psu), dissolved oxygen (mg/l), chlorophyll (µg/l), ph, ta (µmol/kg), and ωarag. to further understand and visualize the data, in addition to the depth profiles, 3-dimensional cross-section models of the shelf were constructed. finally, statistical analyses were conducted to determine significance (p<0.05) of data spread and the relationships between parameters via correlation coefficients (wright-fairbanks et al., 2020). correlation coefficients were calculated using matlab package “statistics and machine learning toolbox,” version 2021a, and tables for both datasets for carbonate parameters were generated using microsoft excel (2008) and the conditional formatting function. aresty rutgers undergraduate research journal, volume i, issue iv 3 results glider data both glider missions exhibited similar trends in both chemical and physical parameters. temperatures above 12 °c and salinities above 35 psu were observed offshore/in-slope waters, while lower temperatures (below 8 °c) and salinities (below 33 psu) were observed nearshore (figure 3 a, b; figure 4 a, b). higher dissolved oxygen (above 9.5 mg/l) and chlorophyll (above 2 µg/l) concentrations were observed nearshore and along shelf, while lower concentrations for both (below 7.5 mg/l; below 1 µg/l) were observed at the shelf break and in-slope water (figure 3 c, d; figure 4 c, d). looking at carbonate chemistry parameters, ph was lower (7.95 – 8) in deeper in-slope water, while ph was higher (8.1-8.15) on shelf (figure 3 e; figure 4 e). higher ta (2300-2350 µmol/kg) and ωarag (2.4 – 2.6) were observed at the shelf break and in deeper inslope water and lower ta (below 2250 µmol/kg) and ωarag (below 2) were observed nearshore and on the shelf (figure 3 f, g; figure 4 f, g). prevalent mixing of the water column and uniform surface and bottom waters on the shelf were also observed in both glider datasets. unique oceanographic features were evident in the rutgers glider dataset (figure 3). past the shelf break, significant water column mixing was reflected in all profiles (3/6 – 3/9), in the upper 200 m where the glider sampled. additionally, when the rutgers glider was returning to the shelf (3/11 – 3/12) on the southernmost transect, a pocket of elevated chlorophyll (3 µg/l) was noted to coincide with elevated ph (8.15; figure 3e) and ωarag (2.4; figure 3g). figure 2: (a) maps illustrating glider mission tracks for the winter 2021 survey along nes. (route r) glider tracks for mab deployment (white) (star = recovery site, green dot = deployment site). (route s) glider tracks for hudson canyon deployment (magenta) (star = recovery site, green dot = deployment site). (b) at right, photography of rutgers glider (ru 30) being deployed (photo credit: elizabeth wright-fairbanks). aresty rutgers undergraduate research journal, volume i, issue iv whd. additionally, when figure 3: rutgers glider profilesdepth profiles for temperature (a), salinity (b), oxygen (c), chlorophyll (d), ph (e), ta (f), aragonite saturation state (g). depth is measured on the y-axis and time (m/d; from start of mission to end) is tracked on the x-axis, while each parameter’s values are plotted across a color gradient in respect to depth and time. aresty rutgers undergraduate research journal, volume i, issue iv figure 4: stony brook glider profilesdepth profiles for temperature (a), salinity (b), oxygen (c), chlorophyll (d), ph (e), ta (f), aragonite saturation state (g). depth is measured on the y-axis and time (m/d; from start of mission to end) is tracked on the x-axis, while each parameter’s values are plotted across a color gradient in respect to depth and time. aresty rutgers undergraduate research journal, volume i, issue iv examining the relationship between ph and ωarag similar patterns in the relationship between ph and ωarag relative to temperature and salinity were observed with both gliders (figure 5). for ph, higher values occurred at higher temperature and salinity, while lower ph values were associated with lower temperature and salinity values (figure 5 a, b). the opposite pattern was observed with ωarag, where low values occurred at low temperature and salinity, while high values were observed at high temperature and salinity (figure 5 c, d). to understand the drivers behind the ph and ωarag, the correlation coefficient tables (tables 1, 2) were used. for both datasets, ph had the strongest positive correlation with oxygen concentrations (0.73; 0.78) and, unlike ωarag, had a stronger positive correlation with chlorophyll concentration (0.59) as well. while ωarag had the strongest positive correlation with salinity and ta (0.94; 0.83), both ph and ωarag had a stronger correlation to temperature. the ph had a negative relationship with temperature (0.56; 0.70) while ωarag had a stronger positive figure 5: temperature-salinity plots for ph (top) and aragonite saturation state (bottom) for both rutgers (a, c) and stony brook (b, d) gliders. temperature is measured on the y-axis and salinity is measured on the x-axis, lines representing isopycnals (points of specific water densities) overlayed to give visuals of ocean water layering. parameter of interest values plotted across a color gradient in respect to temperature and salinity. aresty rutgers undergraduate research journal, volume i, issue iv relationship with temperature (0.91; 0.78). finally, ta had the greatest positive correlations to temperature (0.98) and ωarag (0.94; 0.83) as well as strong negative correlations with oxygen (0.88; 0.94) and latitude (-0.72; -0.86). these trends in the correlation coefficients and drivers for the carbonate parameters align with patterns of high/low values of related parameters, such as salinity and chlorophyll, observed by the gliders in nearshore/on shelf and in-slope waters. 4 discussion in this study, two sets of glider data provided by rutgers university (southern mid atlantic bight) and stony brook university (northern mid atlantic bight/hudson canyon) illustrated carbonate chemistry patterns in the mab. over the course of the winter 2021 mission, ph and ωarag varied across time and space from the influences of biological, chemical, and physical processes. this glider data can then be related to commercial fishery management zones, particularly shellfish fisheries, to understand the potential responses of marine animals to acidification. impact of currents on parameter profiles depth profiles of in-slope waters of both deployments suggest direct influence from the warmer, saltier gulf stream that produced higher temperature and salinity conditions, similar to those observed during a previous winter glider deployment in the mab (wright-fairbanks et al., 2020). these warmer, more saline conditions acted as drivers for high ta and ωarag values in deep in-slope waters for both datasets and surface in-slope waters for rutgers deployment. the spatial variation in ωarag values, with higher values seen in rutgers (southern) compared to stony brook (northern) deployment, is consistent with the “south-to-north decline” of ωarag described in cai et al. (2020) from the weakening influence of highly saturated and buffered gulf stream water. the gulf stream carries warmer saltier water from the tropics, which in effect has a lower dic/ta ratio, thus increasing the buffering capacity of the water maintaining higher ωarag (cai et al., 2020; wright-fairbanks et al., 2020). this northward current traveling alongside the shelf can be pushed up onto and mix with shelf waters via warm-core eddies (fratantoni et al., 2001; zhang et al., 2015; tables 1 and 2: correlation coefficients for both rutgers (left) and stony brook (right) glider data for the different carbonate parameters (ph, ta, ωarag). all coefficients were significant (p << 0.05). coefficient values greater than |0.5| (in bold) indicate stronger correlation. orange shaded cells = positive correlation, blue shaded cells = negative correlation; increased color saturation indicates increased correlation. *ta calculated as a function of glider salinity values, therefore coefficient of 1. aresty rutgers undergraduate research journal, volume i, issue iv wright-fairbanks et al., 2020) and consequently lead to high ta and ωarag, which was especially notable in the more southern rutgers dataset. conversely, nearshore shelf waters were generally colder and fresher, influenced both by coastal freshwater input and potentially the southward flowing labrador current (xu et al., 2017; wright-fairbanks et al., 2020). these fresher and shallow depth conditions of nearshore and shelf water were also characterized by higher chlorophyll and oxygen concentrations, the latter of which was likely a function of both higher productivity (i.e., oxygen production through photosynthesis) and coldwater temperatures (i.e., increased gas solubility). furthermore, these nearshore and shelf conditions translated to high ph and low ta and ωarag values, which is consistent with observations from the previous winter survey wright-fairbanks et al. (2020). the high observed ph values can be potentially attributed to the influence of increased biological uptake of co2 via increased photosynthesis in shallower shelf and nearshore water (cai et al., 2020; wright-fairbanks et al., 2020). ph and ωarag drivers the significant strong correlations of ph with chlorophyll and oxygen indicate that ph was more directly driven by biological influences, such as photosynthesis-respiration rates. conversely, ωarag was more so driven by physical and chemical influences, such as currents and water temperature, with significant strong correlations to ta, salinity, and temperature. this supports the conclusions of cai et al. (2020), who found that ωarag is temperaturedriven in the coastal western atlantic while ph is figure 6: satellite data for surface chlorophyll concentrations in the southern mab (3/9 – 3/15) (nasa goddard space), red box indicating area rutgers glider traveling through during same timeframe. rutgers glider chlorophyll data (top left) included for comparison, red box indicating elevated chlorophyll concentration of interest. aresty rutgers undergraduate research journal, volume i, issue iv more susceptible to short-timescale drivers like phytoplankton blooms. the drivers of ph and ωarag are further understood by looking at the unique, event-based features that occurred during the rutgers glider mission (figure 3). on the southern-most transect, as the glider returned on shelf (3/11 – 3/12), a pocket of both high ph and elevated chlorophyll was observed (figure 3 d, f). satellite data from the same timeframe (figure 6) confirmed the elevated glider chlorophyll measurements, supporting the fact that biological interactions — specifically photosynthesis rates — have a large impact on ph values. additionally, the direct relationships between ωarag and ta, salinity, and temperature were evident on the southern-most transect (3/11 – 3/12; figure 3), likely due to a storm-driven “sloshing event.” during early march strong and variable wind events, such as nor’easters and south/west winds, were observed within the glider data and promoted water column mixing and shifting currents (figure 7). earlier nor’easters (3/1, 3/5) produced strong northeast winds that traditionally promote downwelling of shelf waters due to the buildup of water along the coast. strong south and west winds beginning on 3/9 likely caused currents to shift in the opposite direction and potentially promoted a switch to intrusion of high ωarag in-slope water onto the shelf. these events further support that ωarag was more impacted by physical influences; in this case, ωarag was driven by wind-driven mixing. it is important to acknowledge that this is an observational study for an area of research where limited experimental studies have been conducted, limiting the ability to predict which variables drive ph and ωarag changes the most. the scarcity of experimental studies leaves opportunities for researchers to expand on our understanding of the drivers of ph and figure 7: 3d glider transects for ωarag, rutgers glider tracks (south) and stony brook (north), with shelf floor bathymetry and northeast coastline (drawn around long. -74.5) illustrated. black arrows = nor’easter (northeast winds, winds promoting downwelling on shelf water); magenta arrows = south and west winds (winds triggering shift to upwelling of offshore/in slope water). white box area of interest highlighting potential sloshing event from shifting currents. aresty rutgers undergraduate research journal, volume i, issue iv ωarag through controlled experiments in which significantly correlated parameters (oxygen and chlorophyll concentrations, temperature, etc.) are manipulated individually. strong winter mixing event in any given depth profile nearshore or on the shelf, the water column was fully mixed, as is typical for mab winter conditions (castelao et al., 2010). stratification of in-slope waters occurred during the early part of the deployment (early march) but disintegrated over time, likely due to increased storm-related winds. this event was most pronounced in the southern (rutgers) dataset where strong mixing of the water column was observed (3/6 – 3/9) within deeper in-slope water, revealing full water column homogenization for the maximum depth sampled (200m) (figure 3). this was the first time this degree of slope water mixing event has been observed in a rutgers winter survey. this mixing event highlights the prevalence of interannual variability of the gulf stream-influenced shelf break jet in the mab, which has been described by linder (1996) and linder and gawarkiewicz (1998). typically, the shelf break jet is more stratified for in-slope waters due to differing salinities between denser, saltier gulf stream water and fresher overlaying surface water in the winter, which was seen in the stony brook deployment (figure 4) (linder & gawarkiewicz, 1998). however, in the rutgers deployment, the described stratification of water masses of the winter shelf break jet was not present during 3/6 – 3/9, and instead a uniform, unstratified shelf break jet from full column (200 m) mixing was observed (figure 3). significance to stress the significance and importance of monitoring carbonate chemistry in coastal systems, it is important to relate this data to further biological and economic implications. atlantic surfclams and atlantic sea scallops are both calcifying organisms that reside in and are major fisheries of the mab (national marine fisheries, 2021). atlantic surfclams reside nearshore in waters of 10-50m in depth (pousse et al., 2020), while atlantic sea scallops inhabit midto outer-shelf waters 27-80m deep (hart et al., 2004). referring back to the ωarag plots (figure 7, figure 3g, figure 4g), the observed glider data in these essential habitat regions highlight that in winter surfclams are exposed to relatively lower ωarag while sea scallops inhabit waters with higher ωarag due to greater influence from the gulf stream. the difference in habitat range between these species, along with the natural variation in ωarag across the shelf, is important for understanding which species may be more impacted with ongoing ocean acidification and episodic coastal acidification events. furthermore, recent laboratory oa experiments reveal that surfclams have a higher tolerance to more acidic waters (0.57 ωarag, pousse et al., 2020; 1.09 ωarag, meseck et al., 2021) yet may suffer increased energy expenditure and metabolic losses from the acidic stress (pousse et al., 2020). this signifies the importance of understanding the carbonate systems when designing future laboratory oa perturbation experiments for different species based on the naturally occurring variation already present within these environments. 5 conclusions the mab and other coastal zones have various biological, chemical, and physical processes that impact acidity and carbonate chemistry. acidification in coastal zones is further complicated by the influence of these processes varying temporally and spatially. in the winter 2021 survey described here, slope waters of the mab were strongly influenced by the warm saline gulf stream transporting water with a high buffering capacity (high ta) northward. nearshore and shelf waters were more influenced by terrestrial freshwater inputs and the cold, fresher labrador current, leading to higher chlorophyll and oxygen concentrations. opposing patterns in ph and ωarag were reflected in both glider datasets, indicating ph and ωarag were influenced by different drivers. the results described above indicate that ph was more biologically driven while ωarag was more thermodynamically and chemically driven. short-term biological and physical events inaresty rutgers undergraduate research journal, volume i, issue iv fluenced spatial differences in carbonate parameters over the course of deployment. understanding and monitoring these short-term seasonal variations in ph and ωarag is significant for developing experimental designs for understanding and predicting biological responses to acidification based on naturally occurring fluctuations in acidity as well as projected future acidification. moving forward, the winter deployments analyzed here will contribute to a broader project that will deploy ph gliders in the same region seasonally for the next two years. the integration of my results with those from future missions will allow for a better understanding of the variability and drivers of carbonate chemistry over both space and time in this dynamic, economically important region∎ 6 acknowledgements i would like to give special thanks to assistant professor grace saba and dr. elizabeth wrightfairbanks for their careful guidance and support given throughout the research process. i would also like to thank the members of the rutgers center for ocean observing leadership, including the glider technicians (david aragon, nicole waite, chip haldeman), laura nazzaro, hugh roarty, and theodore thompson; and charlie flagg of stony brook university for their work with glider deployment and recovery making this research possible, as well as daphne monroe for reviewing my work. 7 references [1] anderson, d. m., gilbert, p. m., burkholder, j. m. (2002). harmful algal blooms and eutrophication: nutrient sources, composition, and consequences. estuaries, 25, 704–726. accessed from https://www.whoi.edu/cms/files/anderson_etal_2002_estuaries_29903.pdf [2] cai, w.-j., hu, x., huang, w.-j., murrell, m. c., lehrter, j. c., lohrenz, s. e., chou, w.-c., zhai, w., hollibaugh, j. t., wang, y., zhao, p., guo, x., gundersen, k., dai, m., & gong, g.-c. (2011). acidification of subsurface coastal waters enhanced by eutrophication. nature geoscience, 4(11), 766-770. doi:10.1038/ngeo1297 [3] cai, wj., xu, yy., feely, r.a. et al. (2020). controls on surface water carbonate chemistry along north american ocean margins. nat commun 11, 2691. https://doi.org/10.1038/s41467-020-16530-z [4] castelao, r., glenn, s., and schofield, o. (2010). temperature, salinity, and density variability in the central middle atlantic bight, j. geophys. res., 115, c10005, doi:10.1029/2009jc006082. [5] degrandpre, m.d., olbu, g.j., beatty, c.m., hammar, t.r.. (2002). air–sea co2 fluxes on the us middle atlantic bight. deep sea research part ii: topical studies in oceanography, 49(20), 4355-4367. doi: https://doi.org/10.1016/s0967-0645(02)00122-4 [6] fratantoni, p. s., pickart, r. s., torres, d. j., & scotti, a. (2001). mean structure and dynamics of the shelfbreak jet in the middle atlantic bight during fall and winter, journal of physical oceanography, 31(8), 2135-2156. doi: https://doi.org/10.1175/15200485(2001)031%3c2135:msadot%3e2.0.co;2 [7] goldsmith, k. a., lau, s., poach, m. e., sakowicz, g. p., trice, t. m., ono, c. r., nye, j., shadwick, e. h., stlaurent, k. a., & saba, g. k. (2019). scientific considerations for acidification monitoring in the u.s. mid-atlantic region. estuarine, coastal and shelf science 225: 106189, https://doi.org/10.1016/j.ecss.2019.04.023. [8] hart, d. r., chute, a. s., northeast fisheries science center (u.s.). (2004). essential fish habitat source document. sea scallop, placopecten magellanicus, life history and habitat characteristics, noaa technical memorandum nmfs-ne, 189, https://repository.library.noaa.gov/view/noaa/4031. [9] ioos. (2019). manual for real-time quality control of ph data observations: a guide to quality control and quality assurance for ph observations. 13. doi: 10.25923/111k-br08 [10] ipcc. (2019): summary for policymakers. in: ipcc special report on the ocean and cryosphere in a changing climate [h.-o. pörtner, d.c. roberts, v. masson-delmotte, p. zhai, m. tignor, e. poloczanska, k. mintenbeck, m. nicolai, a. okem, j. petzold, b. rama, n. weyer (eds.)]. in press. https://www.ipcc.ch/site/assets/uploads/sites/3/2019/11/03_srocc_spm_final.pdf [11] kerfoot, john. (2014). spt: slocum power tools. github repository, https://github.com/kerfoot/spt.git [12] kroeker, k. j., kordas, r. l., crim, r., hendriks, i. e., ramajo, l., singh, g. s., duarte, c. m., & gattuso, j. p. (2013). impacts of ocean acidification on marine organisms: quantifying sensitivities and interaction with warming. glob chang biol, 19(6), 1884-1896. doi:10.1111/gcb.12179 [13] kwiatkowski, l., & orr, j. c. (2018). diverging seasonal extremes for ocean acidification during the https://www.whoi.edu/cms/files/anderson_etal_2002_estuaries_29903.pdf https://www.whoi.edu/cms/files/anderson_etal_2002_estuaries_29903.pdf https://doi.org/10.1038/s41467-020-16530-z https://doi.org/10.1016/s0967-0645(02)00122-4 https://doi.org/10.1175/1520-0485(2001)031%3c2135:msadot%3e2.0.co;2 https://doi.org/10.1175/1520-0485(2001)031%3c2135:msadot%3e2.0.co;2 https://doi.org/10.1016/j.ecss.2019.04.023 https://repository.library.noaa.gov/view/noaa/4031 https://repository.library.noaa.gov/view/noaa/4031 https://www.ipcc.ch/site/assets/uploads/sites/3/2019/11/03_srocc_spm_final.pdf https://www.ipcc.ch/site/assets/uploads/sites/3/2019/11/03_srocc_spm_final.pdf https://github.com/kerfoot/spt.git aresty rutgers undergraduate research journal, volume i, issue iv twenty-first century. nature climate change, 8(2), 141-145. doi:10.1038/s41558017-0054-0 [14] lewis, e. and wallace, d. w. r. (1998). program developed for co2 system calculations, ornl/cdiac105, carbon dioxide inf. anal. cent., oak ridge natl. lab., oak ridge, tenn., 38 pp., https://salishsea.pnnl.gov/media/ornl-cdiac-105.pdf [15] linder, christopher. (1996). a climatology of the middle atlantic bight shelfbreak front. doi:10.1575/1912/5671 [16] linder, christopher & gawarkiewicz, glen. (1998). a climatology of the middle atlantic bight shelf break front. journal of geophysical research. 103. 97. doi: 10.1029/98jc01438. [17] meseck, s., mercaldo-allen, r., clark, p., kuropat, c., redman, d., veilleux, d., milke, l. (2021). effects of ocean acidification on larval atlantic surfclam (spisula solidissima) from long island sound in connecticut. fishery bulletin, noaa, 119(1). doi:10.7755/fb.119.1.8 [18] nasa goddard space flight center, ocean ecology laboratory, ocean biology processing group. moderate-resolution imaging spectroradiometer (modis) aqua chlorophyll data; 2018 reprocessing. nasa ob.daac, greenbelt, md, usa. doi: data/10.5067/aqua/modis/l3b/chl/2018. [19] national marine fisheries service (2021). fisheries of the united states, 2019. u.s. department of commerce, noaa. current fishery statistics no. 2019 available at: https://www.fisheries.noaa.gov/national/sustainablefisheries/fisheries-united-states [20] pousse, e., poach, m., redman, d., sennefelder, g., white, l., lindsay, j., munroe, d., hart, d., hennen, d., dixon, m., li, y., wikfors, g., meseck, s.. (2020). energetic response of atlantic surfclam spisula solidissima to ocean acidification. marine pollution bulletin, 161(b). doi: https://doi.org/10.1016/j.marpolbul.2020.111740 [21] saba, g.k., goldsmith, k.a., cooley, s.r., grosse, d., meseck, s.l., miller, w., phelan, b., poach, m., rheault, r., st. laurent, k., testa, j., weis, j.s., zimmerman, r. (2019a). recommended priorities for research on ecological impacts of coastal and ocean acidification in the u.s. mid-atlantic. estuarine, coastal and shelf science 225: 106188, https://doi.org/10.1016/j.ecss.2019.04.022. [22] saba g. k., wright-fairbanks, e., chen, b., cai, w-j, barnard, a.h., jones, c.p., branham, c.w., wang k. and miles t. (2019b). the development and validation of a profiling glider deep isfet-based ph sensor for high resolution observations of coastal and ocean acidification. front. mar. sci. 6:664. doi: 10.3389/fmars.2019.00664 [23] sharp, j.d., pierrot, d., humphreys, m.p., epitalon, j.m., orr, j.c., lewis, e.r., wallace, d.w.r. (2020). co2sysv3 for matlab (version v3.1.1). zenodo. http://doi.org/10.5281/zenodo.3950562 [24] van heuven, s., d. pierrot, j.w.b. rae, e. lewis, and d.w.r. wallace. (2011). matlab program developed for co2 system calculations. ornl/cdiac105b. carbon dioxide information analysis center, oak ridge national laboratory, u.s. department of energy, oak ridge, tennessee. https://doi.org/10.3334/cdiac/otg.co2sys_matlab_v1.1 [25] wanninkhof, r., barbero, l., byrne, r., cai, w.-j., huang, w.-j., zhang, j.-z., baringer, m., langdon, c. (2015). ocean acidification along the gulf coast and east coast of the usa. continental shelf research. 98, 54-71. doi: http://dx.doi.org/10.1016/j.csr.2015.02.008 [26] wright‐fairbanks, e. k., miles, t. n., cai, w.‐j., chen, b., saba, g. k. (2020). autonomous observation of seasonal carbonate chemistry dynamics in the mid‐ atlantic bight. journal of geophysical research: oceans, 125, e2020jc016505. https://doi.org/10.1029/2020jc016505 [27] xu, y.-y., cai, w.-j., gao, y., wanninkhof, r., salisbury, j., chen, b., reimer, j. j., gonski, s. and hussain, n. (2017). short-term variability of aragonite saturation state in the central mid-atlantic bight, j. geophys. res. oceans, 122, 4274–4290, doi:10.1002/2017jc012901. [28] xu, y.‐y., cai, w.‐j., wanninkhof, r., salisbury, j., reimer, j., & chen, b. (2020). long‐term changes of carbonate chemistry variables along the north american east coast. journal of geophysical research: oceans, 125, e2019jc015982. https://doi.org/10.1029/2019jc015982 [29] zhang, w. g., and g. g. gawarkiewicz (2015). dynamics of the direct intrusion of gulf stream ring water onto the mid-atlantic bight shelf, geophys. res. lett., 42, 7687–7695, doi:10.1002/2015gl065530 https://salish-sea.pnnl.gov/media/ornl-cdiac-105.pdf https://salish-sea.pnnl.gov/media/ornl-cdiac-105.pdf https://www.fisheries.noaa.gov/national/sustainable-fisheries/fisheries-united-states https://www.fisheries.noaa.gov/national/sustainable-fisheries/fisheries-united-states https://doi.org/10.1016/j.marpolbul.2020.111740 https://doi.org/10.1016/j.marpolbul.2020.111740 https://doi.org/10.1016/j.ecss.2019.04.022 http://doi.org/10.5281/zenodo.3950562 https://doi.org/10.3334/cdiac/otg.co2sys_matlab_v1.1 http://dx.doi.org/10.1016/j.csr.2015.02.008 https://doi.org/10.1029/2020jc016505 https://doi.org/10.1029/2019jc015982 aresty rutgers undergraduate research journal, volume i, issue iv marissa is a recent graduate (class of 2021) of the school of environmental and biological sciences with a bachelor of science in marine biology and a minor in ecology, evolution, and natural resources. her interest in the marine sciences is focused on conservation and environmental research, as climate change threatens to alter our planet she hopes to be a part of research guiding us through it. during her time at the rutgers, marissa worked with gliders under the rutgers center for ocean observing leadership (rucool) assisting in the preparation and deployment of gliders as well as the data analysis of the glider data her sophomore summer into junior year. her senior year, marissa worked under the saba laboratory ocean acidification focus and guided by mentors grace saba and elizabeth wright-fairbanks she developed her honors thesis, using gliders for monitoring winter ocean chemistry. currently, she is working as a fisheries and data analyst for the marine research non-profit group, beyond our shores foundation, gaining experience before potentially pursuing a graduate degree. marissa can be reached at: marissaguzik@gmail.com. mailto:marissaguzik@gmail.com aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. in silico studies of compounds present in azadirachta indica (neem) and their ability to bind hiv integrase protein zhongrui zhang, yin hei lau sonia arora (faculty advisor) ✵ abstract azadirachta indica (neem) is an evergreen tree that belongs to the meliaceae family. it is native to the indian subcontinent and grows worldwide. it is also known as the "village pharmacy" in india for its wide range of therapeutic and pharmacological properties. an in vitro study indicated that a. indica showed anti-hiv properties. however, the exact mechanism for the supposed anti-hiv properties remains unknown. this study aimed to construct an insilico database of the compounds present in a. indica and propose a computational analysis of these compounds against hiv integrase. we performed a thorough literature search to gather relevant information on the plant compounds, including chemical structure, location within the plant, extraction method, and percent yield of each compound found in the plant. we took a comprehensive approach to closely study the binding pockets of hiv integrase and performed molecular docking on a. indica compounds using molecular operating environment. a deductive analysis of the docking energies of these compounds revealed thirty potential binders against hiv integrase proteins. we further validated these binders by comparing the ligand interactions to known inhibitors using ligplot+, which identified the presence of numerous hydrogen bonds and hydrophobic interactions at the protein binding pocket. in conclusion, we propose an underlying binding potential for several a. indica compounds with hiv integrase, yielding a potential mechanism for the anti-hiv activity of a. indica. key terms: neem, anti-hiv, medicinal plant, molecular docking 1 introduction azadirachta indica (family: meliaceae), commonly known as neem, is a large, fast-growing, tropical evergreen tree that has been widely used in traditional medicine since prehistoric times (abdelhady et al., 2015; paul, prasad, & sah, 2011; sadeghian & mortazaienezhad, 2007). a. indica is indigenous to the indian subcontinent and is cultivated in at least 30 countries worldwide (abdelhady et al., 2015). in india, a. indica is also known as "the village pharmacy," "the wonder tree," "nature's drug store," and "the life-giving tree" (hossain et al., 2013; patel et al., 2016; paul et al., 2011). all parts of the tree can be used in disease treatments due to the presence of various phytochemicals. medicinal plants play an important role in the health of human society. a. indica is a medicinal plant with a broad spectrum of therapeutic applications. bioactive phytochemicals, such as flavonoids, terpenoids, tannins, carbohydrates, and proteins, provide a. indica with its healing properties. many unique compounds have been identified and isolated from all parts of a. indica (sarah et al., 2019). azadirachtin, nimbin, gedunin, and quercetin are some of the most studied compounds in a. indica. these compounds carry various biological and pharmacological properties such as antimicrobial, antiviral, antifungal, antimalarial, anti-inflammatory, antiulcer, and anticancer properties (jerobin et al., 2015 and paul et al., 2011). clinical trials have affirmed various therapeutic properties of neem. for example, neem bark extract was found to control gastric hypersecretion as well as gastroesophageal and gastroduodenal ulcers (bandyopadhyay et al., 2004). a recent randomized controlled trial also found the therapeutic potential of neem in preventing covid-19 infection (nesari et al., 2021). therefore, a. indica has been used in the treatment of fearesty rutgers undergraduate research journal, volume i, issue iv ver, malaria, intestinal infections, inflammation, arthritis, and skin diseases (abdelhady et al., 2015; anyaehie, 2009). because of its medicinal properties in disease prevention and treatment, the united nations proclaimed a. indica as the "tree of the 21st century” (hossain et al., 2013). the human immunodeficiency virus (hiv) is a retrovirus that attacks the human immune system. the genetic material of the retrovirus is inserted into the host genome by the retroviral integrase during the process of integration (komal et al., 2020 and smith & daniel, 2006). as a result of a weakened immune system, various symptoms — such as fever, cough, swollen lymph nodes, mouth ulcers, and muscle aches — can develop. if hiv is not treated properly, it can lead to severe diseases such as tuberculosis, cryptococcal meningitis, cancers, and acquired immunodeficiency syndrome (aids) (world health organization [who], 2021; centers for disease control and prevention [cdc], 2021). hence, urgent treatment is necessary for hiv patients. according to the who, there were approximately 37.7 million hiv cases in 2020. however, due to the development of resistance to current medication targeting integrase, the investigation of new integrase inhibitors is needed (mesplède et al., 2012). in a previous in vitro study conducted by udeinya et al. (2004), a fractionated acetone-water extract prepared from a. indica showed anti-cytoadhesion activity, which protects lymphocytes against invasion by hiv and suggests the anti-retroviral property of this plant. in-silico screening, or computer-aided drug design (cadd), has become a crucial part of the modern drug discovery process. it uses a variety of bioinformatics applications and algorithms to efficiently screen for potential drug candidates and significantly reduce the time and resources needed in the traditional lab-bench-based drug delivery process (rodrigues and schneider, 2015). in addition, the aforementioned algorithms can be used to predict the pharmacological properties and interactions of molecules. molecular docking is a type of cadd that predicts the protein-ligand interaction between the drug target and the drug candidate. it runs computer simulations of the potential drug candidates (ligands) with different 3d postures interacting with the drug targets (proteins) and measures the favorableness of such interactions in terms of binding energy. such predictions could therefore be used as the first step of drug candidate screening and can eliminate unlikely candidates within a relatively short time frame using fewer resources. despite recent findings on a. indica’s antihiv potential, the exact mechanism of action is still unknown. previous studies from our lab have focused on several anti-hiv targets such as hiv protease and reverse transcriptase. however, none of these studies have led to conclusive data (unpublished observations). therefore, this study aimed to construct an in-silico database of the compounds present in a. indica and propose a computational analysis of these compounds against hiv integrase — one of the important proteins in the hiv life cycle — to investigate the potential inhibitory activity in reducing viral load. 2 methods literature review and database building a literature search on a. indica was first conducted to collect common compounds present in this plant with readily available structures. for each of these compounds, the percent yield and the location within the plant containing the highest abundance of compounds were also collected. the three-dimensional (3d) structures of the compounds were collected on pubchem (kim et al., 2020) (sdf format) or chemspider (mol format). for compounds without readily available 3d structures on these public sources, the 2d structures were collected and converted to 3d models using discovery studio (d.s.) visualizer (biovia & dassault systèmes, 2017). hydrogen atoms were added to all 3d compound structures. the geometry of each compound was cleaned using the built-in minimize structure tool in ucsf chimera (pettersen et al., 2004) to reduce the internal energies. all optimized a. indica compounds were saved in a mol2 format and ready for molecular docking. aresty rutgers undergraduate research journal, volume i, issue iv protein visualization and optimization the 3d protein structures of hiv integrase were retrieved from the rcsb protein data bank to serve as the target model of investigation (berman et al., 2000). the human protein structures with higher resolution and known bounded ligands were prioritized during the collection process. minor protein processing was performed to optimize the protein model and minimize errors. all selected proteins underwent energy optimization and geometry cleaning with the structural preparation and protonate 3d tool in molecular operating environment (moe) ("molecular operating environment (moe), 2019.01," 2022), and the processed proteins were saved in moe format. a pre-docking binding pocket analysis was also performed using ligplot+(laskowski & swindells, 2011) to collect baseline protein-ligand interactions between the known inhibitor compounds and the hiv integrase proteins. docking parameter optimization and baseline building the minimized hiv integrase protein files were re-docked using moe docking. re-docking was performed as a suitability control experiment, which involved taking out the originally bounded ligands in each protein file and docking them back into the protein binding pocket with various binding parameters to identify the most optimal conditions for the experiment. the default placement method of triangle matcher with london dg scoring system and the refinement method of rigid receptor with gbvi/wsa dg scoring system were used as the docking methods. these methods and scoring systems are known to give reliable results, and the results were estimated in terms of free energy reported in kcal/mol (corbeil et al., 2012; galli et al., 2014). however, other parameters such as receptor region, docking site, and number of docked poses were tested for the most optimized parameter combinations for each protein target, which was measured by the rmsd values between the original ligand and the re-docked models. the docking energies for the most suitable docking parameters (often resulted in the lowest rmsd values) were also recorded for baseline purposes. a. indica compound docking a moe database file in mdb format was created with the name and the structure in mol2 format of each optimized a. indica compound. the optimized docking parameters were used albeit the moe compound database file, which was used as the docking ligands. the docked poses with the best docking energies (most negative) for each compound were recorded. for each a. indica compound, the average docking energies against all protein targets were compared with the average docking energies in optimization. since the docking energy was measured in free energy, any a. indica compounds with more negative average docking energies than optimization was more thermodynamically favorable to bind and therefore identified as potential hiv integrase binders. post-docking analysis ligplot+ was used to perform a post-docking analysis on all binders in complex with their protein binding pockets. two of the most common and relatively strong protein-ligand interactions — hydrogen bonds and hydrophobic interactions — were examined to help explain and verify the favorable docking energies obtained by the binder compounds. 3 results compound database building a comprehensive database comprised of 50 compounds present in a. indica was created. figure 1 shows the structures of a few compounds present in the database. in addition to the 3d structures of each compound, the database also collected the locations where these compounds are found on the plant, a brief categorization of each compound, the extraction methods, and the corresponding percent yield found in the plant (table 1). aresty rutgers undergraduate research journal, volume i, issue iv azadirachtin nimbin gedunin nimbolide nimbolin salannin cycloeucalenol nimbosterol nimbolicin figure 1. the 3d structures of nine representative a. indica compounds collected in the database. grey: carbon; white: hydrogen; red: oxygen. protein data collection and binding pocket analysis of hiv integrase four hiv integrase proteins (pdb ids: 1qs4, 3nf6, 3nf7, 6wc8) were selected (goldgur et al. 1999; gorman et al. 2020; peats et al. 2010). these pdb files had the highest resolution, were derived from human targets, and contained at least one known hiv integrase inhibitor. clustal omega multiple protein sequence alignment revealed that these selected integrase proteins were at least 95% identical between any two proteins (sievers et al. 2011). the difference in protein sequence was due to the presence of 2-4 unique mutation sites in each protein. the number of files used in this study was a balance between accuracy and resource, as previous studies from our lab had demonstrated that using four target files was sufficient to generate reliable results. binding pocket analysis of each of the four proteins revealed the presence of numerous hydrogen bonds and hydrophobic interactions between the known inhibitor ligands and the integrase proteins, as shown in table 2. aresty rutgers undergraduate research journal, volume i, issue iv table 1: in-silico compound database for a. indica. all data was collected from publicly available journal sources as indicated in reference column. the percent yields corresponded to the extraction method as listed. n/a: data not available. compound source extraction method percent yield reference 1 azadirachtin flower, fruit, leaf, and seed aqueous 0.1-0.3% in seed (biswas et al, 2002; kaushik, 2021; morgan, 2009; paul et al., 2011; ponnusamy et al., 2015; sadeghian & mortazaienezhad, 2007; singh et al., 2017) 2 isomargolonone bark n/a n/a (biswas et al., 2002; singh et al., 2017) 3 azadiradione fruit, leaf, seed aqueous 0.3% in leaf (paul et al., 2011; ponnusamy et al., 2015; sadeghian & mortazaienezhad, 2007) 4 epicatechin bark n/a n/a (biswas et al., 2002; singh et al., 2017) 5 mahmoodin seed oil n/a n/a (biswas et al., 2002) 6 azadirone fruit, leaf, seed oil aqueous 2.46% in leaf (paul et al., 2011; ponnusamy et al., 2015; sadeghian & mortazaienezhad, 2007) 7 flavanone flower n/a n/a (nakahara et al., 2003) 8 margolone bark n/a n/a (biswas et al., 2002; singh et al., 2017) 9 catechin bark n/a n/a (biswas et al., 2002; singh et al., 2017) 10 nimbin leaf, seed oil, trunk and root bark aqueous 2.6% in leaf (biswas et al., 2002; kaushik et al., 2021; paul et al., 2011; ponnusamy et al., 2015; sadeghian & mortazaienezhad, 2007; singh et al., 2017) 11 gedunin leaf, seed oil n/a n/a (anand, 2017; biswas et al., 2002; paul et al., 2011; ponnusamy et al., 2015; sadeghian & mortazaienezhad, 2007; singh et al., 2017) 12 nimbinin leaf, seed oil, trunk and root bark n/a n/a (koul, isman, & ketkar, 1990; paul et al., 2011) 13 nimbolide leaf, seed oil aqueous 2.20% leaf (biswas et al., 2002; kaushik et al., 2021; sadeghian & mortazaienezhad, 2007; singh et al., 2017) 14 nimbidin leaf, seed n/a n/a (biswas et al., 2002; koul et al., 1990; singh et al., 2017) 15 nimbolin a trunk wood n/a n/a (paul et al., 2011) 16 nimbolin b trunk wood n/a n/a (paul et al., 2011) 17 quercetin flower and leaf n/a n/a (kaushik et al., 2021; paul et al., 2011) 18 salannin leaf, seed oil aqueous 5.6% in leaf (paul et al., 2011; ponnusamy et al., 2015; sadeghian & mortazaienezhad, 2007) 19 nimbidol leaf n/a n/a (anand, 2017) 20 cycloeucalenol wood oil n/a n/a (paul et al., 2011) aresty rutgers undergraduate research journal, volume i, issue iv 21 nimbosterol (beta-sitosterol) leaf, wood oil n/a n/a (kaushik et al., 2021) 22 nimbinone bark n/a n/a (ara, siddiqui, faizi, & siddiqui, 1988) 23 nimbolicin bark n/a n/a (read & french, 1993) 24 margocin root bark n/a n/a (ara et al., 1990) 25 gallic acid bark n/a n/a (biswas et al., 2002; singh et al., 2017) 26 2-methyl-5-ethylfuran leaf butanol 4.8273% (hossain et al., 2013) 27 arabinose bark n/a n/a (kumar et al., 2017) 28 m-toluylaldehyde leaf methanol 22.7669% (hossain et al., 2013) 29 2-methyl-benzaldehyde leaf butanol 11.8674% (hossain et al., 2013) 30 levoglucosenone leaf butanol 7.1217% (hossain et al., 2013) 31 methyl isoheptadecanoate leaf hexane chloroform methanol 2.1921% 11.6299% 12.2749% (hossain et al., 2013) 32 methyl petroselinate leaf hexane 11.2380% (hossain et al., 2013) 33 phytol leaf hexane ethyl acetate chloroform 2.6170% 61.2401% 10.0515% (hossain et al., 2013) 34 butyl palmitate leaf hexane 6.6981% (hossain et al., 2013) 35 isobutyl stearate leaf hexane 4.2521 % (hossain et al., 2013) 36 oxalic acid leaf hexane 13.7094% (hossain et al., 2013) 37 methyl 14-methylpentadecanoate leaf methanol ethyl acetate chloroform butanol 38.1251% 6.4278% 31.8674% 13.4471% (hossain et al., 2013) 38 hexahydrofarnesyl acetone leaf ethyl acetate 2.5888% (hossain et al., 2013) 39 lineoleoyl chloride leaf methanol chloroform butanol 26.8329% 11.3587% 13.6057% (hossain et al., 2013) 40 nonacosane leaf chloroform butanol 20.6575% 12.8752% (hossain et al., 2013) 41 stearic acid kernel oil n/a 18% (do et al., 2022 ) 42 palmitic acid kernel oil n/a 16.9% do et al., 2022 ) 43 oleic acid kernel oil n/a 45.9% (do et al., 2022 ) 44 linoleic acid kernel oil n/a 15.69% (do et al., 2022 ) 45 pyroligneous acid heartwood n/a 38.4% (kumar et al., 2017) 46 hentriacontane leaf butanol 13.9887 (hossain et al., 2013) 47 heptacosane leaf hexane 8.1010% (hossain et al., 2013) 48 octacosane leaf hexane 7.0926 (hossain et al., 2013) 49 eicosane leaf hexane 10.0136 (hossain et al., 2013) 50 nonadecane leaf hexane 3.7587% (hossain et al., 2013) aresty rutgers undergraduate research journal, volume i, issue iv table 2: binding pocket analysis of the four selected hiv integrase proteins. the hydrogen bonds and hydrophobic interactions were identified using ligplot+. all ligands present in these proteins had shown existing binding activity toward hiv integrase. 100: 1-(5-chloroindol-3-yl)-3-hydroxy-3-(2h-tetrazol-5-yl)-propenone imv: 5-[(2-oxo-2,3-dihydro-1h-indol-1-yl)methyl]-1,3-benzodioxole-4-carboxylic acid ciw: 5-[(5-chloro-2-oxo-2,3-dihydro-1h-indol-1-yl)methyl]-1,3-benzodioxole-4-carboxylic acid tqm: {5-(3-fluorophenyl)-2-[(thiophen-2-yl)ethynyl]-1-benzofuran-3-yl}acetic acid pdb # ligand hydrogen bond hydrophobic 1qs4 100 (5cltep) thr66 asp64 asn155 lys159 gln148 lys156 ile151 glu152 3nf6 imv glu170 gln95 ala128 lys173 his171 tyr99 ala129 met178 thr174 leu102 trp132 thr125 ala169 3nf7 ciw val77 val150 leu158 val79 ser153 his183 gly82 met154 ile84 glu157 6wc8 tqm gln95 leu102 glu96 ala128 ala98 ala129 tyr99 trp132 docking parameter optimization and baseline building docking optimization was performed on each of the four hiv integrase targets. the redocked models were compared with the original ligands to identify the best docking parameters to use for a. indica compounds. after the optimization process, the protein atoms without the surrounding solvent were set as the docking receptor. the protein residues within the 5å space of the original ligand were defined as the protein active sites for docking. thirty placement poses and five refinement poses were deemed the best parameters for later studies. for each of the four protein targets, these parameters yielded rmsd values of 1.99, 0.35, 0.73, 1.63, respectively, which were low enough to generate accurate docked results (figure 2). the average docking energy of redocked ligands was -5.6358 kcal/mol. molecular docking of a. indica compounds into hiv integrase binding pocket molecular docking was performed on all 50 a. indica compounds against each of the four hiv integrase proteins using the optimized docking parameters. after a comparison between the average docking energy for the plant compounds and the redocked ligands, 30 a. indica compounds were predicted to have a more favorable binding energy and were identified as potential hiv integrase binders (table 3). aresty rutgers undergraduate research journal, volume i, issue iv figure 2: ligand-bounded hiv integrase binding pocket (pdb# 3nf7). (a) docking optimization with original ligand (green) and redocked ligand (pink). (b) a. indica compound salannin (blue) and gedunin (orange). table 3: the docking energy for each of the a. indica compounds and the redocked ligands. each of these energy measures was the average docking energy of each compound against all four hiv integrase proteins (pdb# 1qs4, 3nf6, 3nf7, 6wc8). highlighted compounds were identified as binders. * redocked ligands from docking optimization. compound average docking energy (kcal/mol) compound average docking energy (kcal/mol) 0 redocked* -5.6358 26 2-methyl-5-ethylfuran -4.1258 1 azadirachtin -6.3591 27 arabinose -4.1509 2 isomargolonone -5.3900 28 m-toluylaldehyde -4.1068 3 azadiradione -5.4921 29 2-methyl-benzaldehyde -3.9820 4 epicatechin -5.2782 30 levoglucosenone -3.5953 5 mahmoodin -5.9337 31 methyl isoheptadecanoate -6.2995 6 azadirone -5.6070 32 methyl petroselinate -6.4152 7 flavanone -5.0024 33 phytol -6.4347 8 margolone -5.3343 34 butyl palmitate -6.5009 9 catechin -5.2375 35 isobutyl stearate -6.7462 10 nimbin -6.0113 36 oxalic acid -3.0358 11 gedunin -5.8379 37 methyl 14-methylpentadecanoate -6.3077 12 nimbinin -5.6339 38 hexahydrofarnesyl acetone -6.0264 13 nimbolide -5.6445 39 lineoleoyl chloride -6.2090 14 nimbidin -5.4625 40 nonacosane -7.4238 15 nimbolin a -6.6793 41 stearic acid -6.3773 16 nimbolin b -6.3832 42 palmitic acid -6.1378 17 quercetin -4.9746 43 oleic acid -6.2499 18 salannin -5.9750 44 linoleic acid -6.3096 19 nimbidol -5.1154 45 pyroligneous acid -3.3098 20 cycloeucalenol -6.0074 46 hentriacontane -7.3965 21 nimbosterol -6.3417 47 heptacosane -7.0865 22 nimbinone -5.1792 48 octacosane -7.3098 23 nimbolicin -6.4237 49 eicosane -6.4206 24 margocin -5.4672 50 nonadecane -6.1945 25 gallic acid -4.2497 figure 2a figure 2b aresty rutgers undergraduate research journal, volume i, issue iv post-docking analysis a post-docking analysis was performed on all thirty a. indica binders against each of the four protein targets. the ligplot+ images revealed the intermolecular interactions of these binders in complex with the hiv integrase binding pockets (figure 3). most of the binders were shown to be surrounded by large, hydrophobic clusters. some hydrogen bonds were also observed with some a. indica compounds. the common interacting residues of a few representative compounds are shown in table 4. figure 3: representative figure of the post-docking analysis on a. indica compound bounded hiv integrase proteins. semi-circles indicate protein residue involved in hydrophobic interactions. arrow-pointed orange compounds indicate protein residue involved in hydrogen bonds. the purple compounds are a. indica compounds of interest. aresty rutgers undergraduate research journal, volume i, issue iv table 4: the common interacting hiv integrase residues of a few representative a. indica binders and the original redocked ligand via either hydrophobic interactions or hydrogen bonds. *residues involved in hydrogen bonds. compound average docking energy (kcal/mol) common interacting residues re-docked -5.6358 val77, val79, gly82, ile84, val150, ser153, met154, glu157, leu158 azadirachitin -6.3591 val79, gly82, val150, ser153, met154, glu157, his183, lys188*, arg199* isobutyl stearate -6.7462 val77, val79, gly82, val150, met154, glu157, his183, lys186, lys188 nimbolicin -6.4237 val77, val79, ala80*, gly82, val150, met154, glu157, his183, lys186, lys188, arg199 nimbolide -5.6445 ser81, gly82, val150, ser153, met154, glu157, his183, lys188*, arg199* nimbolin a -6.6793 val77, val79, ala80, gly82, val150, ser153, met154, glu157, his183, lys188, arg199 nimbosterol -6.3417 val77, val79, gly82, val150, met154, glu157, his183, lys186, lys188 octacosane -7.3098 val77, gly82, val150, ser153, met154, glu157, lys188 salannin -5.975 ser153, met154, glu157, his183, lys186*, lys188*, arg199 4 discussion & conclusion a comprehensive in-silico database of a. indica compounds was created, providing detailed information on many compounds regarding their sources, extraction methods, and percent yields extracted from the plant. although the exact percent yield of these compounds varies greatly depending on the extraction method, this information could provide valuable insights for later drug discovery stages. the 3d structures collected for each of these compounds were also extensively used in the molecular docking studies against hiv integrase. in the docking optimization process, the original ligands in each protein file were docked back into the protein binding pockets; the resulting model was referred to as re-docked ligands. the relative position and identity of the redocked ligands were visually compared with the original ligands to determine the reliability of the docking methods. the re-docked ligands were shown to occupy a highly similar 3d space with the original ligand. this was also quantified via the rmsd values, which measured the average distance between the atoms of the original ligand and the re-docked ligand. therefore, the low rmsd values also reflected highly similar postures between the predicted model and the original ligand. both verification methods indicated that the optimized docking parameters and algorithm were highly accurate in predicting the binding affinity of the a. indica compounds. aresty rutgers undergraduate research journal, volume i, issue iv meanwhile, the average docking energy in optimization was also an important baseline for identifying the potential a. indica binders. a closer look at the 30 identified a. indica binders revealed that they were bound at the same binding pocket as the original ligands, suggesting a potentially similar allosteric effect. interestingly, the average docking energies of six compounds (nonacosane, hentriacontane, octacosane, heptacosane, isobutyl stearate, and nimbolin a) were one standard deviation more favorable than the redocked ligands, indicating more efficient binding activity than the original ligands. post-docking analysis was performed to explain and validate the favorable docking energies predicted by the docking algorithm. the large hydrophobic clusters surrounding most a. indica binders and the presence of hydrogen bonds with some binders were both excellent indicators of strong intermolecular interactions. some interacting residues in the original ligands, such as val77, val79, gly82, val150, ser153, met154, glu157, and his183, were commonly retained across many a. indica binders. in addition to the retained interactions, most binders also gained new hydrophobic interactions; some of these binders, such as azadirachitin and nimbolide, also gained new hydrogen bonds. the a. indica binders that gained new interactions suggest a more potent binding ability to the target proteins compared to the corresponding binding ability of the original ligands. therefore, the intermolecular interactions present in these new protein-ligand complexes validated the favorable binding energies predicted by the docking algorithms, which supported the identified a. indica binders against hiv integrase. overall, we have identified 30 out of 50 a. indica compounds as binders of hiv integrase proteins. the large proportion of the binder compounds present in this plant provides a feasible explanation of this plant's hiv viral reducing potential. therefore, we propose a potential mechanism for the anti-hiv activity for a. indica which could offer insights into a novel hiv treatment candidate. however, the fact that the docking energies of the existing known binders were used as the cut-off point in a. indica binder identification may be a potential limitation this study. the possibility of the a. indica compounds with less-than-ideal docking energies binding to the targets still exists. this study only provides a computational screening of the potential hiv drug candidates; further bench testing on promising candidates is still necessary to validate the results. therefore, future goals include further testing these compounds in a wet lab setting to validate the potential inhibitory potential against hiv integrase. the database created in this study may play an important role in future studies of this plant compared to other biological targets, which in turn enables exploration of other therapeutic targets∎ 5 acknowledgements we would like to express our sincere gratitude to our mentor and research advisor dr. sonia arora for her continuous support throughout the duration of the project. this research work would not be possible without her enthusiasm and knowledge to the topic. her guidance was always inspiring and this project was a great learning opportunity on in-silico approaches, which could have endless applications in drug discovery works. 6 references [1] abdelhady, m. i. s., bader, a., shaheen, u., elmalah, y., abourehab, m. a. s., & barghash, m. f. (2015). azadirachta indica as a source for antioxidant and cytotoxic polyphenolic compounds. biosciences biotechnology research asia, 12(2), 1209-1222. doi:10.13005/bbra/1774 [2] alzohairy, m. a. (2016). therapeutics role of azadirachta indica (neem) and their active constituents in diseases prevention and treatment. evid based complement alternat med., 2016. doi:10.1155/2016/7382506 [3] anand, n. (2017). antifungal, contraceptive, anti-cancer, mosquito repellent properties of aresty rutgers undergraduate research journal, volume i, issue iv azadirachta indica: a review. texila international journal of basic medical sciences, 2(1). doi:10.21522/tijbms.2016.02.01.art006 [4] anyaehie, u. b. (2009). medicinal properties of fractionated acetone/water neem [azadirachta indica] leaf extract from nigeria: a review. niger j physiol sci., 24(2). doi:10.4314/njps.v24i2.52926 [5] ara, i., siddiqui, b. s., faizi, s., & siddiqui, s. (1988). terpenoids from the stem bark of azadirachta indica. phytochemistry, 27(6), 18011804. doi:10.1016/0031-9422(88)80447-3 [6] ara, i., siddiqui, b. s., faizi, s., & siddiqui, s. (1990). tricyclic diterpenoids from root bark of azadirachta indica. phytochemistry, 29(3), 911914. doi:10.1016/0031-9422(90)80044-h [7] bandyopadhyay, u., biswas k., sengupta, a., moitra, p., dutta, p., sarkar, d., debnath, p., ganguly, c. k. & banerjee, r. k. (2004). clinical studies on the effect of neem (azadirachta indica) bark extract on gastric secretion and gastroduodenal ulcer. life sci, 75(24):2867-78. [8] berman, h. m., westbrook, j., feng, z., gilliland, g., bhat, t. n., weissig, h., . . . bourne, p. e. (2000). the protein data bank. nucleic acids research, 28, 235-242. [9] biovia, & dassault systèmes. (2017). discovery studio visualizer. san diego: dassault systèmes. [10] biswas, k., chattopadhyay, i., banerjee, r. k., & bandyopadhyay, u. (2002). biological activities and medicinal properties of neem (azadirachta indica). current science, 82(11), 1336-1345. [11] centers for disease control and prevention [cdc]. (2021). about hiv. retrieved from https://www.cdc.gov/hiv/basics/whatishiv.html [12] corbeil, c. r., williams, c. i., & labute, p. (2012). variability in docking success rates due to dataset preparation. journal of computer-aided molecular design, 26(6), 775–786. https://doi.org/10.1007/s10822-012-9570-1 [13] council, n. r. (1992). neem: a tree for solving global problems. washington, dc: the national academies press. [14] do, d. n., pham, h. d., le, x. t., le, m. t., nguyen, n. p., bui, q. m. & luu, x. c. (2022).mechanical extraction of neem seed kernel (azadirachta indica a. juss.) harvesting in ninh thuan, viet nam by using hydraulic pressing: effect of processing parameters. materials science forum, 1048, 437-444. [15] galli cl, sensi c, fumagalli a, parravicini c, marinovich m, eberini i. a computational approach to evaluate the androgenic affinity of iprodione, procymidone, vinclozolin and their metabolites. plos one. 2014 aug 11;9(8):e104822. doi:10.1371/journal.pone.0104822. [16] goldgur, y., craigie, r., fujiwara, t., yoshinaga, t., davies, d.r. (1999). core domain of hiv-1 integrase complexed with mg++ and 1-(5chloroindol-3-yl)-3-hydroxy-3-(2h-tetrazol-5yl)propenone. doi: 10.2210/pdb1qs4/pdb [17] gorman, m.a. & parker, m.w. (2020). hiv integrase core domain in complex with inhibitor 2(5-(3-fluorophenyl)-2-(2-(thiophen-2-yl)ethynyl)1benzofuran-3-yl)ethanoic acid. doi:10.2210/pdb6wc8/pdb [18] hossain, m. a., al-toubi, w. a. s., weli, a. m., alriyami, q. a., & al-sabahi, j. n. (2013). identification and characterization of chemical compounds indifferent crude extracts from leaves of omani neem. journal of taibah university for science, 7(4), 181-188. doi:10.1016/j.jtusci.2013.05.003 [19] jerobin, j., makwana, p., kumar, r. s., sundaramoorthy, r., mukherjee, a., & chandrasekaran, n. (2015). antibacterial activity of neem nanoemulsion and its toxicity assessment on human lymphocytes in vitro. int j nanomedicine, 10(suppl 1), 77-86. doi:10.2147/ijn.s79983 [20] kaushik, p., ahlawat, p., singh, k., & singh, r. (2021). chemical constituents, pharmacological activities, and uses of common ayurvedic medicinal plants: a future source of new drugs. advances in traditional medicine. doi:10.1007/s13596-021-00621-3 [21] kim, s., chen, j., cheng, t., gindulyte, a., he, j., he, s., . . . bolton, e. e. (2020). pubchem in 2021: new data content and improved web interfaces. nucleic acids research, 49(d1), d1388-d1395. doi:10.1093/nar/gkaa971 [22] komal, d., khushboo, j., aaftaab, s., lakshmi, s., & mallika, a. (2020). targeting integrase enzyme: a therapeutic approach to combat hiv resistance. mini-reviews in medicinal chemistry, 20(3). doi:10.2174/1389557519666191015124932 [23] koul, o., isman, m. b., & ketkar, c. m. (1990). properties and uses of neem, azadirachta indica. can. j. bot., 68(1), 1-11. doi:10.1139/b90-001 [24] kumar, b., bhaskar, d., rajadurai, m., & sathyamurthy, b. (2017). in vitro studies on the effect of azadirachta indica linn. in lung cancer a549 https://www.cdc.gov/hiv/basics/whatishiv.html https://doi.org/10.1007/s10822-012-9570-1 aresty rutgers undergraduate research journal, volume i, issue iv cell lines. world j pharm pharm sci, 6(9), 16271640. [25] laskowski, r. a., & swindells, m. b. (2011). ligplot+: multiple ligand-protein interaction diagrams for drug discovery. journal of chemical information and modeling, 51, 2778-2786. [26] mesplède, t., quashie, p. k. & wainberg, m. a. (2012). resistance to hiv integrase inhibitors. curr opin hiv aids, 7(5), 401-408. doi:10.1097/coh.0b013e328356db89 [27] molecular operating environment (moe), 2019.01. (2022). chemical computing group ulc, 1010 sherbooke st. west, suite #910, montreal, qc, canada, h3a 2r7. [28] morgan, e. d. (2009). azadirachtin, a scientific gold mine. bioorganic & medicinal chemistry, 17(12), 4096-4105. doi:10.1016/j.bmc.2008.11.081 [29] nakahara, k., roy, m. k., ono, h., maeda, i., ohnishi-kameyama, m., yoshida, m., & trakoontivakorn, g. (2003). prenylated flavanones isolated from flowers of azadirachta indica (the neem tree) as antimutagenic constituents against heterocyclic amines. j agric food chem., 51(22), 6456-6460. doi:10.1021/jf034666z [30] nesari, t. m., bhardwaj, a., shrikrishna, r., ruknuddin, g., ghildiyal, s., das, a., pandey, a. k., chaudhary, n., soman, g. & barde, m. (2021). neem (azadirachta indica a. juss) capsules for prophylaxis of covid-19 infection: a pilot, double-blind, randomized controlled trial. altern ther health med, 27(s1):196-203. [31] patel, s. m., venkata, k. c. n., bhattacharyya, p., sethi, g., & bishayee, a. (2016). potential of neem (azadirachta indica l.) for prevention and treatment of oncologic diseases. semin in cancer biol., 40-41, 100-115. doi:10.1016/j.semcancer.2016.03.002 [32] paul, r., prasad, m., & sah, n. k. (2011). anticancer biology of azadirachta indica l (neem): a mini review. cancer biol ther., 12(6), 467-476. doi:10.4161/cbt.12.6.16850 [33] peat, t.s., newman, j., deadman, j.j., rhodes, d. (2010). structural basis for a new mechanism of inhibition of hiv integrase identified by fragment screening and structure based design. doi:10.2210/pdb3nf7/pdb [34] pettersen, e., goddard, t., huang, c., couch, g., greenblatt, d., meng, e., & ferrin, t. (2004). ucsf chimera--a visualization system for exploratory research and analysis. journal of computational chemistry, 25(13), 1605-1612. [35] ponnusamy, s., haldar, s., mulani, f., zinjarde, s., thulasiram, h., & ravikumar, a. (2015). gedunin and azadiradione: human pancreatic alpha-amylase inhibiting limonoids from neem (azadirachta indica) as anti-diabetic agents. plos one., 10(10). doi:10.1371/journal.pone.0140113 [36] read, michael d. & french, h. james h., eds. (1993). genetic improvement of neem: strategies for the future. proc. of the international consultation on neem improvement held at kasetsart university, bangkok, thailand, 18 22 january 1993. bangkok, thailand: winrock international. 194 + x pp. [37] rodrigues, t., schneider, g. (2015). chapter 6 in silico screening: hit finding from database mining. the practice of medicinal chemistry, 4, 141-160. https://doi.org/10.1016/b978-0-12-417205-0.00006-7 [38] sadeghian, m. m., & mortazaienezhad, f. (2007). investigation of compounds from azadirachta indica (neem). asian journal of plant sciences, 6, 444-445. doi:10.3923/ajps.2007.444.445 [39] sarah, r., tabassum, b., idrees, n., & hussain, m. k. (2019). bioactive compounds isolated from neem tree and their applications. in natural bio-active compounds: springer. [40] sievers, f., wilm, a., dineen, d., gibson, t. j, karplus, k., li, w., lopez, r., mcwilliam, h., remmert, m., söding, j., thompson, j. d, higgins, d. g, (2011) fast, scalable generation of high-quality protein multiple sequence alignments using clustal omega. molecular systems biology, 7. 539. doi: accession:10.1038/msb.2011.75 [41] singh, h., kaur, m., dhillon, j. s., batra, m., & khurana, j. (2017). neem: a magical herb in endodontics. stomatological dis sci, 1, 50-54. doi:10.20517/2573-0002.2016.10 [42] smith, j. a., & daniel, r. (2006). following the path of the virus: the exploitation of host dna repair mechanisms by retroviruses. acs chem. biol., 1(4), 217-226. doi:10.1021/cb600131q [43] udeinya, i. j., mbah, a. u., chijioke, c. p., & shu, e. n. (2004). an antimalarial extract from neem leaves is antiretroviral. transactions of the royal society of tropical medicine and hygiene, 98(7), 435-437. doi:10.1016/j.trstmh.2003.10.01 https://doi.org/10.1016/b978-0-12-417205-0.00006-7 aresty rutgers undergraduate research journal, volume i, issue iv [44] world health organization [who]. (2021). hiv/aids. retrieved from https://www.who.int/news-room/fact-sheets/detail/hiv-aids zhongrui zhang is a recent graduate from rutgers university-new brunswick. he has a b.s. in biotechnology, bioinformatics from the school of environmental and biological science. he had been conducting research in dr. sonia arora’s lab on two projects over the course of two years. one was to utilize in-silico techniques to study the compounds present in ocimum sanctum against inflammatory pathways. the other was to investigate the potential anti-hiv activity of azadirachta indica in reducing viral loads. he also assisted phd students in dr. james simon's lab in conducting synthetic and analytical organic chemistry work, where he gained hands-on experience working with plant compounds. these research experiences sparked his interest in drug discovery, and he is currently working at bristol myers squibb within the biologics department. in the future, zhongrui would like to pursue graduate studies in the drug development-related field. zhongrui can be reached at zhongrui.zhang@rutgers.edu. yin hei lau is a graduate of rutgers university. she has a b.s. degree in biotechnology – bioinformatics from the school of environmental and biological science. she has a broad interest in health and medicine and would like to conduct further research in the medical field. her research, under the guidance of dr. sonia arora, investigated the anti-hiv properties of compounds in a. indica through in-silico approach. she also worked in dr. judith storch's research lab for more than two years. she assisted a phd student in functional analysis of enterocyte fatty acid binding proteins (fabp). for independent projects, she studied the hepatic lipid metabolism in the intestine-specific liver fabp (lfabp) knockout mice and the intestinal lipid metabolism in the liver-specific lfabp knockout mice. yin hei can be contacted at: yinhei.lau@rutgers.edu. https://www.who.int/news-room/fact-sheets/detail/hiv-aids https://www.who.int/news-room/fact-sheets/detail/hiv-aids mailto:zhongrui.zhang@rutgers.edu mailto:yinhei.lau@rutgers.edu aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. making a dent in the obesity equation via coupling sugar sweetened beverage taxes with fruit and vegetable subsidies sarah pfreundschuh ✵ abstract obesity rates continue to rise in children and adults alike in the united states and represent a significant threat to public health and economic well-being. many factors have contributed to the obesity equation, including the widespread availability and appeal of ultra-processed food and drink. sugar-sweetened beverages (ssbs) represent one such drink, as a critical examination of the available evidence reveals a clear link between their consumption and increased risks of obesity and related conditions such as type 2 diabetes. taxing ssb purchases therefore presents a potentially valuable means of making a dent in the contribution of one key risk factor to the obesity equation, though the beverage industry has fought against the enactment of these taxes and has instead promoted a generally unclear public health stance on ssbs. this paper explores existing ssb excise taxes that have been implemented in recent years, focusing especially on philadelphia’s tax as a case study for examining the behavioral changes associated with ssb taxes and the management of ssb tax revenue. it then suggests that ssb tax revenue be directly funneled into the subsidization of fruits and vegetables to maximize the obesity-fighting potential of these relatively novel excise taxes. 1 introduction since the late 1970s, obesity has been steadily rising in the united states (office of the surgeon general, 2010; dixon, 2020), with its prevalence reaching 42.4% among adults ages 20 and over from 2017-2018 (cdc3, 2022). even amongst american children, 1 in 5 were obese as of 2018 (cdc, 2022). obesity is a significant risk factor for a multitude of chronic health conditions, including type 2 diabetes, coronary heart disease, hypertension, high ldl cholesterol, low hdl cholesterol, dyslipidemia, gallbladder disease, and osteoarthritis (cdc1, 2022; must et al., 1999). these chronic health conditions reduce the quality of life of millions of americans and present an enormous economic burden, amounting to nearly $173 billion annually in obesity-related medical costs (ward et. al., 2021). additional economic costs stem from the lost productivity associated with overweight and obese status (cdc1, 2022), though such costs pale in comparison to the medical costs of obesity. given the enormous cost of obesity to both the economic vitality and public health of america, addressing its complex causes is of paramount importance. factors such as increasingly sedentary jobs and lifestyles, low levels of intentional physical activity, and the wide availability and popularity of calorie-dense, ultra-processed foods have been linked most consistently to obesity (hruby & hu, 2015; dixon, 2020). nonetheless, numerous other more biologically and socially complex factors — such as exposure to obesogenic endocrine-disrupting chemicals, sleep quantity and quality, psychological conditions, poverty, poor education, lack of access to healthy food, and genetics — also play a role (hruby & hu, 2015; dixon, 2020). this overwhelming number of factors that cumulatively contribute to the overall risk and prevalence of obesity, forming the so-called “obesity equation,” as i call it, necessitates that the issue be addressed from multiple angles. figure 1 depicts the multifactorial nature of obesity. aresty rutgers undergraduate research journal, volume i, issue iv figure 1: several diverse factors that contribute to the “obesity equation”. in the sections of this exploratory synthesis paper that follow, i address one such angle whereby the obesity problem can be intercepted: the consumption of sugar-sweetened beverages (ssbs). first, i delineate the evidence linking consumption of ssbs to obesity and related chronic conditions. i subsequently discuss in the context of ssbs the use of excise taxes, or taxes waged at the distributorend on specific goods and activities that tend to discourage purchases by resulting in higher prices at the consumer-end (chaloupka et al., 2019), as a tactic to help make a dent in the obesity equation in the united states. laced throughout is a discussion of the soda industry’s role in promoting increased ssb consumption and opposing public health initiatives to implement ssb excise taxes, despite strong evidence for the negative health effects of ssbs. finally, i propose that revenue from existing ssb excise taxes such as philadelphia’s be funneled into initiatives that further fight obesity to maximize their public health efficacy and i highlight evidence of fruit and vegetable subsidization as one potentially suitable initiative. aresty rutgers undergraduate research journal, volume i, issue iv 2 sugar-sweetened beverages, obesity, and related chronic disease: the evidence sugar-sweetened beverages include any beverage sweetened with a form of added sugar, such as sucrose or high-fructose corn syrup, including regular soda, sports drinks, and fruit drinks (cdc, 2022). while ssbs represent only one of many determinants of obesity, prospective cohort studies have consistently linked increased consumption of ssbs to weight gain and long-term risk of obesity and related conditions like type 2 diabetes (hu, 2013; malik et al., 2013). increases in ssb consumption in recent decades have also directly paralleled increases in obesity (hu & malik, 2010), further suggesting a likely link between the two trends. though some trials are limited by small sample sizes and relatively short intervention periods, a handful of clinical trials have solidified a likely causal relationship between ssbs and obesity. these studies demonstrate that in children, reducing ssb intake slows weight gain, and in adults, ssb consumption can promote weight gain (wolff and dansinger, 2008; de ruyter et al., 2012). moreover, studies showing no such apparent ssb-obesity connection should be analyzed critically. for instance, in a clinical trial of 47 overweight subjects randomly assigned to drink 1l per day of regular soda, semiskim milk with an equivalent number of calories to regular soda, aspartame-sweetened soda, or water for 6 months, maersk et al. (2012) found no significant differences in the changes in body weight and fat mass between groups over a 6-month period. on a smaller scale, however, the group consuming regular soda had significantly higher increases in liver fat mass, visceral fat mass, muscle fat mass, plasma triglycerides, and total cholesterol compared to the other 3 groups, indicating that ssbs impair cardiometabolic health with uncertain implications on long-term health, including body weight. the most common mechanism to explain the link between ssbs and obesity is that the consumption of liquid calories does not yield proportional caloric reductions in solid food intake (malik, 2010), thus leading to excess calorie intake that promotes weight gain. lending support to this theory, dimeglio and mattes (2000) performed a crossover study in 15 normal-weight adults and found that 4 weeks of dietary supplementation with 450 calories of regular soda led to an increase in total caloric consumption and bmi, yet no such changes were found when diets were supplemented with 450 calories of jellybeans. moreover, research has shown that the oversized portions of sugary drinks sold at many fast food chains and other outlets such as movie theaters not only cause people to consume more than they would from a smaller portion even if they do not finish the entire beverage, but these larger portion sizes also lead individuals to underestimate the amount of sugary drinks they actually consume (flood et al., 2006; nestle et al., 2015). perhaps people believe that they have not consumed in excess simply because there is still soda left in their cups; whatever the rationalization, such a lack of awareness of one’s consumption of a calorically -dense beverage surely undermines individuals’ efforts to control their weight and may contribute to the failure of the consumption of liquid calories to reduce consumption of solid food calories (malik, 2010). given that 12 oz. of the average ssb contains about 140-150 calories (malik, 2010), relatively small increases in consumption can yield significant increases in caloric intake. beyond promoting the excess consumption of calories that may lead to weight gain and obesity, ssbs may pose more nuanced chronic disease risks due to their ability to promote visceral adiposity (ma, 2016; odegaard, 2012; maersk et al., 2012) as well as the high glycemic load associated with their consumption (hu & malik, 2010). in a recent longitudinal cohort study with 1003 adults of mean age 45.3 years, jiantao ma and colleagues (2016) found that six years after initial biometric measurements were recorded, increasing ssb consumption frequency correlated with increasing visceral adipose tissue (vat) volume, or fat accumulatixon around the abdominal organs. notably, no such correlation was found between diet soda consumption and vat volume (ma, 2016). interestingly, ssb consumers were more likely to be engaged in slightly more aresty rutgers undergraduate research journal, volume i, issue iv physical activity than ssb-non-consumers, yet this nuance did not protect them from the bmi increases and vat volume increases observed at the end of the six-year interval of time (ma, 2016). the significantly higher increase in vat mass found in the regular soda group in the aforementioned clinical trial by maersk et al. (2012) further validates that ssbs induce a vat phenotype. vat cells are known to secrete a variety of the pro-inflammatory type of small cell-signaling proteins known as cytokines (alexopoulos et al.; 2014, alvehus et al., 2010; ohman et al., 2009). such cytokines may increase the risk of atherosclerosis and associated cardiac events, given that cardiac events are associated with increased circulating levels of inflammatory markers (ohman et al., 2009). proinflammatory cytokines are also capable of causing insulin resistance in adipose, muscle, and liver tissues due to their inhibitory effects on insulin signaling pathways (de luca & olefsky, 2008) which may explain reported associations between proinflammatory, cytokine-releasing vat and type 2 diabetes (hanley et al., 2009). additionally, the high glycemic loads associated with ssb consumption—which reflect a high content of simple carbohydrates capable of quickly raising blood sugar in standard portion sizes—are further linked to rises in inflammatory markers (cerf, 2013). these high glycemic loads are further associated with pancreatic beta cell dysfunction and insulin resistance, both of which are implicated in the pathogenesis of type 2 diabetes (cerf, 2013; leroith, 2002). unfortunately, methodological issues often inhibit the confidence with which one can draw conclusions from the results of studies addressing a possible link between ssbs and obesity (bucher della torre et al., 2016; pereira, 2006). such issues may accordingly explain why a complete consensus on the relationship between ssbs and obesity is lacking despite the large body of evidence that suggests a positive correlation. in a meta-analysis addressing the bias, generalizability, rigor of data analysis, and the conclusions drawn from 32 studies focused on ssbs and obesity in children and adolescents, bucher della torre and colleagues (2016) found that 23 of these studies had at least one methodological issue. these methodological issues most commonly reflected incomplete or confounded study definitions of ssbs, unreliable measurements of exposure, or participant loss before follow-up in cohort studies. while only 9 of the analyzed studies had no methodological issues, all studies reported either a positive correlation between ssbs and obesity or mixed results. on the other hand, the studies with methodological issues gave more inconsistent findings with some supporting the hypothesized link between ssbs and obesity, some reporting mixed results, and some finding no link at all. moreover, industry funding may also contribute to poor study quality. following an analysis of 133 articles published between 2001 and 2013 on the health effects of ssbs, litman and colleagues reported that 82% of independently funded articles found strong evidence that ssbs have adverse effects on health, whereas only 7% of industry-funded articles reached such a conclusion (2018). industryfunded articles were more likely to report a weak or absent correlation between ssbs and adverse health effects, potentially owing to the non-experimental basis of most of the industry-funded studies, which allows more room for bias (litman et al., 2018). litman et al. acknowledge that no conclusions about the quality of any individual article included in the study can be drawn, as they did not analyze the scientific quality of the articles against any objective standards (2018). nonetheless, the findings of litman et al. (2018) and bucher della torre et al. (2016) highlight the importance of scanning for elements of potential bias and carefully analyzing the methodology behind ssb studies before drawing conclusions, and further suggest that biases and flaws in the methodology of some past studies may understate the true contribution of ssbs to obesity and other adverse health effects. taken altogether, the consensus becomes quite clear: ssbs are not health-promoting but health-harming, contributing to obesity and related chronic diseases. given the economic and quality of life costs of obesity-related health conditions, public health initiatives to reduce consumption of ssbs are of vital importance. unfortunately, however, existing aresty rutgers undergraduate research journal, volume i, issue iv initiatives aimed at reducing consumption have generally been weak. the following section explores a possible reason for the lack of strong antissb action, briefly surveying the soda industry’s monetary involvement in perpetuating ssb consumption. 3 the soda industry’s likely role in promoting growth in ssb consumption & unclear public health stances while the need to reduce ssb consumption to better protect against obesity and related health conditions is clear, industry opposition has made this reduction a challenging task. indeed, ssb consumption has not increased independently of industry involvement. coca-cola, for instance, was originally sold in 6.5 oz. bottles, and in 1934, pepsi-cola made the first move in inflating portion sizes when it introduced a 12 oz. pepsi for the same price of a nickel as its competitor’s 6.5 oz. drink, inflating its profits in the process (dough, 2006). today, the minimum size soda can is 7.5 oz. and bottle sizes range up to 2 liters (nestle et al., 2015). moreover, the standard soda size sold at gas stations and in vending machines is approximately 20 oz. and while one must account for the lost soda volume due to the addition of ice in fountain drinks, even a kid-size soda from a fast-food restaurant is twelve ounces (nestle et al, 2015). citing nationwide surveys of the portions of food americans consume as justification for the change, the u.s. food and drug administration even raised its soda serving size from 8 to 12 ounces in 2014 (center for food safety and applied nutrition, 2022). while the agency states that “the serving size is not a recommendation of how much to eat or drink,” (center for food safety and applied nutrition, 2022) one must question why a nutritionfocused branch of government would not speak about soda in less ambiguous terms given that the large body of research suggests that it is not a health-promoting beverage. unsurprisingly, the food industry has been known to spend millions in political lobbying to influence government nutrition regulations and recommendations (gostin, 2016; nestle et al., 2015), suggesting a possible cause for the ambiguity in the fda’s stance on soda and other ssbs. perhaps further complicating the interests served by the fda is the so-called “revolving door” between the food industry and related government agencies by which executives from the food industry transition to high-power positions in government or vice versa, with little time in between. (nestle et al., 2015; piller, 2018; hyman, 2020). since 2009, the year a since-dropped federal soda tax was proposed, annual lobbying expenditures from the soda industry alone have sat at $60 million (du et al, 2018). clearly, the soda industry seems to have a vested interest in retaining its profits at the potential expense of the larger health interest of americans. 4 united states sugar-sweetened beverage excise taxes: progress despite industry opposition despite the dropped proposal to institute a nationwide ssb excise tax in 2009 and continued industry lobbying and opposition against ssb taxes (nestle et al., 2015; mcgranahan & whitmore schanzenbach, 2011; gostin, 2017), several citylevel ssb excise taxes have since passed. in 2014, berkeley, california became the first u.s. city to pass an excise tax of $0.01 per ounce on ssbs on account of public health interests (falbe et al. 2016; kane & malik, 2019). november 2016 proved to be a remarkable month, with five new ssb excise taxes ranging from $0.01 to $0.02 passed in cook county, illinois; boulder, colorado; and three cities in california’s bay area: san francisco, albany, and oakland (gostin, 2017; kane & malik, 2019). stockton, california and philadelphia, pennsylvania joined this growing list in june 2016, and seattle, washington in june 2017 (kane & malik, 2019). unsurprisingly, however, the american beverage association and the big soda companies it represents have fought many such taxes, spending $19 million to fight the tax proposal in san francisco alone (gostin, 2017). the soda industry is known to fight excise tax proposals through claims that such taxes hurt small businesses and establish a “nanny state,” aresty rutgers undergraduate research journal, volume i, issue iv in which the government infringes on the personal freedom individuals have to make their own choices as consumers (gostin, 2017; brownell & warner, 2009). ironically, companies from this same industry often aggressively market their products to children and teens through television ads and digital media, hooking them at a young age and denying the addictive nature of the sugar and caffeine they contain (brownell & warner, 2009; nestle et al., 2015; falbe et al., 2019; sylvetsky et al., 2020). surely this very practice interferes with the ability of individuals to make their own autonomous decisions about whether or not to purchase ssbs, making the soda industry’s stance against ssb excise taxes hypocritical. yet, while funding from philanthropists offered sufficient support to fight back and uphold the tax proposal in san francisco (gostin, 2017), proposals in other cities such as new york city, new york have nonetheless been dropped due to industry opposition of this sort (kansagra et al., 2015). to examine the efficacy of ssb excise taxes more closely, philadelphia’s tax will be examined in further detail. 5 philadelphia as a case study for evaluating the efficacy of ssb excise taxes officially known as the philadelphia beverage tax (city of philadelphia1, 2022), philadelphia’s tax is particularly interesting, as the primary motive behind its proposal was to provide funding to support mayor jim kenney’s initiative to make access to pre-kindergarten education universal (kane & malik, 2019). moreover, the tax was not only restricted to ssbs but also included any sweetened beverages (sbs), regardless of whether or not the sweetener contained calories (city of philadelphia1, 2022; kane & malik, 2019). examples of taxed beverages include both diet and regular soda, non-100%-fruit drinks, pre-sweetened tea and coffee drinks, and other pre-packed beverages sweetened with natural or artificial sweeteners (city of philadelphia1, 2022). the tax sits at $0.015 per ounce and is charged to all distributors of sbs, including restaurants, grocery stores, schools, hospitals, and even non-profit organizations (city of philadelphia1, 2022). distributors have passed 43-104% of the philadelphia sb excise tax onto consumers, depending on the outlet of purchase (madsen et al., 2019). the cost of sbs increased by $0.0065 per ounce in supermarkets, $0.0087 per ounce in mass merchandiser outlets, and $0.0156 per ounce in pharmacies (roberto et al., 2019). just one year after the excise tax was implemented on january 1, 2017 (madsen et al., 2019), an analysis of retail sales from january 1, 2014, to december 31, 2017, revealed that total volume retail sales of sbs in philadelphia had declined by 51% compared to sales prior to tax implementation (roberto et al., 2019). however, it should be noted that sales of sbs in pennsylvania zip codes surrounding philadelphia increased, offsetting 24.4% of this 51% decline (roberto et al., 2019). the net decline in sb sales brought about by implementation of the excess tax therefore appears to be approximately 38% (roberto et al., 2019),. however, data on sb sales were not evaluated in new jersey and could have further offset this decline, despite the disincentive of toll charges to travel across the border (roberto et al., 2019). nonetheless, the effectiveness of the sb tax in deterring sb purchases is promising. henceforth, the following question arises: do reduced purchases equate to significantly reduced consumption across the population in philadelphia? one longitudinal survey-based study examined soda drinking in one adult and one child from several hundred households in philadelphia and comparison areas in delaware; montgomery county, pennsylvania; or bucks county, pennsylvania before and after implementation of the philadelphia beverage tax. the study revealed that 10 to 11 months after implementation of the tax, philadelphian adults consumed about 1 less regular soda every 3 days (cawley et al., 2019). amongst african american adults in particular, the sb tax resulted in the consumption of 1 less regular soda every 2 days, suggesting that the effects of the tax may differ amongst sociodemographic groups (cawley et al., 2019). interestingly, however, cawley et al. did not aresty rutgers undergraduate research journal, volume i, issue iv report any statistically significant differences between socioeconomic groups (2019) despite the fact that ssb consumption tends to be higher amongst low-income individuals (cdc2, 2022). while no significant differences in soda consumption were reported amongst all philadelphian children after the tax, children who consumed the amount of sugar equivalent to that in a 20 oz. regular soda on a daily basis, or 67 grams per day, consumed 22% less added sugar after implementation of the tax (cawley et al., 2019), suggesting that the tax may have a more significant impact on individuals who were preexisting high-sugar consumers. it should be noted that cawley et al. used different samples of participants to make preand post-tax comparisons, potentially blurring the true, unadulterated effects of the tax of sb consumption. another cross-sectional survey-based study conducted at drexel university used random-dialing phone call data from 899 philadelphian participants and 878 nearby-comparison-area participants to determine that philadelphians were 40% less likely to consume regular soda, 64% less likely to consume energy drinks, and 58% more likely to consume bottled water following implementation of the sb tax. however, this reported change was based on data collected only 1-2 months after the tax had been put into practice and therefore may not accurately reflect its long-term effects (zhong et al., 2018). notably, a recent, more long-term study with a similar random-dialing phone call survey design conducted by some of the same drexel university researchers revealed that there were no significant overall differences in either ssb or diet sb consumption before and 12 months after tax implementation (zhong et al., 2020). further, there were no significant differences in such consumption between philadelphians (n=357) and non-philadelphians (n=158), though a slightly higher percentage of philadelphian participants decreased their ssb consumption compared to the non-philadelphians participants (zhong et al., 2020). yichen zhong and colleagues do however acknowledge that the sample size for this study was small (2020). moreover, the survey-based design of the study carries risk of bias, though accurately measuring changes in soda consumption (rather than purchases) via other methods would likely prove difficult. table 1 summarizes the basic characteristics of the aforementioned studies, examining the effect of philadelphia’s excise tax on sb consumption. studies examining the impact of the ssb tax instituted in berkeley, california on ssb consumption have reported a similar mix of findings, thus supplementing the relatively small pool of studies on philadelphia’s sb tax to strengthen the degree to which conclusions can be drawn about the general effects of ssb/sb taxes on purchasing behavior. one short-term study found that the tax in berkeley brought about a 21% reduction in ssb consumption in low-income neighborhoods (falbe et al., 2016), and another longer-term study reported no significant difference in self-reported ssb consumption one-year after tax implementation despite significant declines in ssb sales (silver et al, 2017). on the other hand, the results of one exceptionally longterm study that surveyed differences in consumption before and after implementation of the tax amongst individuals from demographically diverse berkeley neighborhoods (n = 1513) and comparison neighborhoods in san francisco and oakland (n = 3712) revealed that the frequency of ssb consumption in berkeley decreased by 0.55 times per day after tax-implementation based on a comparison of pre-tax consumption to a weighted average of consumption during the first 3 years post-tax-implementation. water consumption comparably increased by 1.02 times per day during the same time period (lee et al., 2019). while it should be noted that the cities from which the comparison neighborhoods were selected both passed ssb taxes in 2016, only a portion of the final 2017 surveys in oakland were collected after these taxes were implemented (lee et al., 2019). table 2 summarizes the basic characteristics of the aforementioned studies examining the effect of berkeley’s excise tax on ssb consumption. aresty rutgers undergraduate research journal, volume i, issue iv table 1: population characteristics and methodology for studies evaluating post-taxation changes in sb consumption in philadelphia, pa the similarly varied findings of the studies in berkeley and philadelphia and the limitations of the survey-based studies generally used to measure ssb consumption necessitate that further long-term studies with larger sample sizes be conducted before more conclusive claims are drawn on the true efficacy of ssb or sb excise taxes on ssb consumption. nonetheless, the existing evidence suggest that such taxes likely have some degree of efficacy. setting conclusions about the true efficacy of ssb and sb excise taxes aside, the philadelphia beverage tax has generated an impressive amount of revenue despite the significant reductions in soda volume purchased in the city after tax implementation (roberto et al., 2019). as of the end of the fourth fiscal quarter of 2021, the tax had generated $333.9 million since its enactment in 2017 (rhynhart, 2022). taking annual revenue to be consistent for the purposes of generating an average, the tax has generated roughly $66.8 million annually (rhynhart, 2022). $122 million, or 37% of the tax’s overall revenue, has been spent on pre-kindergarten education, and significantly smaller amounts have gone towards community schools, the city’s office of education and rebuild, the city’s new program focused on rebuilding community parks, libraries, and recreation centers (rhynhart, 2022; city of philadelphia2, 2022). while the city claims that rebuild will devote hundreds of millions of dollars to its initiatives thanks to the sb tax (city of philadelphia2, 2022), $179 million, or 54%, of philadelphia’s sb tax revenue still remains in the city’s general fund (rhynhart, 2022). despite city controller rebecca rhynhart’s repeated pushes for sb tax funds to be authors, year participant recruitment analytic sample characteristics data collection method length of time between tax implementation & evaluation cawley et al., 2019 separate groups of participants recruited outside of grocery stores at baseline and follow-up timepoints. baseline 2016 data for n=600 philadelphian households, n=705 comparison households from delaware and other pennsylvania areas; follow-up 2017 data for n=763 philadelphian households, n=738 comparison households from delaware and other pennsylvania areas online survey (by phone for those who could not complete online) 10-11 months zhong et al., 2018 part of the drexel university beverage choice research study. recruitment via random-digit phone dialing. n=899 philadelphian households, n=878 comparison households from cities in new jersey and delaware phone survey 1-2 months zhong et al., 2020 participants from the drexel university beverage choice research study recontacted 1 year after recruitment via random-digit phone dialing for initial surveying. participants paid $20 for follow-up survey only. n=479 philadelphian households, n=384 comparison households from cities in new jersey and delaware phone survey 1 year aresty rutgers undergraduate research journal, volume i, issue iv separated from the general fund to ensure transparent spending, no changes have been made (mccrystal, 2019; rhynhart, 2022). the spending timeline and specific allocation of remaining revenue remains unclear, with only vague commentary from mayor jim kenney’s office claiming that it would take time for the rate of revenue influx from the sb tax to become outpaced by the rate at which programs intended to be supported by such revenue can utilize the funds to develop (mccrystal, 2019). in the meantime, of course, tax revenue will continue to accumulate so long as the tax remains in effect. in the final section to follow, i propose the allocation of at least a portion of the revenue generated by the philadelphia sb excise to fund consumer-level fruit and vegetable subsidies as a possible means of not only ensuring that tax revenue is spent in a timely manner, but also enhancing the tax’s potential to combat obesity. table 2: population characteristics and methodology for studies evaluating post-taxation changes in sb consumption in berkeley, ca authors, year participant recruitment analytic sample characteristics data collection method length of time between tax implementation & evaluation falbe et al., 2016 separate groups of participants recruited via intercept surveying at high-traffic intersections at baseline and follow-up timepoints. baseline 2014 data for n=285 berkeley participants, n=606 comparison participants from oakland and san francisco, ca; follow-up 2015 data for n=501 philadelphian participants, n=1045 comparison participants from oakland and san francisco, ca. participants recruited from low-income neighborhoods with large minority populations. in-person interview 4 months silver et al., 2017 participants recruited via landline and cellular random-digit phone dialing for initial survey baseline 2014 measurements for n=649 berkeley participants; follow-up 2015 measurements for n=654 participants. low-income census blocks were disproportionately favored during random-digit dialing recruitment. phone survey 1 year lee et al., 2019 separate groups of participants recruited via intercept surveying at high-traffic intersections at baseline and follow-up timepoints. baseline 2014 data for n=1513 berkeley participants, n=3712 comparison participants from oakland and san francisco, ca; unspecified sample sizes for followup data collected in 2015, 2016, and 2017. participants recruited from low-income neighborhoods with large minority populations. in-person interview 3 years, plus annual follow-up aresty rutgers undergraduate research journal, volume i, issue iv 6 fruit & vegetable subsidies: an alternative obesity-minded investment for philadelphia’s sweetenedbeverage excise tax revenue though additional research is needed to know the full extent to which ssb taxes reduce soda consumption, existing evidence from u.s.-based studies in relevant cities like berkeley and philadelphia suggests that such taxes are effective to at least some extent (bleich & long, 2020). and while the philadelphia sb excise tax in particular was, again, not proposed with obesity as a primary motive (kane & malik, 2019), positive externalities are never unwelcomed. yet, one has to question why the $150.9 million in tax revenue in the general fund has not yet been devoted to fulfilling the mayor’s mission of funding universal pre-k or supporting the mission of rebuild (kane & malik, 2019; city of philadelphia2, 2022). the philadelphian public ought to demand that this revenue be reinvested within a set period of time, yet given the present lack of action and the burden of obesity in america, perhaps at least a portion of this revenue might be better spent directly funneled into programs that further fight obesity. i suggest a program that funds fruit and vegetable subsidies, whereby government funding would artificially lower the cost of such foods. it is generally well accepted that increasing consumption of whole fruits and vegetables is inversely associated with weight gain and obesity (guyenet, 2019; nour et al., 2018; dreher, 2018; he et al., 2004). possible causes for this correlational trend are the relatively low energy and high fiber content of fruits and vegetables, as well as their constituent plant phytochemicals (dreher, 2018). moreover, increasing fruit and vegetable consumption may reduce the risk of cardiovascular disease, hypertension, stroke, and certain cancers (boeing et al., 2012; hung et al., 2004). despite the fact that most data on the health effects of fruits and vegetables is correlational, there is no significant evidence to suggest that fruits and vegetables are harmful, and the science-backed 2020-2025 dietary guidelines for america recommend filling half your plate with whole fruits and vegetable (usda, 2022). yet according to nationwide 2013 behavioral risk factor surveillance system (brfss) data, only 12.3% of americans meet the fruit intake recommendations laid out by the 2015-2020 dietary guidelines for americans, and only 9.3% of americans meet the vegetable intake recommendations (lee-kwan et al., 2015). for many americans, however, complex socioeconomic factors, such as the relatively high cost of fruits and vegetables and limited access, make their low intake of such foods more than a matter of personal preference (lee-kwan et al., 2015). regarding cost barriers to accessing fruits and vegetables, relatively few interventional studies have explored the effectiveness of subsidies in promoting the purchase of healthy food (an, 2013; herman et al., 2008; pearson-stuttard et al., 2017). significantly, however, in one los angeles-based study amongst 602 postpartum women enrolled in the special supplemental nutrition program for women, infants, and children (wic), researcher dena herman and colleagues issued participants in the experimental group $10 worth of vouchers per week on a bimonthly basis to spend at either a farmers market site or a supermarket site over a 6-month intervention period (2008). survey data revealed that compared to control participants, who were given a monthly $13 dollar voucher to spend on disposable diapers, farmer’s market participants in the experimental group consumed 1.4 additional servings of fruits and vegetables per 1000 kcal in contrast to baseline, and supermarket participants consumed 0.8 additional servings per 1000 kcal (herman et al., 2008). remarkably, the reported increases in consumption were maintained 6 months after participants stopped receiving the vouchers (herman et al., 2008). such a finding suggests that even short periods of fruit and vegetable subsidization might have sufficiently positive effects on individuals to establish resilient changes in long-term purchasing behavior. if drawn out for a year, herman et al.’s fruit and vegetable subsidy intervention would cost $480 per wic participant (2018). according to u. s. department of agriculture (usda) data, there were aresty rutgers undergraduate research journal, volume i, issue iv 233,500 enrolled wic participants in pennsylvania as of 2016 (usda, 2019). no further data is available on the breakdown of this state-wide data, however (usda, 2019). employing herman et al.’s intervention amongst all pennsylvania wic participants for one year would theoretically cost approximately $112 million, almost double the $65.1 million average annual revenue generated by the philadelphia sb excise tax (rhynhart, 2019). further information about the number of wic participants in philadelphia is needed before conclusions can be drawn about the feasibility of establishing such an intervention. however, even if the vast majority of the 233,500 enrolled wic participants in pennsylvania were enrolled in philadelphia, funding might still be feasible with sb tax revenue if weekly subsidies were cut by a small percentage or if funding was received more intermittently throughout the year. importantly, the findings of a recent longitudinal cohort study of snap beneficiaries throughout north carolina further support the efficacy of fruit and vegetable subsidies in encouraging more healthful dietary consumption (berkowitz et al., 2021). compared to those registered for standard snap benefits alone (n=33246), those receiving an additional $40 per month through enrollment in supersnap, established to fund only the purchase of fresh, frozen, or canned fruits and vegetables free of added sugar and salt, purchased an average of $31.84 more of fruits, vegetables, nuts, and legumes per month. while it follows that enrollment in supersnap did not increase out-of-pocket spending on fruits, vegetables, and other whole foods, it nonetheless produced a significant shift in purchasing behavior. this shift also included a significant decrease in spending on both processed foods and sugar-sweetened beverages, though because supersnap beneficiaries were enrolled in the program by clinical staff due to pre-existing health conditions (e.g., diabetes or obesity), there is the possibility that such beneficiaries may have been relatively biased towards healthy eating. nonetheless, this study further supports the theory that lowering the cost barrier to accessing fruits and vegetables can potentially induce a measurable change in dietary consumption, which one can presume aligns reasonably well with food purchasing behavior. ultimately, wic participants pose a more economically feasible target population for a subsidy program in philadelphia than snap participants, as significantly more individuals are enrolled in the latter: as of 2021, 464, 008 individuals were receiving snap benefits in philadelphia alone (u.s. census bureau, 2022). additionally, given that research suggests that parental modeling and exposure to foods during childhood plays a significant role in shaping long-term feeding behavior (birch et al., 2007; vollmer & baietto, 2017; ventura & worobey, 2013; wadhera et al., 2015), subsidizing fruits and vegetables for wic participants might generate self-perpetuating positive effects on diets across generations. beyond subsidizing fruit and vegetables for wic participants, tax revenue from the philadelphia sb excise tax might alternatively be used to lower the cost of fruit and vegetable subsidies on a somewhat broader scale. amongst other findings, a meta-analysis of 23 interventional studies and 7 prospective cohort studies selected from 3,163 reviewed abstracts related to the impact on price changes on diet identified that just a 10% decrease in fruit and vegetable prices resulted in a 14% increase in fruit and vegetable purchases (pearsonstuttard et al., 2017). such findings suggest that using a portion of the sb tax revenue to subsidize select fruits and vegetables in philadelphia might produce a measurable rise in purchases of these healthpromoting foods. perhaps such subsidies might be better selectively applied to grocery stores or farmer’s markets serving lower-income neighborhoods, where the existing costs of fruits and vegetables pose more of a significant access barrier. while additional expertise in economics, data science, and public health beyond the scope of this synthesis paper is needed to propose a specific and practicable plan for using tax revenue from the philadelphia sb excise tax to subsidize fruits and vegetables, the aforementioned studies suggest that subsidizing fruits and vegetables could proaresty rutgers undergraduate research journal, volume i, issue iv duce measurable results. indeed, the available evidence indicates that targeting smaller populations or select fruits and vegetables could make the costs of the subsidies realistic, given existing revenue from the taxes. moreover, such subsidies might help support philadelphia grocery stores, which appear to have suffered losses after implementation of the tax in 2017. a few months after the tax had been put into practice, the owner of 6 shoprite stores in philadelphia, jeff brown, told national public radio that his sales were down 15%, leaving him with no choice but to cut hours for his union employees (aubrey, 2017). sales presumably underwent such a remarkable drop due to some individuals choosing to shop outside of philadelphia to avoid the tax (aubrey, 2017), and whether or not this drop lessened over time is unclear. nonetheless, the number is striking. subsidizing fruits and vegetables while simultaneously taxing sbs might help to bring back some of this lost business in philadelphia. 7 conclusion obesity has been on the rise since the late 1970s, and its massive health and economic consequences cannot be ignored. while ssb excise taxes are by no means capable of resolving this public health threat altogether, existing evidence reasonably supports the use of such taxes, as ssbs have been shown to promote weight gain and chronic disease, and this scientific consensus might be even more clear if methodological issues and probable industry-biases in some studies and reviews were eliminated. ssb excise taxes, though still few and far between, seem to hold some efficacy in reducing the purchase and consumption of ssbs. yet, while these taxes generate valuable revenue, their obesity-fighting potential might be optimized if such revenue were devoted to addressing other factors that contribute to the obesity equation. fruit and vegetable subsidies pose one such promising possibility and might help to combat lost business generated by sb/ssb excise taxes in cities like philadelphia, where the majority of sb tax revenue generated since its 2017 inception remains unspent. despite inevitable industry opposition, bringing back a federal proposal for a ssb excise ought to be the ultimate goal, as this more unified approach would maximize public health benefits and likely rid retail stores in cities with existing sb/ssb taxes of the issue of lost revenue that may emerge when some individuals resultingly choose to shop for groceries in areas with no such taxes. until then, however, cities like philadelphia ought to continue to act as role models and work to improve their sb/sbb tax programs to optimize their obesityfighting potential, setting examples that will hopefully ignite a larger movement∎ 8 references [1] alexopoulos, n., katritsis, d., & raggi, p. (2014). visceral adipose tissue as a source of inflammation and promoter of atherosclerosis. atherosclerosis, 233(1), 104–112. https://doi.org/10.1016/j.atherosclerosis.2013.12.023 [2] alvehus, m., burén, j., sjöström, m., goedecke, j., & olsson, t. (2010). the human visceral fat depot has a unique inflammatory profile. obesity (silver spring, md.), 18(5), 879–883. https://doi.org/10.1038/oby.2010.22 [3] an, r. (2013). effectiveness of subsidies in promoting healthy food purchases and consumption: a review of field experiments. public health nutrition, 16(7), 1215–1228. https://doi.org/10.1017/s1368980012004715 [4] aubrey, a. (2017). judges take up big soda's suit to abolish philadelphia's sugar tax. national public radio. https://www.npr.org/sections/thesalt/2017/04/05/522626223/judges-take-up-big-sodas-suit-toabolish-philadelphias-sugar-tax [5] berkowitz, s. a., curran, n., hoeffler, s., henderson, r., price, a., & ng, s. w. (2021). association of a fruit and vegetable subsidy program with food purchases by individuals with low income in the us. jama network open, 4(8), e2120377. https://doi.org/10.1001/jamanetworkopen.2021.20377 [6] birch, l., savage, j. s., & ventura, a. (2007). influences on the development of children's eating behaviours: from infancy to adolescence. canadian journal of dietetic practice and research: a publication of dietitians of canada = revue canadienne de la pratique et de la recherche en dietetique : une publication des dietetistes du canada, 68(1), s1–s56. [7] boeing, h., bechthold, a., bub, a., ellinger, s., haller, d., kroke, a., leschik-bonnet, e., müller, m. j., oberritter, h., schulze, m., stehle, p., & watzl, b. (2012). critical review: vegetables and fruit in the prevention of chronic diseases. european journal of nutrition, 51(6), 637–663. https://doi.org/10.1007/s00394-012-0380-y [8] brownell, k. d., & warner, k. e. (2009). the perils of ignoring history: big tobacco played dirty and https://doi.org/10.1016/j.atherosclerosis.2013.12.023 https://doi.org/10.1016/j.atherosclerosis.2013.12.023 https://doi.org/10.1038/oby.2010.22 https://doi.org/10.1017/s1368980012004715 https://www.npr.org/sections/thesalt/2017/04/05/522626223/judges-take-up-big-sodas-suit-to-abolish-philadelphias-sugar-tax https://www.npr.org/sections/thesalt/2017/04/05/522626223/judges-take-up-big-sodas-suit-to-abolish-philadelphias-sugar-tax https://www.npr.org/sections/thesalt/2017/04/05/522626223/judges-take-up-big-sodas-suit-to-abolish-philadelphias-sugar-tax https://doi.org/10.1001/jamanetworkopen.2021.20377 https://doi.org/10.1007/s00394-012-0380-y aresty rutgers undergraduate research journal, volume i, issue iv millions died. how similar is big food? the milbank quarterly, 87(1), 259–294. https://doi.org/10.1111/j.1468-0009.2009.00555.x [9] bucher della torre, s., keller, a., laure depeyre, j., & kruseman, m. (2016). sugar-sweetened beverages and obesity risk in children and adolescents: a systematic analysis on how methodological quality may influence conclusions. journal of the academy of nutrition and dietetics, 116(4), 638–659. https://doi.org/10.1016/j.jand.2015.05.020 [10] cawley, j., frisvold, d., hill, a., & jones, d. (2019). the impact of the philadelphia beverage tax on purchases and consumption by adults and children. journal of health economics, 67, 102225. https://doi.org/10.1016/j.jhealeco.2019.102225 [11] centers for disease control and prevention (2022). childhood overweight and obesity. https://www.cdc.gov/obesity/childhood/index.html [12] centers for disease control and prevention1 (2022). adult obesity causes & consequences. https://www.cdc.gov/obesity/adult/causes.html [13] centers for disease control and prevention2 (2022). get the facts: sugar-sweetened beverages and consumption https://www.cdc.gov/nutrition/data-statistics/sugar-sweetened-beverages-intake.html [14] centers for disease control and prevention3 (2022). prevalence of obesity and severe obesity among adults: united states, 2017–2018. https://www.cdc.gov/nchs/products/databriefs/db360.htm [15] center for food safety and applied nutrition (2022). serving size on the new nutrition facts label. u.s. food and drug administration (fda). www.fda.gov/food/new-nutrition-facts-label/serving-size-newnutrition-facts-label. [16] cerf m. e. (2013). beta cell dysfunction and insulin resistance. frontiers in endocrinology, 4, 37. https://doi.org/10.3389/fendo.2013.00037 [17] chaloupka, f. j., powell, l. m., & warner, k. e. (2019). the use of excise taxes to reduce tobacco, alcohol, and sugary beverage consumption. annual review of public health, 40, 187–201. https://doi.org/10.1146/annurev-publhealth-040218-043816 [18] city of philadelphia1 (2022). payments, assistance & taxes: philadelphia beverage tax (pbt). https://www.phila.gov/services/payments-assistance-taxes/business-taxes/philadelphia-beverage-tax/ [19] city of philadelphia2 (2022). rebuild: building stronger communities, one park, recreation center, and library at a time. https://www.phila.gov/programs/rebuild/ [20] de luca, c., & olefsky, j. m. (2008). inflammation and insulin resistance. febs letters, 582(1), 97– 105. https://doi.org/10.1016/j.febslet.2007.11.057 [21] de ruyter, j. c., olthof, m. r., seidell, j. c., & katan, m. b. (2012). a trial of sugar-free or sugarsweetened beverages and body weight in children. the new england journal of medicine, 367(15), 1397–1406. https://doi.org/10.1056/nejmoa1203034 [22] dimeglio, d. p., & mattes, r. d. (2000). liquid versus solid carbohydrate: effects on food intake and body weight. international journal of obesity and related metabolic disorders: journal of the international association for the study of obesity, 24(6), 794–800. https://doi.org/10.1038/sj.ijo.0801229 [23] dixon, j. l. (2020). obesity explosion: the collision of nutrition, culture, politics, & food. kendall hunt. [24] dough, w. (2006). pepsi-cola. ncpedia. https://www.ncpedia.org/pepsi-cola [25] dreher, m. l. (2018). whole fruits and fruit fiber emerging health effects. nutrients, 10(12), 1833. https://doi.org/10.3390/nu10121833 [26] du, m., tugendhaft, a., erzse, a., & hofman, k. j. (2018). sugar-sweetened beverage taxes: industry response and tactics. the yale journal of biology and medicine, 91(2), 185–190. [27] falbe, j., thompson, h. r., becker, c. m., rojas, n., mcculloch, c. e., & madsen, k. a. (2016). impact of the berkeley excise tax on sugar-sweetened beverage consumption. american journal of public health, 106(10), 1865–1871. https://doi.org/10.2105/ajph.2016.303362 [28] falbe, j., thompson, h. r., patel, a., & madsen, k. a. (2019). potentially addictive properties of sugar-sweetened beverages among adolescents. appetite, 133, 130–137. https://doi.org/10.1016/j.appet.2018.10.032 [29] flood, j. e., roe, l. s., & rolls, b. j. (2006). the effect of increased beverage portion size on energy intake at a meal. journal of the american dietetic association, 106(12), 1984–1991. https://doi.org/10.1016/j.jada.2006.09.005 [30] gostin l. o. (2016). "big food" is making america sick. the milbank quarterly, 94(3), 480–484. https://doi.org/10.1111/1468-0009.12209 [31] gostin l. o. (2017). 2016: the year of the soda tax. the milbank quarterly, 95(1), 19–23. https://doi.org/10.1111/1468-0009.12240 [32] guyenet s. j. (2019). impact of whole, fresh fruit consumption on energy intake and adiposity: a systematic review. frontiers in nutrition, 6, 66. https://doi.org/10.3389/fnut.2019.00066 [33] hanley, a. j., wagenknecht, l. e., norris, j. m., bryer-ash, m., chen, y. i., anderson, a. m., bergman, r., & haffner, s. m. (2009). insulin resistance, beta cell dysfunction and visceral adiposity as predictors of incident diabetes: the insulin resistance atherosclerosis study (iras) family study. diabetologia, 52(10), 2079–2086. https://doi.org/10.1007/s00125-009-1464-y [34] he, k., hu, f. b., colditz, g. a., manson, j. e., willett, w. c., & liu, s. (2004). changes in intake of fruits and vegetables in relation to risk of obesity and weight gain among middle-aged women. international journal of obesity and related metabolic disorders : journal of the international association for the study of obesity, 28(12), 1569–1574. https://doi.org/10.1038/sj.ijo.0802795 [35] herman, d. r., harrison, g. g., afifi, a. a., & jenks, e. (2008). effect of a targeted subsidy on intake of fruits and vegetables among low-income women in the special supplemental nutrition program for https://doi.org/10.1111/j.1468-0009.2009.00555.x https://doi.org/10.1016/j.jand.2015.05.020 https://doi.org/10.1016/j.jhealeco.2019.102225 https://www.cdc.gov/obesity/childhood/index.html https://www.cdc.gov/nutrition/data-statistics/sugar-sweetened-beverages-intake.html https://www.cdc.gov/nutrition/data-statistics/sugar-sweetened-beverages-intake.html https://www.cdc.gov/nchs/products/databriefs/db360.htm http://www.fda.gov/food/new-nutrition-facts-label/serving-size-new-nutrition-facts-label http://www.fda.gov/food/new-nutrition-facts-label/serving-size-new-nutrition-facts-label https://doi.org/10.3389/fendo.2013.00037 https://doi.org/10.1146/annurev-publhealth-040218-043816 https://www.phila.gov/services/payments-assistance-taxes/business-taxes/philadelphia-beverage-tax/ https://www.phila.gov/services/payments-assistance-taxes/business-taxes/philadelphia-beverage-tax/ https://www.phila.gov/programs/rebuild/ https://www.phila.gov/programs/rebuild/ https://doi.org/10.1016/j.febslet.2007.11.057 https://doi.org/10.1056/nejmoa1203034 https://doi.org/10.1038/sj.ijo.0801229 https://www.ncpedia.org/pepsi-cola https://doi.org/10.3390/nu10121833 https://doi.org/10.2105/ajph.2016.303362 https://doi.org/10.1016/j.appet.2018.10.032 https://doi.org/10.1016/j.appet.2018.10.032 https://doi.org/10.1016/j.jada.2006.09.005 https://doi.org/10.1111/1468-0009.12209 https://doi.org/10.1111/1468-0009.12240 https://doi.org/10.3389/fnut.2019.00066 https://doi.org/10.1007/s00125-009-1464-y https://doi.org/10.1038/sj.ijo.0802795 aresty rutgers undergraduate research journal, volume i, issue iv women, infants, and children. american journal of public health, 98(1), 98–105. https://doi.org/10.2105/ajph.2005.079418 [36] hruby, a., & hu, f. b. (2015). the epidemiology of obesity: a big picture. pharmaco economics, 33(7), 673–689. https://doi.org/10.1007/s40273-0140243-x [37] hu f. b. (2013). resolved: there is sufficient scientific evidence that decreasing sugar-sweetened beverage consumption will reduce the prevalence of obesity and obesity-related diseases. obesity reviews: an official journal of the international association for the study of obesity, 14(8), 606–619. https://doi.org/10.1111/obr.12040 [38] hu, f. b., & malik, v. s. (2010). sugar-sweetened beverages and risk of obesity and type 2 diabetes: epidemiologic evidence. physiology & behavior, 100(1), 4754. https://doi.org/https://doi.org/10.1016/j.physbeh.2010.01.0 36 [39] hung, h. c., joshipura, k. j., jiang, r., hu, f. b., hunter, d., smith-warner, s. a., colditz, g. a., rosner, b., spiegelman, d., & willett, w. c. (2004). fruit and vegetable intake and risk of major chronic disease. journal of the national cancer institute, 96(21), 1577–1584. https://doi.org/10.1093/jnci/djh296 [40] hyman, m. (2020). food fix: how to save our health, our economy, our communities, and our planet-one bite at a time. little, brown spark. [41] kane, r. m., & malik, v. s. (2019). understanding beverage taxation: perspective on the philadelphia beverage tax's novel approach. journal of public health research, 8(1), 1466. https://doi.org/10.4081/jphr.2019.1466 [42] kansagra, s. m., kennelly, m. o., nonas, c. a., curtis, c. j., van wye, g., goodman, a., & farley, t. a. (2015). reducing sugary drink consumption: new york city's approach. american journal of public health, 105(4), e61–e64. https://doi.org/10.2105/ajph.2014.302497 [43] lee-kwan, s. h., moore, l. v., blanck, h. m., harris, d. m., & galuska, d. (2017). disparities in statespecific adult fruit and vegetable consumption united states, 2015. mmwr. morbidity and mortality weekly report, 66(45), 1241–1247. https://doi.org/10.15585/mmwr.mm6645a1 [44] leroith d. (2002). beta-cell dysfunction and insulin resistance in type 2 diabetes: role of metabolic and genetic abnormalities. the american journal of medicine, 113 suppl 6a, 3s–11s. https://doi.org/10.1016/s0002-9343(02)01276-7 [45] litman, e. a., gortmaker, s. l., ebbeling, c. b., & ludwig, d. s. (2018). source of bias in sugarsweetened beverage research: a systematic review. public health nutrition, 21(12), 2345–2350. https://doi.org/10.1017/s1368980018000575 [46] ma, j., mckeown, n. m., hwang, s. j., hoffmann, u., jacques, p. f., & fox, c. s. (2016). sugar-sweetened beverage consumption is associated with change of visceral adipose tissue over 6 years of follow-up. circulation, 133(4), 370–377. https://doi.org/10.1161/circulationaha.115.018704 [47] madsen, k. a., krieger, j., & morales, x. (2019). sugar-sweetened beverage taxes: emerging evidence on a new public health policy. jama, 321(18), 1777–1779. https://doi.org/10.1001/jama.2019.5344 [48] malik, v. s., popkin, b. m., bray, g. a., després, j. p., & hu, f. b. (2010). sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. circulation, 121(11), 1356–1364. https://doi.org/10.1161/circulationaha.109.876185 [49] malik, v. s., pan, a., willett, w. c., & hu, f. b. (2013). sugar-sweetened beverages and weight gain in children and adults: a systematic review and meta-analysis. the american journal of clinical nutrition, 98(4), 1084–1102. https://doi.org/10.3945/ajcn.113.058362 [50] mccrystal, l. (2019, november 13). philly’s controller says only 30% of soda tax revenue has been spent. she wants more transparency. the philadelphia inquirer. https://www.inquirer.com/news/soda-tax-beverage-rebecca-rhynhart-jim-kenney-revenue-spending-prek-rebuild-20191113.html [51] mcgranahan, l., & whitmore schanzenbach, d. (2011). who would be affected by soda taxes? chicago fed letter, no. 284, 1. [52] must, a., spadano, j., coakley, e. h., field, a. e., colditz, g., & dietz, w. h. (1999). the disease burden associated with overweight and obesity. jama, 282(16), 1523–1529. https://doi.org/10.1001/jama.282.16.1523 [53] nestle, m., bittman, m., & baer, m. (2015). soda politics: taking on big soda (and winning). in soda politics. oxford university press, incorporated. [54] nour, m., lutze, s. a., grech, a., & allman-farinelli, m. (2018). the relationship between vegetable intake and weight outcomes: a systematic review of cohort studies. nutrients, 10(11), 1626. https://doi.org/10.3390/nu10111626 [55] odegaard, a. o., choh, a. c., czerwinski, s. a., towne, b., & demerath, e. w. (2012). sugar-sweetened and diet beverages in relation to visceral adipose tissue. obesity (silver spring, md.), 20(3), 689–691. https://doi.org/10.1038/oby.2011.277 [56] office of the surgeon general us (2010). the surgeon general’s vision for a healthy and fit nation. https://www.ncbi.nlm.nih.gov/books/nbk44656/ [57] ohman, m. k., wright, a. p., wickenheiser, k. j., luo, w., & eitzman, d. t. (2009). visceral adipose tissue and atherosclerosis. current vascular pharmacology, 7(2), 169–179. https://doi.org/10.2174/157016109787455680 [58] pearson-stuttard, j., bandosz, p., rehm, c. d., penalvo, j., whitsel, l., gaziano, t., conrad, z., wilde, p., micha, r., lloyd-williams, f., capewell, s., mozaffarian, d., & o'flaherty, m. (2017). reducing us cardiovascular disease burden and disparities through national and targeted dietary policies: https://doi.org/10.2105/ajph.2005.079418 https://doi.org/10.1007/s40273-014-0243-x https://doi.org/10.1007/s40273-014-0243-x https://doi.org/10.1111/obr.12040 https://doi.org/https:/doi.org/10.1016/j.physbeh.2010.01.036 https://doi.org/https:/doi.org/10.1016/j.physbeh.2010.01.036 https://doi.org/10.1093/jnci/djh296 https://doi.org/10.4081/jphr.2019.1466 https://doi.org/10.2105/ajph.2014.302497 https://doi.org/10.15585/mmwr.mm6645a1 https://doi.org/10.1016/s0002-9343(02)01276-7 https://doi.org/10.1017/s1368980018000575 https://doi.org/10.1161/circulationaha.115.018704 https://doi.org/10.1001/jama.2019.5344 https://doi.org/10.1161/circulationaha.109.876185 https://doi.org/10.3945/ajcn.113.058362 https://www.inquirer.com/news/soda-tax-beverage-rebecca-rhynhart-jim-kenney-revenue-spending-prek-rebuild-20191113.html https://www.inquirer.com/news/soda-tax-beverage-rebecca-rhynhart-jim-kenney-revenue-spending-prek-rebuild-20191113.html https://www.inquirer.com/news/soda-tax-beverage-rebecca-rhynhart-jim-kenney-revenue-spending-prek-rebuild-20191113.html https://doi.org/10.1001/jama.282.16.1523 https://doi.org/10.3390/nu10111626 https://doi.org/10.1038/oby.2011.277 https://www.ncbi.nlm.nih.gov/books/nbk44656/ https://doi.org/10.2174/157016109787455680 aresty rutgers undergraduate research journal, volume i, issue iv a modelling study. plos medicine, 14(6), e1002311. https://doi.org/10.1371/journal.pmed.1002311 [59] pereira, m. (2006). the possible role of sugarsweetened beverages in obesity etiology: a review of the evidence. international journal of obesity, 30(s3), s28–s36. https://doi.org/10.1038/sj.ijo.0803489 [60] pereira, m. a. (2013). diet beverages and the risk of obesity, diabetes, and cardiovascular disease: a review of the evidence. nutrition reviews, 71(7), 433–440. https://doi.org/10.1111/nure.12038 [61] piller, c. (2018). is fda’s revolving door open too wide? science, 361(6397), 21–21. https://doi.org/10.1126/science.361.6397.21 [62] rhynhart, r. (2022). data release: beverage tax revenue and expenditures. city of philadelphia office of the controller: rebecca rhynhart, city controller. https://controller.phila.gov/philadelphia-audits/data-release-beverage-tax/ [63] roberto, c. a., lawman, h. g., levasseur, m. t., mitra, n., peterhans, a., herring, b., & bleich, s. n. (2019). association of a beverage tax on sugarsweetened and artificially sweetened beverages with changes in beverage prices and sales at chain retailers in a large urban setting. jama, 321(18), 1799–1810. https://doi.org/10.1001/jama.2019.4249 [64] silver, l. d., ng, s. w., ryan-ibarra, s., taillie, l. s., induni, m., miles, d. r., poti, j. m., & popkin, b. m. (2017). changes in prices, sales, consumer spending, and beverage consumption one year after a tax on sugar-sweetened beverages in berkeley, california, us: a before-and-after study. plos medicine, 14(4), e1002283. https://doi.org/10.1371/journal.pmed.1002283 [65] sylvetsky, a. c., parnarouskis, l., merkel, p. e., & gearhardt, a. n. (2020). children’s sugar-sweetened beverage consumption: striking parallels with substance use disorder symptoms. frontiers in pediatrics, 8, 594513. https://doi.org/10.3389/fped.2020.594513 [66] usda (2022). dietary guidelines. u.s. department of agriculture: choosemyplate. https://www.choosemyplate.gov/eathealthy/dietary-guidelines [67] usda (2019). wic 2016 eligibility and coverage rates. u.s. department of agriculture: food and nutrition service. https://www.fns.usda.gov/wic/wic-2016eligibility-and-coverage-rates#chart1 [68] u.s. census bureau (2022). snap benefits recipients in philadelphia county/city, pa. federal reserve bank of st. louis. https://fred.stlouisfed.org/series/cbr42101paa647ncen [69] ventura, a. & worobey, j. (2013). early influences on the development of food preferences. current biology, 23(9), r401–r408. https://doi.org/10.1016/j.cub.2013.02.037 [70] vollmer, b. & baietto, j. (2017). practices and preferences: exploring the relationships between food-related parenting practices and child food preferences for high fat and/or sugar foods, fruits, and vegetables. appetite, 113, 134–140. https://doi.org/10.1016/j.appet.2017.02.019 [71] wadhera, d., capaldi phillips, e. d., wilkie, l. m., & boggess, m. m. (2015). perceived recollection of frequent exposure to foods in childhood is associated with adulthood liking. appetite, 89, 22–32. https://doi.org/10.1016/j.appet.2015.01.011 [72] ward, z. j., bleich, s. n., long, m. w., & gortmaker, s. l. (2021). association of body mass index with health care expenditures in the united states by age and sex. plos one, 16(3), e0247307. https://doi.org/10.1371/journal.pone.0247307 [73] wolff, e. & dansinger, m. l. (2008). soft drinks and weight gain: how strong is the link? the medscape journal of medicine, 10(8), 189 [74] zhong, y., auchincloss, a. h., lee, b. k., & kanter, g. p. (2018). the short-term impacts of the philadelphia beverage tax on beverage consumption. american journal of preventive medicine, 55(1), 26–34. https://doi.org/10.1016/j.amepre.2018.02.017 [75] zhong, y., auchincloss, a. h., lee, b. k., mckenna, r. m., & langellier, b. a. (2020). sugar-sweetened and diet beverage consumption in philadelphia one year after the beverage tax. international journal of environmental research and public health, 17(4), 1336. https://doi.org/10.3390/ijerph17041336 https://doi.org/10.1371/journal.pmed.1002311 https://doi.org/10.1038/sj.ijo.0803489 https://doi.org/10.1111/nure.12038 https://doi.org/10.1126/science.361.6397.21 https://controller.phila.gov/philadelphia-audits/data-release-beverage-tax/ https://controller.phila.gov/philadelphia-audits/data-release-beverage-tax/ https://doi.org/10.1001/jama.2019.4249 https://doi.org/10.1371/journal.pmed.1002283 https://www.choosemyplate.gov/eathealthy/dietary-guidelines https://www.choosemyplate.gov/eathealthy/dietary-guidelines https://www.fns.usda.gov/wic/wic-2016-eligibility-and-coverage-rates#chart1 https://www.fns.usda.gov/wic/wic-2016-eligibility-and-coverage-rates#chart1 https://fred.stlouisfed.org/series/cbr42101paa647ncen https://fred.stlouisfed.org/series/cbr42101paa647ncen https://doi.org/10.1016/j.cub.2013.02.037 https://doi.org/10.1016/j.appet.2017.02.019 https://doi.org/10.1016/j.appet.2015.01.011 https://doi.org/10.1371/journal.pone.0247307 https://doi.org/10.1016/j.amepre.2018.02.017 https://doi.org/10.1016/j.amepre.2018.02.017 https://doi.org/10.3390/ijerph17041336 aresty rutgers undergraduate research journal, volume i, issue iv this work is licensed under a creative commons attribution-noncommercial-sharealike 4.0 international license. sarah pfreundschuh is a class of 2022 graduate from rutgers honors college and the school of environmental and biological sciences. she majored in nutritional science and minored in biochemistry, completing wet lab research on the metabolic functions of stearoyl coa desaturase enzymes. separately, she pursued literature-based research on the adverse health effects of sugarsweetened beverages and the leveraging of tax-based solutions to best address such effects as a term project for a class taught by nutritional sciences professor dr. joseph l. dixon, titled “obesity: biology, behavior, and management.” sarah currently interns at a functional medicine practice and hopes to attend medical school to pursue a career as a functional medicine physician, focusing on disease prevention and root-cause-oriented, systems-based approaches to the diagnosis and treatment of existing chronic health conditions. she holds related side interests in sustainable agriculture and initiatives that increase access to nutrient-dense foods and enjoys cooking, yoga, and spending time outdoors in her free time. aresty rutgers undergraduate research journal, volume i, issue ii introduction to volume i, issue ii of the aresty rutgers undergraduate research journal ✵ rutgers university is a research powerhouse, fueled by the thousands of undergraduates who invest themselves in the research community each year. the aresty rurj has been established to showcase the cutting-edge research being conducted at this institution by providing a platform for undergraduate researchers to publish their work. not only does this journal showcase high-quality undergraduate research, but its undergraduate-led founding and development serves as a testament to the heights that the students at rutgers are capable of reaching. the aresty rurj has proven to be a valuable learning experience for all those involved. the undergraduate reviewers are provided unique exposure to the reviewer side of the peer review process, which undergraduates seldom have the opportunity to experience. the authors undergo the peer review process, which is modeled after other peer reviewed journals and includes several stages of rigorous reviews. the pilot year of this journal has fortunately been a success, and we have high hopes for the future. the editorial team of volume i, issue ii has set a high precedent for subsequent iterations. we would like to express our thanks and appreciation to the junior founders, authors, undergraduate reviewers, graduate reviewers, faculty reviewers, copy editors, chief copy editor, creative director, aresty research center administrators, rutgers libraries team, and faculty advisors—all of whom have contributed significantly to this issue. we hope that the aresty rurj continues to weave itself into the fabric of rutgers research and acts as a reminder of undergraduate achievement and potential. aresty rurj founders, prachi srivastava frederric kelada