573 American Academic Scientific Research Journal for Engineering, Technology, and Sciences ISSN (Print) 2313-4410, ISSN (Online) 2313-4402 http://asrjetsjournal.org/ Prediction of Genomic Signature of Ngs Sequences and Comparative Drug-Likeness Angela U. Makolo a , Festus Segun Ajiboye b* a,b University of Ibadan Bioinformatics Group, Department of Computer Science, University of Ibadan, Ibadan 200132, Nigeria Department of Computer Science, Federal School of Statistics Ibadan a Email: aumakolo@gmail.com, b Email: ajiboyefestussegun@gmail.com Abstract Developing a drug or particular immunotherapy medication for a worldwide epidemic illness caused by viruses (current pandemic) necessitates comprehensive evaluation and annotation of the metagenomic datasets to filter nucleotide sequences quickly and efficiently. Because of the homologs' origin of aligning sequences, space complexity, and time complexity of the analyzing system, traditional sequence alignment procedures are unsuccessful. This necessitates employing an alignment-free sequencing approach in this research that solves the foregoing issue. We suggest a distance function that compresses performance metrics for automatically identifying Short nucleotide sequences used by SARS coronavirus variants to identify critical features in genetic markers and genomic structure. This method provides easy recognition of data compressed by using a set of mathematical and computational tools in the study. We also show that by using our suggested technique to examine extremely short regions of nucleotide sequences, we can differentiate SAR-CoV-2 from SAR-CoV-1 viruses. Later, the Lipinski descriptor (rule of 5) was used to predict the drug-likeness of the target protein in SARS-CoV-2. A regression model using random forest was created to validate the machine learning model for computational analysis. This work was furthered by comparing the regressor model to other machine learning models using lezypredict, allowing scientists to swiftly and accurately identify and describe the SARS coronavirus strains. Keywords: SAR-CoV-1;SAR-CoV-2; compression complexity; Lempel-Ziv;Lipinski descriptor; Regression model. 1. Introduction An avalanche of DNA data is being generated by the Human Genome Project, as well as sequences in life sciences and pharmaceuticals. With so much data, software that can extract relevant information from raw DNA sequences is in high demand. ------------------------------------------------------------------------ * Corresponding author. http://asrjetsjournal.org/ https://sciprofiles.com/profile/2401869 American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 574 It's critical to have tools that can recognize protein-coding regions and predict complete genes. WHO recently proclaimed a viral pandemic known as COV19, which shares the same disease symptoms as past active coronaviruses and is the century's third significant human coronavirus [2] The COVID-19 risk level was established by the World Medical Association evaluation for the geographical and international level to “Extremely high “on February 28, 2020.” The COVID-19 infection is suspected to be a member of the influenza virus family to Beta coronavirus lineage B (Sarbecovirus) based on scientifically confirmed facts and viral protein-based similarities [14]. According to phylogenetic analyses, RNA, and the whole nucleotide of the COVID-19 viral infection and similar coronaviruses, the COVID-19 viral infection is clearly linked to two bat SARS-like coronaviruses reported in China, SARS-COV-1: Urbanni and SAR-CoV-1: BJ01[20]. According to a phylogenetic investigation of whole-genome alignment and similarity plots, the COVID-19 virus shows the greater similarity to SAR-COV-1: Urbanni and SAR-COV-1: BJ01, both of which have been accepted as alignmentsequence phylogenetic distribution [11]. The current pandemic virus nucleotide show almost 99 percent nucleotide alignment and a paucity of variability with the variant, implying a shared homologues and origin, and pointing to the human strain's recent appearance. It's uncertain if the COVID-19 viral originated independently as a unique variant capable of infecting people or as a result of recombination with previously described bat and unknown coronaviruses. According to the research done so far [14], the COVID-19 virus most likely originated in bats. [2], a Corresponding research, had made a lot of effort into using various machine learning techniques to quickly analyze multiple DNA sequences. The types of genome sequences and the numerous alignment techniques used for genomic analysis were recognized in the early study [5]. Also, they used MLDSP (guided algorithm and digital signal analysis) and a machine learningbased alignment-free algorithm were employed for genomic analysis in their work distance similarity, decision trees, genetic algorithms, clustering, and CNN networks was among the strategies utilized for the classification of new pathogens, according to their research [5]. The paper [8] suggested an Effort-To- Compress (ETC) estimate method to obtain information contained in a sequence data that is predicated on the concept of compression-complexity, as well as a compression-similarity average measure for accurate classification of SAR coronavirus dwarf varieties in a set of viruses by only using brief gene sequences remnants. Also, Single nucleotide polymorphisms (SNPs), whole-genome sequence phylogeny, protein mutations, and microsatellites was among the genetic markers employed in their investigation [3]. Discovered that the normal hidden Markov model has an inherent protein mutation problem, so the method they advanced was Hidden Markov Model (AHMM) that show to be more prolific in identifying genetic markers than the previous model. Using genomic data, this publication [6] performed a phylogenetic tree-based approach to investigate the homologues connection of SARS-CoV2 with other beta coronaviruses [8]. To generate scores for probable functional sites, a statistical technique called Reference point logistics (RPL) regression was proposed. 2. Research Methodology The dataset contains four different viruses, with one reference virus having between 27,608 and 29,751 copies. To differentiate between SAR-COV-1 and SAR-COV-2, 300 short reads were picked at random from each American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 575 virus. The K-means clustering algorithm will be used to map the selected DNA base primary sequence into 0-1 sequences using the three metrics described below [24]. Table 2.1: various methods for categorizing DNA sequences into various sets of 0-1 sequences. Seq No Steps based on sequence arranged to 0 sets mapped to 1 sets 1 Molecular weight(MW) {A,G}set to 0} Pyrimidine {{C, T}set to 1} 2 Carboxylic acid Amino{{A,C}set 0} Keto- {{G,T}set to1} 3 Strength of hydrogen electron {A,T set 0} Strong h-b electron {G,C set 1} The research method is divided into seven (7) stages which are: 2.1 Analyzed the Primary We first analyzed the primary DNA of the proposed sequence which are the under listed virus are: SAR coronavirus Urbanni (AY278741.1), SAR coronavirus BJ01 (AY278488.2) and Ebola olavirus (NC_002549.1) SAR-CoV-2: Reference genome (NC_004718.3). The initial two viruses belong to the SAR-CoV-1 variant. Then, nucleotide data was collected from Genbank database, Table 2.2: DNA Sequences and Length. Before we can evaluate complexity levels, we must first convert primary sequences to 0-1 in our study. We looked at three distinct ways for categorizing genome sequences [21] and mapping to the numbers 0 and 1 depending on their: • Molecular mass (MW) American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 576 • Carboxylic acid • Hydrogen bond electron 2.2 Complexity Measures for Lempel-Ziv (LZ) We employ Lempel-Ziv [1] compression measures to assess the distribution of short-length regions of nucleotide map sequences. Lempel–Ziv compression (LZ) is a common and universally used metric for determining how many steps the the input pattern must be compressed to an uniform length in the Non- Sequential Recursive Pair Substitution machine leaning (or a DNA of zero entropy). The compressibility of an input sequence is estimated using the complexity measure. 2.3 Measure and Identification of Distance The concatenations approach was then used to determine the compressed sequences identification distance using a computed compression complexity measure (LZ ). We had three virus genome sequences (W1, W2, and W3). Concatenation is used to create new sequences W1W2 and W2W1. The complexity measurements LZ(W1), LZ(W2), LZ(W1W2), and LZ(W1W2) are then computed (W2W1). In accordance with [7]'s methodology. The mean of the similar difference among the compression values of the two concatenated sequences W1W2 and W2W1 is proposed as a distance measure. They are mathematically described as: dLZ(W1; W2) = (LZ(W1W2) LZ(W1)) + (LZ(W2W1) LZ(W2)) (2.1) 2 dLZ(W1:W3) = (LZ(W1W3) - LZ(W1)) + (LZ(W3W1)-LZ(W3). (2.2) 2 dLZ(W2:W3) = (LZ(W2W3) - LZ(W2)) + (LZ(W2W3)-LZ(W3). (2.3) 2 The minimum distance of the set d(W1; W2); d(W1; W3); d(W2; W3) is then determined. The distances in the mathematical equations 2.1, 2.2, and 2.3 will always be non-negative and symmetric, yet genomic complexity is not uniform. The time complexity of regions with more regular genes is lower than that of regions without a American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 577 gene. For each of the three sets, the LZ distances are determined independently. The distance between the two sequences is calculated using the average values. The two SAR coronaviruses (Urbanni and BJ01) ought to have the smallest mean average (in complexities). In order to uniquely differentiate the SAR virus, the distance between Ebola and any of the SARS coronaviruses is compared to the distance between Ebola and any of the SARS coronaviruses (Urbanni and BJ01). If there are more than three sequences, this steps of display can readily be expanded. Perform Gen translation Figure 2.1: Comparative Analysis Procedure. 2.4 Analyze Short SARS-COV-1 VS. SARS-COV-2 We proceed with the research to analyze SAR-CoV-2 virus from SAR-CoV-1 virus after demonstrating the performance of LZ-based measures in effectively differentiating virus by examining extremely small portions of nucleotide sequences (the result presented in result and analysis). The sequences below are used to accomplish this. Perform Gen translation American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 578 Table 2.3: Genbank accession numbers, names, abbreviations and lengths of coronaviruses. S.NO Accessions Number Genome Abbreviations Lengths 1 AY278741 SAR-CoV-1: Urbanni Urbanni 29728 2 AY278488 SAR-CoV-1: BJ01 BJ01 29726 3 NC_004718.3 SAR-CoV-2: Reference Sequence SAR-CoV-2 29754 2.5 Machine Learning Implementation The target amino acid is extracted from the detected SARS-COV-2, and we use Lipinski's Rule for analyzing the drug-likeness of the nucleotide to predict the drug-likeness of the extracted protein. According to Christopher Lipinski, such drug-likeness is predicated on Absorption, Distribution, Metabolism, and Excretion (ADME). He also looked at all orally active FDA-approved medications in order to come up The following is rule-of-five: Molecular Structure < 500 Dalton • Hydrogen Donor < 5 • Hydrogen acceptor < 10 • Octanol-water partition (LogP) < 5 Figure 2.2: Machine Learning Implementation Chart. American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 579 2.6 Step 6: Model Performance This section describes how to classify new observations using the data set provided. The procedures in developing the model for the section, on the other hand, are divided into four (4) phases: i. Data gathering/preprocessing ii. Model training iii Model construction iv Model comparison 2.2.1 PRE-PROCESSING OF DATA Following the collection of data, the dataset is preprocessed to match the desired purpose, as stated in the previous heading. The proposed model was created with the help of the padel descriptors, which contain approximately 25400 characteristics. However, to obtain an ideal model, a bash script was used to generate the descriptor features, which were then normalized. Figure 2.3: Dataset before fingerprint descriptor. American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 580 Figure 2.4: Dataset after fingerprint descriptor. 3 .Computational Drug-Likeness Regression Models with Random Forest Using the fingerprint descriptor bioactivity data, we created an algorithm model. The random forest approach were adopted to create a regression model of SAR-COV-2 inhibitors. 3.1 Input features The SAR-Cov-2 data set contains 881 input features and 1 output variable (pIC50) values Figure 3.1: Input features. American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 581 4. Results and Analysis 4.1 SARS-COV-1 Identification For LZ measurements, the average difference between the two SAR-CoV-1 virus, BJ01 and Urbanni is the greatest. At p 0:05, BJ01 and Ubanni signal statistical significance. The complexity of a genome is not uniform. The time complexity of regions with more regular genes is higher than that of regions without a gene. Table 4.1: Pairwise average measure for the selected viruses – Ebola, BJ01 and Urbanni. We discover mean difference between segments of lengths 300 DNA bases each. Those 300 was selected at random point of the whole sequences. The average difference among the two SAR-CoV-1 virus BJ01 and Urbanni is shown to be higher for LZ. [3] Table 4.1: pairwise mean average for the three viruses – Ebola, BJ01 and Urbanni. Pair of Virus Distances: dLZ (µ ± Ꝺ) Ebola and BJ01 0.34685 ± 0.0386 Ebola and Urbanni 0.3302 ± 0.0351 BJ01 and Urbanni 0.37185 ± 0.0445* *shows significance value at p < 0:05 Figure 4.1 shows that outcome were statistically confirmed using 95 percent confidence interval plots. Based on the sample data, we can conclude that LZ can distinguish the SAR coronaviruses among the specified collection of viruses using just small contiguous segments comprises of 300 DNA bases picked at random among the whole sequence, with a total error rate of 0.05 American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 582 Figure 4.1: 95 percent confidence intervals for mean dLZ distance estimates for Ebola, Fig 4.1: 95% confidence interval for mean dLZ mean distance for the selected virus (pairwise) – Ebola, BJ01 together with Urbanni. The mean position among the two SAR-CoV-1 virus is proven to be higher for LZ. 4.2. SARS-CoV-2 Identification We expand the research to distinguish SAR-CoV-2 virus from SAR-CoV-1 virus, having established the effectiveness of LZ-mean difference successfully differentiating viruses by evaluating very small regions of DNA sequences. For this experiment, we'll use the sequences (Table 3.1). Table 4.2 shows the results. For the selected viruses – SAR-CoV-2, BJ01, and Urbanni – pairwise mean distances were calculated. We calculated an average based on 300 short, contiguous segments. These 300 bases was picked at random sites across the whole sequences. For LZ measurements, the average position among the selected SAR-CoV-1 virus BJ01 and Urbanni is the greatest. This is a statistically significant outcome. Table 4.2: pairwise average for LZ measures between the three viruses – SAR-CoV-1: BJ01, SAR-CoV-1: Urbanni and SAR-CoV-2. Pair of Viruses Distances: dLZ (µ ± Ꝺ) SAR-CoV-2 and BJ01 0.4031 ± 0.0524 SAR-CoV-2 and Urbanni 0.38375 ± 0.0474 BJ01 and Urbanni 0.37185 ± 0.0446* *shows significance value at p < 0:05 American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 583 Table 4.2 shows pairwise distances for LZ [3] measures for the selected viruses – SAR-CoV-1: BJ01, SAR- CoV-1: Urbanni, also SAR-CoV-2 – calculation of the selected DNA sequences, randomly selected from the whole genome sequence. Table 4.2 shows the mean pairwise distances (and variance).The sequence SAR-CoV-1 viral match (BJ01 and Urbanni), the LZ-pair mean position provided the least value (statistically significant). The pairwise conclusion of the outcome is deduced at p-value 0.05 interval graph as in Figure 4.2. with reference to the sample dataset, with a total error rate of 0.05, it can be deduced that LZ can discriminate between SAR-CoV-1 and SAR-CoV-2 virus adopting just small segment section comprising of 300 segments) of the whole sequence. The graph of mean average measure for LZ for the selected viruses for a one selectively segment of minimal length three hundred nucleotide shown at Figure 4.2 showing the outcome of variant length differentiating SAR-CoV-1 viruses from SARCoV-2 viruses, the graph of mean average measure for LZ of the selected virus of a pair selected lengths 300 sequences of the table Figure 4.2: Distance Measure Of Sars-Cov-2 Vs Viruses. The sequences (SARS-CoV-2 BJ01), and (SARS-CoV-2 Urbanni) and (BJ01 and Urbanni) calculated and plotted using LZ for the selected viruses for a pair selectively segment of minimal length three hundred nucleotide shown from the DNA genome. LZ accurately display the least average measure for those pair SARS- CoV-1 virus (BJ01 and Urbanni) among those three pair. This clearly indicate that the presence of SAR-COV-2 with other SAR-COV-1 make the average distance measure higher and clearly shows the distinguished features of SARS-COV-2 from SAR-COV-1 viral sequence. 4.3 Identification of Target Amino Acid Having identify the SARS-CoV-2, the DNA sequences was converted to protein American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 584 Figure 4.3: Bar plot of highly express protein. sequences in order to identify the target amino acid. The target amino acid is the highly express amino acid in the SAR-CoV-2 virus, which can either inhibit or allow modular solubility of oral vaccine. The bioactivity chemical compound that can either activate the highly express protein for possible oral solubility or to inhibit the highly express protein, using the Lipinski rule of five [13] will then be determine. 4.4 Prediction of Drug-Likeness The bioactivity of the highly expressed protein is downloaded from chemble database. We performed data Pre- Processing on the downloaded bioactivity compound, all redundant chemical was removed and the curated data was divided into two major activity “active and intermediate”. The chemical data are label as follows: The bioactivity dataset is in the IC50 unit. Chemical Compounds having bioactivities of less than 1000 nM will be determine to be active while bioactivities higher than 10,000 nM will be team to be inactive. As for those values among 1,000 and 10,000 nM will be tag intermediate 4.4.1 Lipinski descriptors I Christopher Lipinski, a chemist at Pfizer, he developed a set of rule-of-five for concluding the drug- likeness of compounds. Such drug-likeness is based on the Absorption, Distribution, Metabolism and Excretion (ADME) that is also known as the pharmacokinetic profile. The Lipinski's Rule-of-five state: Ii Molecular weight < 500 Dalton Ii Octanol-water partition coefficient (LogP) < 5 American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 585 Iii Hydrogen acceptors < 10 IV Hydrogen donors < 5 Statistical analysis | Mann-Whitney U Test was later used to test the percentage significant. The rule of the significant factor is as follows: Figure 4.4.1: (Molecular weight). Different MW 1058779.0 2.072256e-57 0.05 distribution (reject H0) Figure 4.4:1i (Log P). Lipinski rule 2: (Log P < 5) American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 586 Different Log P 10419900.0 2.318667e-61 0.05 distribution (reject H0) Figure 4.4.Iii: (NumHAcceptos). Lipinski rule 4: (Hydrogen bond acceptor < 10) 1407572.5 0.000004 0.05 Figure 4.4.Iii: (NumHDonors). Lipinski rule 3: (Hydrogen bond donors < 5) Different NumHDonors 1361005.5 2.520096e-10 0.05 distribution (reject H0) American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 587 5. Conclusion Compression-complexity metrics and grouping methods can be quite useful in the domain of NGS sequences. An entire genome sequence contains a number of intrinsic characteristics that aid in the detection of expressed genes and protein sequences, which can help with the development of oral vaccines while also allowing for proper taxonomic classification. The Lipinski description concept lays out the different scenarios for the emergence of any viral disease everywhere on the earth. This study's regressor algorithm were evaluated and equated to the performance of several algorithm approaches. Our approach appears to be more accurate than support vector machines, decision trees, and other models in terms of accuracy, which is crucial for predicting drug-likeness genomic sequences. The regressor model shows that it is more accurate than other comparison models, and thus may be enhanced and used as a backend model not just in protein sequence oral vaccine discovery, but also in other fields of bioinformatics and computational biology where data integrity is a crucial problem. This research used a clustering and complexity technique to detect viral genomes, and then used the k-means and Lempel-Ziv (LZ) algorithm to create a regressor model that might provide biological interpretation, making drug-likeness prediction and vaccine development easier. 6. Funding This research was funded by the National Institutes of Health, grant number U41HG006941; H3Africa Bioinformatics Network, H3ABioNet,. Acknowledgements The authors acknowledge the suggestions of Dr Temitayo Olagunju and Dr Taiwo Adigun in the initial review of the manuscript 7. Conflicts of Interest The authors declare no conflict of interest. The funders or the H3Africa Initiative had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. References [1] Abraham, L. & Jacob, Z. (2019). The difficulties of discrete sequencing is discussed. IEEE Transactions on Cognitive Science, 22(1):75–81 [2] Alagaili, AN., Briese, T., Mishra, N., Kapoor, V., Sameroff, SC., & deWit, E., et al.(2014). Middle east respiratory syndrome coronavirus infection in dromedary camels in Saudi Arabia. MBio. 2014; 5. https://doi.org/10. 1128/mBio.00884-14 [3] Bin Li, Yi-Bing Li, and Hong-Bo He (2005). LZ Nucleotide sequence functionality position and its implementation in phylogenetic analysis restructuring. Genomics Bioinformatics & Proteomics, http://www.h3abionet.org/ https://doi.org/10.%201128/mBio.00884-14 https://doi.org/10.%201128/mBio.00884-14 https://doi.org/10.%201128/mBio.00884-14 American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 588 3(4):206–212. [4] DR,P., Bose, P. (2021). Comparative study of Sars, Mers, Bat-sars and Sars- cov-2. News medical Life sciences [5] Gurjit, R., & Maximillian P. (2020). Machine learning using intrinsic genomic signatures for rapid classification of novel. International Journal of natural Science and Engineering Research Council of Canada.: 10:1371. [6] Hafiz., A., & Farheen., R. (2021). Comprehensive comparative genomic and microsatellite analysis of SARS, MERS, BATSARS, and COVID‐19 coronaviruses. International Journal of natural Science and Engineering Research Council of Canada, Vol 10 issue 1002. [7] Hasan, H., & Khalid, S,. (2003). A new sequence distance measure for phylogenetic tree construction. Bioinformatics, 19(16):2122–2130, 2003. [8] Karthi, B., Nithin, N. (2020). Compression-complexity measurements: Automatic identification of SARS coronavirus. International Journal of natural Science and Engineering Research Council of Canada.: Volume 3 issue 24. [9] Vijayaragavan., S.P. Kumar., B. & Ajay., p. (2000). Prediction of genetic structure in eukayotic DNA using refrence point logistic regression and sequence alignment. International Journal of Advanced Research in Computer Engineering & Technology (IJARCET). Vol 16 Issue 5 [10] Mitchell, T,. (1999). Machine learning and data mining. Communications of the ACM, 42(11), 30-36. [11] Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22(22),:4673–80. [12] Vinga, S. & Almeida, J. (2003). Alignment-free sequence comparison—a review. Bioinformatics; 19(4):513–23. [13] Christopher, L,. (2002). Capture the Untapped Value of Therapeutics. Melior Pharmaceuticals. https://www.meliordiscovery.com/christopher-lipinski/ [14] Karthi, B. & Nithin, N. (2020). Compression-complexity measurements: Automatic identification of SARS coronavirus bioRxiv preprint doi: https://doi.org/10.1101/2020.03.24.006007. [15] Kumar, S., Stecher, G., Tamura, K,. (2016). MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol; 33(7):1870–4. [16] Lempel, A., Ziv, J. (1976). On the complexity of finite sequences. IEEE Transactions on information theory;22(1):75–81. LI, M. & Paul, M. (2014). Kolmogorov complexity and its applications. Algorithms and Complexity, 1:187. Liwei Liu, Dongbo Li, and Fenglan Bai. Application of a relative Lempel-Ziv complexity to comparing biological nucleotide. Letters in Chemical Physics, 530:107–112. [17] Lu, H., Yang, L., Yan, K., Xue, Y. & Gao, Z. (2017). A cost-sensitive rotation forest algorithm for gene expression data classification. Neurocomputing; 228:270–6 [18] Lu, R., Zhao, X., Li, J., Niu, P., Yang, B. & Wu, H., et al (2020). Genomic characterization and epidemiology of novel coronavirus: implications for virus origins and receptor binding. Lancet;. https://doi.org/10.1016/S0140-6736(20)30251-8 Luk, H., Li, X., Fung, J., Lau, S. & Woo, P. (2019). Molecular epidemiology, evolution and phylogeny of SARS coronavirus. Infection, Genetics and https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Rehman,+Hafiz+Abdul https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Rehman,+Hafiz+Abdul https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Ramzan,+Farheen https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Ramzan,+Farheen https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Ramzan,+Farheen https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Ramzan,+Farheen American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS)(2023)Volume 90, No 1, pp 573-589 589 Evolution; 71: 21–30. https://doi.org/10.1016/j.meegid. [19] Maguire, P., Moser, P., Maguire, R. & Griffith, V. (2014). Is it possible to program consciousness? Using algorithmic information theory to quantify integrated data. arXiv preprint arXiv:14050126. [20] Mitchell, T,. (1999). Machine learning and data mining. Communications of the ACM, 42(11), 30-36. Ming, Li., Jonathan, H., Badger, X., Chen., Sam, K., Paul, K., & Haoyong, Z. (2001). The application of an information- based sequencing distance to the phylogeny of the entire mitochondrial genome. Bioinformatics, 17(2):149–154.