Highlights in BioScience ISSN:2682-4043 DOI:10.36462/H.BioSci.202203 Research Article Open Access 1 Biotechnology Department, Science Faculty, Helwan university, Helwan, Egypt. 2 Biotechnology Department, Agriculture Faculty, Al-Azhar university, Cairo, Egypt. * To whom correspondence should be addressed: saifeldeenmib99@gmail.com Editor: Hatem Zayed, College of Health and Sciences, Qatar University, Doha, Qatar. Reviewer(s): Jean Legeay, Université de Lorraine, Inra, UMR IAM - Interactions Arbres-Microorganismes, F- 54000 Nancy, France. Vivek Chaudhary, Motilal Nehru National Institute of Technology, Allahabad, Prayagraj, Uttar Pradesh 211004, India. Received: September 20, 2022 Accepted: December 3, 2022 Published: December 15, 2022 Citation: Ibrahim MS, Ibrahim SM. Identifying biotic stress-associated molecular markers in wheat using differential gene expression and machine learning techniques. 2022 Dec 15;5:bs202203 Copyright: © 2022 Ibrahim and Ibrahim. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and supplementary materials. Funding: The authors have no support or funding to report. Competing interests: The authors declare that they have no competing interests. Identifying biotic stress-associated molecular markers in wheat using differential gene expression and machine learning techniques Manar S. Ibrahim1 ><, Saifeldeen M. Ibrahim*2 >< Abstract Wheat is an important crop for global food security and a key crop for many devel- oping countries. Thanks to next-generation sequencing (NGS) technologies, researchers can analyze the transcriptome of wheat and reveal differentially expressed genes (DEGs) responsible for essential agronomic traits and biotic stress tolerance. In addition, machine learning (ML) methods have opened new avenues to detect patterns in expression data and make predictions or decisions based on these patterns. We used both techniques to identify potential molecular markers in wheat associated with biotic stress in six gene expression studies conducted to investigate powdery mildew, blast fungus, rust, fly larval infection, greenbug aphid, and Stagonospora nodorum infections. A total of 24,152 threshold genes were collected from different studies, with the highest threshold being 7580 genes and the lowest being 1073 genes. The study identified several genes that were differentially expressed in all comparisons and genes that were present in only one data set. The study also discussed the possible role of certain genes in plant resistance. The Ta-TLP, HBP- 1, WRKY, PPO, and glucan endo-1,3-beta-glucosidase genes were selected by the inter- pretable model-agnostic explanation algorithm as the most important genes known to play a significant role in resistance to biotic stress. Our results support the application of ML analysis in plant genomics and can help increase agricultural efficiency and production, leading to higher yields and more sustainable farming practices. Keywords: Wheat, gene expression, machine learning, biotic stress, disease Introduction Agriculture is the primary source of food production on a global scale. In the new era, sus- tainable agriculture is used to preserve or improve food quality while protecting the environment [1]. wheat (Triticum aestivum) is a member of the Poaceae family and is considered the oldest and most commonly cultivated crop [2], with an annual cultivation area of 217 million hectares. It is the most commonly grown crop in the world [3]. Also, it is one of three primary cereal crops with the greatest impact on global food security, second only to maize (Zea mays) and third only to rice [4]. Most experts assume that wheat was the first crop to be produced between 10,000 and 8,000 B.C. [2], and this is considered evidence of the importance of wheat from ancient times. Wheat is one of the most essential crops for ensuring global food security [5]. Considering the position of wheat in the world trade, wheat accounts for over 50% of the worldwide grain trade and 30% of the global grain yield [6]. And for its role in nutrition, wheat is a staple crop in more than 40 nations worldwide, feeding 82% to 85% of the world’s population [6]. According to the estimations, the annual wheat production is expected to reach 750.4 million metric tonnes in the next few years. Africa would consume 76.5 MMT of wheat, with imported grain accounting for 48.3 MMT, or 63.3 percent of total consumption [7]. Powerful yet user-friendly bioinformatics tools are required by the expanding field of gene expression profiling in order to facilitate systems-level data comprehen- sion [8]. Studies on transcriptomics have a significant potential for unique and exploratory research aimed at new insights into molecular pathways [9]. Highlights in BioScience Page 1 of 11 December 2022|Volume 5 https://doi.org/10.36462/H.BioSci.202203 https://creativecommons.org/licenses/by/4.0/ manar.samir249@science.helwan.edu.eg https://orcid.org/0000-0003-0143-2596 saifeldeenmib99@gmail.com https://orcid.org/0000-0002-6593-378X http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Next-generation sequencing (NGS) technologies greatly im- prove studies on genome-wide mRNA expression data. In com- parison to microarray technology, NGS provides higher reso- lution data and more precise transcript level measurements for gene expression studies [10]. By studying the wheat transcrip- tome, it is possible to identify the differentially expressed genes, genomic annotations, regulatory elements, molecular markers, and expression quantitative trait loci (eQTLs), as well as their sequence variants, which are responsible for significant charac- teristics [11]. With the advancement of next-generation sequencing tech- nology, the genetic diversity of wheat has recently been studied using a high-throughput, low-cost molecular technique [12].The significant genetic variation seen in the primary, secondary, and tertiary wheat gene pools provides the starting point for the breed- ing of nutrient-rich wheat genotypes [13]. A variety of molec- ular markers have been used to study the genetic diversity of wheat [12] to measure biological population divergence’s magni- tude and each component character’s proportionate contribution to the overall divergence [14]. Machine learning (ML) is a rapidly growing field of com- puter science that has the potential to revolutionize many indus- tries, including plant science. Machine learning algorithms use statistical techniques to identify patterns in data and make pre- dictions or decisions based on those patterns. This makes them well-suited to solving complex problems that are difficult for hu- mans to tackle, such as analyzing vast amounts of data from experiments or monitoring the health of crops. In plant science, machine learning can be used to predict the best time to plant crops, optimize irrigation systems, identify pests and diseases, and even breed new varieties of plants. This could help im- prove the efficiency and productivity of agriculture, leading to better yields and more sustainable farming practices. Overall, the use of machine learning in plant science has the potential to significantly advance the field and improve our understanding of plants. Consequently, in this study, the aim of our investigation is to perform expression profiling analysis of wheat to identify un- known molecular markers. We aim to use bioinformatics tools to ;(a) Study the gene expression of wheat under various envi- ronmental and developmental conditions. (b) Study the genetic diversity of these genes across the wheat genome. (c) Identify important genes and their interactions. (d) Use machine learn- ing methods to identify genes associated with wheat response to biotic stresses. By doing so, we hope to gain a better under- standing of the genetic basis of wheat response to biotic stresses, which could ultimately lead to the development of more resilient and productive varieties of wheat. This is important because bi- otic stresses, such as pests and diseases, can cause significant yield losses and can threaten the global food supply. By identi- fying molecular markers that are associated with wheat response to biotic stresses, we can improve our ability to breed new vari- eties of wheat that are better able to withstand these stresses. This could help to improve the efficiency and productivity of agriculture, leading to better yields and more sustainable farm- ing practices. Methodology Data acquisition The Gene Expression Omnibus (GEO) database, first estab- lished in 2005 [15], has continued to be a vital resource for the global scientific community. With over 20 years of experience, GEO remains the most widely used public repository for high- throughput gene expression and other functional genomics data sets. This valuable resource allows researchers to quickly and easily access a wide range of raw and processed data, along with detailed experimental descriptions, at no cost. The con- venience and accessibility of GEO make it an invaluable tool for advancing scientific research. In this study, GEO has been used to download the data of six different experiments on wheat for further analysis. These experiments have been conducted un- der different biotec stresses. such as the response to powdery mildew infection, the response to blast fungus magnaporthe, the resistance pathways of transgenic wheat lines, the response to a Hessian fly larval attack, the response to greenbug aphid feed- ing, and the response to the necrotrophic effector SnTox3. All GSE accessions are provided with the experiment name of each accession in (table 1). Differential gene expression analysis With the help of the R/Bioconductor and Limma package, the online statistical tool GEO2R [16] was used to analyze the raw six gene expression data sets of wheat under different bi- otic stresses from GEO datasets [17] (table1). The samples have been clustered into two groups (control and infected / treatment). Differential expression genes were analyzed using R packages to create a volcano plot highlighting all significant genes and a heatmap of the top differentially expressed genes based on each group in all samples according to log2FC (fold change) ≥ 1 and adjusted p-value < 0.05 as the threshold for DEGs. Whereby up-regulated DEGs were considered if the logFC (fold change) ≥ 1 and down-regulated DEGs were considered if the logFC ≤ -1. A volcano plot shows statistical significance (P value) versus magnitude of change (fold change). It enables quick visual iden- tification of genes with large fold changes that are also statisti- cally significant. These may be the most biologically significant genes [18]. First, a volcano plot was generated using the R pack- age (ggplot2) after reading the data and filtering it according to the adjusted p-value and logFC (blue: down-regulated, red: up- regulated). Second, a heatmap of the top differentially expressed genes in the RNA-seq data set was then generated. To do this, we need to extract the differentially expressed genes from the DE results file. The heatmap for the gene expression data was generated using R package (pheatmap) after filtering the top dif- ferential gene expression from the data. Then, a Venn diagram was used to show all the interrelationships of the six wheat exper- iments. The Venn diagram of the overlapping DEGs was output Highlights in BioScience Page 2 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat by the interactiVenn website [19]. Protein-protein interaction and and functional enrichment analyses The STRING database was utilised for protein-protein inter- action analysis [20] to evaluate the link between genes associ- ated with wheat gene expression. The STRING database was chosen because this database seeks to integrate all known and predicted relationships between proteins, including functional and physical interactions [20]. All information about GO, an- notated keywords, and protein domains for each GSE dataset has been provided in (table 2) Machine Learning Model The results of the differential gene expression study of pu- tative biotic-associated gene biomarkers were employed as at- tribute values for machine learning (ML) analysis. We retrieved gene expression data from genes that were strongly expressed in gene expression. Their expression data were utilised as training and validation sets for the ML analysis. The gene expression data were adjusted before to the ML analysis. The sklearn and lime0.2.0.1 Python libraries were used to run the "Extra-tree re- gressor" and local interpretable model-agnostic explanations al- gorithms. Results and Discussion Identification of DEGs A total of 24,152 threshold genes were collected from dif- ferent studies. The highest threshold of genes was collected by GSE27320, with 7580 genes, while GSE31760 had the lowest threshold of genes, with only 1073 genes. The volcano plot indicates the upregulated (red color) and downregulated (blue color) genes in Triticum aestivum samples. The horizontal axis represents the fold change (log2FC), and the vertical axis rep- resents the adjusted p-values. A total of 3701 upregulated and 3879 downregulated DEGs were identified from the GSE27320 dataset, while 241 upregulated and 832 downregulated DEGs were identified from the GSE31760 dataset. GSE32151 yielded 3533 upregulated and 5138 downregulated DEGs, GSE34445 yielded 658 upregulated and 1442 downreg- ulated DEGs, GSE45995 yielded 2054 upregulated and 1459 downregulated DEGs, and GSE59723 yielded 74 upregulated and 1141 downregulated DEGs. The volcano plot of each gene expression profile data is shown in (Figure 1). Users can use heatmaps to examine the expression of a sub- set of genes. This can provide helpful insight into the expression of various groups and samples without losing sight of the larger study or losing clarity when evaluating patterns averaged over hundreds of genes at once [18]. According to the adjusted p- value ranking, a heatmap is created for the top DE genes.The heatmap clusters the samples into control (blue) and infected / treatment (red) groups. Red indicates samples with relatively high gene expression, while blue indicates samples with rela- tively low gene expression. Genes with intermediate expression levels are represented by lighter tones and white. A dendrogram hierarchical clustering has been used to reorder the samples and genes (Figure 2). Venn diagrams of the DEGs between the integrated six GEO data sets are shown in (Figure 2). The number in each circle indicates the quantity of differentially expressed genes across the various comparisons. The overlapping number refers to genes that are differently expressed across all comparisons, while the non-overlapping number refers to genes that are unique to each sample. Genes were identified as being up- or down-regulated. The GSE32151 and GSE45995 data sets shared 12 genes, while the GSE45995 and GSE59723 data sets shared 6 genes (PDI2, pepc, gstf5, gstf5:1, and ltp9.2). According to the study of Guo-Tian Liu et al.[21], gstf5 is a member of the (GSTs) family, which is responsible for catalysing the conjugation of the reduced form of glutathione to a number of electrophilic substrates, they also reported that GSTs contribute to resistance against powdery mildew. The GSE31760, GSE45995, and GSE59723 data sets shared 3 genes (rgp, PR4, and pox3). The relationship between methyl jasmonate (MeJA) and the expression profiles of nine pathogenesis- related protein genes (PR genes) was examined in the Zongbiao Duan et al. [22] study to determine the interactive role in pow- dery mildew resistance. This investigation revealed that PR4 is one of the pathogenesis-related protein genes (PR genes), and the expressions of PR4 and other PR genes were most signif- icantly activated by MeJA and showed a significant resistance to powdery mildew. The GSE32151 and GSE59723 datasets shared 1 gene (Xip-R1). Xip-R1 is one of the xylanase inhibitors. R.-J. SUN et al. [23], xylanase inhibitors (XIs) are plant cell wall proteins found mostly in monocots that limit microbial xy- lanases’ hemicellulose degrading ability. Silvio Tundo et al. [24]. The GSE34445 and GSE59723 data sets shared 4 genes (WRKY45, Cht2, WRKY, and ccd1), and the GSE31760 and GSE59723 shared 2 genes (TaAOS and tamdr1). The allene oxide synthase (TaAOS) gene has been identified as being en- gaged in the JA signalling system, which increases plant resis- tance to Fusarium head blight (FHB) [25], Tamdr1 has also been identified as a Fusarium head blight resistance gene [26]. The GSE27320 and GSE31760 shared 2 genes (ald and pSBGer4), and the GSE27320 and GSE32151 data sets shared 2 genes ( hsp16.9-3LC2 and S276 ) . The GSE32151 and GSE34445 data sets shared 1 gene (TaAML15), and the GSE27320 and GSE34445 data sets shared 3 genes (Ss1, wpi6, and gstf6b). The GSE34445 and GSE45995 data sets shared 4 genes (TaGlb2b, TaAKT1, TaMRP2, and omet). Pratiksha Singh et al. [27] re- ported that TaGlb2b showed a significant response to Erysiphe graminis and Fusarium graminearum. The GSE27320, GSE31760, and GSE34445 data sets shared 1 gene (HSP101c) which has an important role in heat tolerance in hexaploid wheat [28]. The GSE27320 and GSE45995 data sets shared 13 genes (Tra2, Tra2:1, TAc23, zip1, TaGI1, PsbP, ctpA, Wcor726, pre-FBPase, lgul, GAPN, 6-FEH, and lgul:1). In a study by Yu Liu et al. to examine the response of PsbP to Highlights in BioScience Page 3 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Table 1. Information for the Six GEO datasets for wheat. Unique identifier for the dataset within GEO (accession), full name of the experiment (experiment name), shortened version of the experiment name (short name), ans the abbreviated version of the experiment name (abbrev). accession experiment name short name abbrev GSE27320 Expression data in wheat (T. aestivum L.) near isogenic lines in response to powdery mildew infection wheat responses to powdery mildew WRPM GSE31760 Transcription profiling wheat responses to adapted and non-adapted isolates of the blast fungus, Magnaporthe wheat responses to blast fungus WRBF GSE32151 Lr1-mediated leaf rust resistance pathways of transgenic wheat lines mediated leaf rust resistance MLRR GSE34445 Expression data from wheat following Hessian fly larval attack. wheat following fly larval WFFL GSE45995 Transcriptomics of induced defense responses to Greenbug aphid feeding in near isogenic wheat lines wheat responses to Greenbug aphid WRGA GSE59723 Transcriptional analyses of wheat responses to the necrotrophic effector SnTox3 wheat responses to SnTox3 WRST Table 2. The SRING database table provides information about GO, annotated keywords, and protein domains for each GSE dataset. GSE Category ID Description FDR GSE27320 Molecular Function GO:0005200 Structural constituent of cytoskeleton 0.0046 Cellular Component GO:0110165 Cellular anatomical entity 2.86E-24 KEGG Pathways map04145 Phagosome 0.0343 Annotated Keywords KW-0963 Cytoplasm 0.00025 KW-0206 Cytoskeleton 0.0015 KW-0493 Microtubule 0.0016 KW-0342 GTP-binding 0.0065 KW-0809 Transit peptide 0.0256 Protein Domains PF03953 Tubulin C-terminal domain 0.00055 PF00091 Tubulin/FtsZ family, GTPase domain 0.0012 Protein Features IPR000217 Tubulin 2.43E-05 IPR002453 Beta tubulin 2.43E-05 IPR003008 Tubulin/FtsZ, GTPase domain 2.43E-05 IPR008280 Tubulin/FtsZ, C-terminal 2.43E-05 IPR013838 Beta tubulin, autoregulation binding site 2.43E-05 IPR017975 Tubulin, conserved site 2.43E-05 IPR018316 Tubulin/FtsZ, 2-layer sandwich domain 2.43E-05 IPR023123 Tubulin, C-terminal 2.43E-05 IPR037103 Tubulin/FtsZ, C-terminal superfamily 2.43E-05 IPR036525 Tubulin/FtsZ, GTPase domain superfamily 2.71E-05 IPR000877 Proteinase inhibitor I12, Bowman-Birk 0.0046 Protein Domains SM00269 Bowman-Birk type proteinase inhibitor 0.004 GSE31760 Cellular Component GO:0110165 Cellular anatomical entity 0.001 GSE32151 Cellular Component GO:0110165 Cellular anatomical entity 6.46E-37 Annotated Keywords KW-0325 Glycoprotein 3.38E-05 KW-1015 Disulfide bond 0.00015 KW-0732 Signal 0.00024 KW-0809 Transit peptide 0.0023 KW-0963 Cytoplasm 0.0023 KW-0326 Glycosidase 0.0028 KW-0378 Hydrolase 0.0062 KW-0676 Redox-active center 0.0206 KW-0119 Carbohydrate metabolism 0.0217 KW-0624 Polysaccharide degradation 0.0386 Protein Features IPR000877 Proteinase inhibitor I12, Bowman-Birk 0.00055 Protein Domains SM00269 Bowman-Birk type proteinase inhibitor 4.17E-05 GSE34445 Cellular Component GO:0110165 Cellular anatomical entity 1.42E-07 GSE45995 Cellular Component GO:0110165 Cellular anatomical entity 1.53E-19 Subcellular localization GOCC:0005576 Extracellular region 0.0056 Annotated Keywords KW-0809 Transit peptide 0.0011 KW-0150 Chloroplast 0.0022 KW-0732 Signal 0.0039 KW-0676 Redox-active center 0.0081 KW-0597 Phosphoprotein 0.0241 KW-0963 Cytoplasm 0.0338 Protein Domains SM00269 Bowman-Birk type proteinase inhibitor 0.004 GSE59723 Cellular Component GO:0110165 Cellular anatomical entity 1.75E-08 Annotated Keywords KW-1015 Disulfide bond 0.00015 KW-0732 Signal 0.00097 KW-0325 Glycoprotein 0.0088 KW-0326 Glycosidase 0.0214 Protein Features IPR000877 Proteinase inhibitor I12, Bowman-Birk 0.0199 Protein Domains SM00269 Bowman-Birk type proteinase inhibitor 0.0016 Table 3. The results of the machine learning model represent the most signifi- cant genes with their corresponding values. Accsession Marker P/N Value Gene.Symbol GSE27320 Ta.85.1.S1 positive 28944.5 PsbP Ta.1842.1.S1_a positive 309.3 WPEAMT Ta.1848.2.S1 positive 4463.8 pip1:2 Ta.2907.1.S1 positive 4340.4 LOC543101 Ta.10.1.S1_a positive 47759.5 LOC543334 TaAffx.120000.1.S1 positive 465.7 TLK1 Ta.10.2.S1 positive 63930.8 LOC543334 Ta.21350.2.S1 positive 216.2 wrsi5-1 Ta.21348.1.S1 positive 292.1 LOC543233 Ta.23758.1.S1 positive 265.9 Wcor518 GSE31760 Ta.28.1.S1 negative 16942 LOC543330 Ta.27762.1.S1 negative 3687.7 Ta-TLP Ta.24501.1.S1 negative 2852.3 LOC543292 Ta.2788.1.A1 negative 768.65 1-SST Ta.82.1.S1 negative 8065.5 LOC543287 TaAffx.115935.1.S1 negative 375.52 LOC543157 Ta.87.1.S1 negative 654.58 pSBGer4 Ta.24254.3.S1 negative 333.75 LOC606342 Ta.25053.1.S1 negative 5220.3 LOC542887 Ta.22673.1.S1_s negative 14994 LOC543498 GSE32151 Ta.9320.1.S1 negative 11.7 CCoAMT Ta.9320.1.S1 negative 11.81 CCoAMT Ta.9402.1.S1 negative 8.97 LOC542918 Ta.8629.1.A1 negative 5.9 WLTP1 TaAffx.39351.2.A1 negative 9.64 HrBP1-1 Ta.9402.1.S1 negative 8.92 LOC542918 Ta.24806.1.S1 negative 10.79 WLIP19 Ta.28700.1.S1 negative 4.55 CHS Ta.2907.1.S1 negative 11.76 LOC543101 Ta.28734.1.S1 negative 15.08 TaGRP2 GSE34445 Ta.8614.1.S1 negative 46.58 WRKY45 Ta.27312.1.S1 negative 63.75 AMT1 Ta.4050.1.S1 negative 136.86 LOC100037581 Ta.203.1.S1 negative 86.99 LOC543416 Ta.1058.1.S1 negative 11290.8 SAMDC1 Ta.192.1.S1 negative 79 LOC543235 Ta.4678.2.S1 negative 61.95 WRKY71 Ta.217.1.S1 negative 1033.27 LOC543244 Ta.1058.3.S1 negative 15078.5 SAMDC1 TaAffx.129134.2.S1 negative 281.44 a2b GSE45995 Ta.28.1.S1 negative 37166.6 LOC543330 Ta.9226.1.S1 negative 41349.6 PR4 Ta.2784.1.A1 negative 42724.1 Chi 1 Ta.14183.1.S1 negative 304.86 LOC543077 Ta.21342.1.S1 negative 34995.6 Chi 3 Ta.22871.1.S1_s negative 13472.6 gamma-TIP TaAffx.39351.2.A1 negative 947.06 HrBP1-1 Ta.27762.1.S1 negative 31690.6 Ta-TLP Ta.278.1.S1 negative 41378.7 LOC543422 Ta.278.1.S1 negative 41179.1 LOC543422 GSE59723 Ta.5428.1.S1 negative 7.41 acT3 Ta.8228.1.S1 negative 4.69 LOC543097 TaAffx.128418.102.S1 negative 2.53 ltp9.2 Ta.706.1.S1_s negative 7.5 g6pdh Ta.81.1.S1 negative 6.86 pepc Ta.24501.1.S1 negative 10.2 LOC543292 Ta.4725.1.S1 negative 5.08 WRKY53-b Ta.27762.1.S1 negative 7.64 Ta-TLP Ta.234.1.S1 negative 4.48 LCT1 TaAffx.115935.1.S1 negative 5.11 LOC543157 Highlights in BioScience Page 4 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Figure 1. Visualization of DEGs volcano plots.The representations are as follows: x-axis, log2FC; y-axis, -log10 of an adjusted p-value. Highlights in BioScience Page 5 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Figure 2. Heatmap and Venn diagram of the top DE genes data Highlights in BioScience Page 6 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Figure 3. PPI networks show the interaction of DEGs. Highlights in BioScience Page 7 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Figure 4. Machine learning results and venn diagrams of the markers, genes, and gene titles that were shared between GSEs. Highlights in BioScience Page 8 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat wheat dwarf virus (WDV), they noticed that the PsbP gene was down-regulated under WDV infection [29]. The GSE27320, GSE45995, and GSE59723 data sets shared 3 genes (pox4, wrsi5-1, and WAS-2). The GSE27320, GSE31760, GSE45995, and GSE59723 data sets shared 2 genes (Ta-TLP and Chi 3). The GSE27320, GSE32151, and GSE45995 data sets shared 14 genes, while the GSE27320, GSE32151, GSE45995 , and GSE59723 data sets shared 5 genes (Taxi-III, Taxi-III:1, wali6, Taxi-IV, and POX2). The GSE27320, GSE34445, and GSE45995 data sets shared 1 gene (Wcab) which was highly significant in the response to stripe rust in wheat [30]. The GSE27320, GSE31760, GSE34445, and GSE45995 data sets shared 1 gene (rbcl), in an investigation by Ceyda Ozfidan -Konakci et al., the rbcl gene was observed at higher transcription levels in the chloroplasts of wheat as a response to hydrogen sulphide (H2S) and nitric oxide (NO) alleviate cobalt toxicity. The GSE31760, GSE34445, and GSE45995 data sets shared 1 gene (prx), and the GSE27320, GSE34445, and GSE59723 data sets shared 1 gene (FKBP77) which has a high expression under heat stress [31]. The GSE27320, GSE32151, GSE34445, and GSE59723 data sets shared 1 gene (taxi-I). GSE32151, GSE34445, GSE45995, and GSE59723 data sets shared 3 genes (xipI:1, Tamyb7, and xipI). The Tamyb7 is one of the MYB family, which is considered to have a significant role in resisting stripe rust [32]. GSE27320, GSE34445, GSE45995, and GSE59723 data sets shared 1 gene (WRN2). GSE31760, GSE34445, GSE45995, and GSE59723 data sets shared 1 gene (S85). The GSE31760, GSE32151, GSE34445, and GSE59723 data sets shared 1 gene (WCK-1), WCK-1 ex- pression is reported to be induced by the fungal elicitor cal- cium ionophore A23187 as well as drought [33]. GSE31760, GSE34445, and GSE59723 data sets shared 2 genes (Cht4 and gstu2), Cht-4 expression was substantially faster in the resistant cultivar than in the susceptible one, in the response to Fusar- ium graminearum in wheat [34]. GSE32151, GSE45995, and GSE59723 data sets shared 4 genes (Tamyb7:2, wali5, Chi 2, and Tamyb7:1), Guozhang Kang et al. [35] discovered that freezing stress increased the expression of the wali5 protein gene. GSE31760, GSE32151, GSE45995, and GSE59723 data sets shared 2 genes (WRKY53-b and Chi 1). The GSE31760, GSE32151, and GSE45995 data sets shared 1 gene (DBP), and the GSE27320, GSE32151, and GSE34445 data sets shared 2 genes (AMT1 and AMT1:1). PPI network construction Protein-protein interaction (PPI) networks were created us- ing the STRING tool. Six modules were identified in this con- structed network, which was made up of 115 nodes and 56 edges. The top significant module was GES32151, which was com- posed of 44 nodes and 30 edges, and had a PPI enrichment p- value of 0.0565. The lowest module was GSE34445, which was made up of only 11 nodes, 1 edge, and had a PPI enrichment p-value of 0.174 (Figure 3). Machine learning The machine learning model identified 53 markers, 48 genes, and 46 gene titles, as shown in (table 3). Also see (Figure 4). Each GSE accession consisted of ten markers. Only GSE27320 had positive values, ranging from 216.2 to 63930.8. The rest of the GSEs had negative values, as the table shows. Some of the GSEs shared the same markers, genes, and gene titles. Accord- ing to the venn diagram (figure 4) of the markers, GSE32151 and GSE45995 shared TaAffx.39351.2.A1. GSE31760, GSE45995, and GSE59723 shared Ta.27762.1.S1. GSE31760 and GSE59723 shared Ta.24501.1.S1 and TaAffx.115935.1.S1. GSE27320 and GSE32151 shared Ta.2907.1.S1. GSE31760 and GSE45995 shared Ta.28.1.S1. For genes, GSE32151 and GSE45995 shared HrBP1-1. GSE31760 , GSE45995 , and GSE59723 shared Ta-TLP. GSE31760 and GSE59723 shared LOC543292 and LOC543157. GSE27320 and GSE32151 shared LOC543101. GSE31760 and GSE45995 shared LOC543330. For gene titles, GSE32151 and GSE45995 shared harpin binding protein 1. GSE31760, GSE45995, and GSE59723 shared thaumatin-like protein. GSE34445 and GSE59723 shared WRKY transcription factor. GSE31760 and GSE59723 shared polyphenol oxidase. GSE27320 and GSE32151 shared methylmalonate semialdehyde dehydrogenase. GSE31760 and GSE45995 shared glucan endo-1 and 3-beta-D-glucosidase. Danielle et al. [36] mention in their study that the Ta-TLP gene is a member of pathogenesis-related proteins and has an important role in plant resistance to powdery mildew. Another study by Zhi-Hui He et al. [37] reported that harpin-binding protein 1 is thought to be involved in plant disease resistance and drought resilience because it is responsible for plastid glutamine synthase (GS). It was shown to be down-regulated in wheat’s single seed descent line 10 (SSDL 10). Furthermore, Thaumatin- like proteins (TLPs), as reported by Weibo Sun et al. [38], play roles in plant resistance to pathogen stress by acting as a positive factor in transgenic poplars with enhanced resistance to spots disease. For WRKY transcription factors, Patel P et al [39] men- tioned in their study that WRKY transcription factors are thought to have a role in contributing to the level of thermotolerance. It was reported [39] that polyphenol oxidase (PPO) has a role in plant resistance to osmotic stress-tolerant bacteria. They found that bacterized seedlings showed a slight improvement in wheat roots and shoots, and the polyphenol oxidase was greater in the roots and shoots. Hegedus et al. [40] reported that the overex- pression of glucan endo-1,3-beta-glucosidase has been linked to a variety of physiological and developmental processes, as well as resistance to biotic and abiotic stress. This suggests that these genes may play an important role in plant resistance to various stresses and diseases. Highlights in BioScience Page 9 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat Reference 1. Gimenez E, Salinas M, Manzano-Agugliaro F. World- wide research on plant defense against biotic stresses as improvement for sustainable agriculture. Sustainability. 2018;10(2):391. 2. Igrejas G, Branlard G. The importance of wheat. In: Wheat quality for improving processing and human health. Springer; 2020. p. 1-7. 3. Erenstein O, Jaleta M, Mottaleb KA, Sonder K, Donovan J, Braun HJ. Global Trends in Wheat Production, Consump- tion and Trade. In: Wheat Improvement. Springer; 2022. p. 47-66. 4. Hayta S, Smedley MA, Clarke M, Forner M, Harwood WA. An efficient Agrobacterium-mediated transformation proto- col for hexaploid and tetraploid wheat. Current Protocols. 2021;1(3):e58. 5. Zhou Y, Zhao X, Li Y, Xu J, Bi A, Kang L, et al. Triticum population sequencing provides insights into wheat adapta- tion. Nature genetics. 2020;52(12):1412-22. 6. Poudel PB, Poudel MR. Heat stress effects and tolerance in wheat: A review. Journal of Biology and Today’s World. 2020;9(3):1-6. 7. Jasrotia P, Kashyap PL, Bhardwaj AK, Kumar S, Singh G. Scope and applications of nanotechnology for wheat pro- duction: A review of recent advances. Wheat Barley Res. 2018;10(1):1-14. 8. Zhou G, Soufan O, Ewald J, Hancock RE, Basu N, Xia J. NetworkAnalyst 3.0: a visual analytics platform for com- prehensive gene expression profiling and meta-analysis. Nu- cleic acids research. 2019;47(W1):W234-41. 9. Kumar Kushwaha S, Vetukuri RR, Odilbekov F, Pareek N, Henriksson T, Chawade A. Differential gene expression analysis of wheat breeding lines reveal molecular insights in yellow rust resistance under field conditions. Agronomy. 2020;10(12):1888. 10. Wang T, Li B, Nelson CE, Nabavi S. Comparative analysis of differential gene expression analysis tools for single-cell RNA sequencing data. BMC bioinformatics. 2019;20(1):1- 16. 11. Paul S, Duhan JS, Jaiswal S, Angadi UB, Sharma R, Raghav N, et al. RNA-Seq Analysis of Developing Grains of Wheat to Intrigue Into the Complex Molecular Mechanism of the Heat Stress Response. 2022. 12. Yang X, Tan B, Liu H, Zhu W, Xu L, Wang Y, et al. Ge- netic diversity and population structure of Asian and Eu- ropean common wheat accessions based on genotyping-by- sequencing. Frontiers in Genetics. 2020;11:580782. 13. Velu G, Crespo Herrera L, Guzman C, Huerta J, Payne T, Singh RP. Assessing genetic diversity to breed competitive biofortified wheat with enhanced grain Zn and Fe concen- trations. Frontiers in Plant Science. 2019;9:1971. 14. Devesh P, Moitra P, Shukla R, Pandey S. Genetic diversity and principal component analyses for yield, yield compo- nents and quality traits of advanced lines of wheat. Journal of Pharmacognosy and Phytochemistry. 2019;8(3):4834-9. 15. Barrett T, Suzek TO, Troup DB, Wilhite SE, Ngau WC, Ledoux P, et al. NCBI GEO: mining millions of expres- sion profilesdatabase and tools. Nucleic acids research. 2005;33(suppl_1):D562-6. 16. Zhang X, Wang Z, Hu L, Shen X, Liu C. Identification of potential genetic biomarkers and target genes of peri- implantitis using bioinformatics tools. BioMed Research International. 2021;2021. 17. Udhaya Kumar S, Thirumal Kumar D, Bithia R, Sankar S, Magesh R, Sidenna M, et al. Analysis of differentially expressed genes and molecular pathways in familial hy- percholesterolemia involved in atherosclerosis: a system- atic and bioinformatics approach. Frontiers in Genetics. 2020;11:734. 18. Liu S, Wang Z, Zhu R, Wang F, Cheng Y, Liu Y. Three differential expression analysis methods for RNA sequenc- ing: limma, EdgeR, DESeq2. JoVE (Journal of Visualized Experiments). 2021;(175):e62528. 19. Sivasakthi P, Sabarathinam S, Vijayakumar TM. Network pharmacology and in silico pharmacokinetic prediction of Ozanimod in the management of ulcerative colitis: A com- putational study. Health Science Reports. 2022;5(1). 20. Szklarczyk D, Gable AL, Nastou KC, Lyon D, Kirsch R, Pyysalo S, et al. The STRING database in 2021: customiz- able protein–protein networks, and functional characteriza- tion of user-uploaded gene/measurement sets. Nucleic acids research. 2021;49(D1):D605-12. 21. Liu GT, Wang BB, Lecourieux D, Li MJ, Liu MB, Liu RQ, et al. Proteomic analysis of early-stage incompatible and compatible interactions between grapevine and P. viticola. Horticulture research. 2021;8. 22. Duan Z, Lv G, Shen C, Li Q, Qin Z, Niu J. The role of jas- monic acid signalling in wheat (Triticum aestivum L.) pow- dery mildew resistance reaction. European journal of plant pathology. 2014;140(1):169-83. 23. Sun RJ, Xu Y, Hou CX, Zhan YH, Liu MQ, Weng XY. Ex- pression and characteristics of rice xylanase inhibitor Os- XIP, a member of a new class of antifungal proteins. Biolo- gia plantarum. 2018;62(3):569-78. Highlights in BioScience Page 10 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Ibrahim and Ibrahim, 2022 Identifying biotic stress associated molecular markers in wheat 24. Tundo S, Mandalà G, Sella L, Favaron F, Bedre R, Kalunke RM. Xylanase Inhibitors: Defense Players in Plant Immunity with Implications in Agro-Industrial Pro- cessing. International Journal of Molecular Sciences. 2022;23(23):14994. 25. Hu C, Chen P, Zhou X, Li Y, Ma K, Li S, et al. Arms Race between the Host and Pathogen Associated with Fusarium Head Blight of Wheat. Cells. 2022;11(15):2275. 26. Wang Q, Shao B, Shaikh FI, Friedt W, Gottwald S. Wheat resistances to Fusarium root rot and head blight are both associated with deoxynivalenol-and jasmonate-related gene expression. Phytopathology. 2018;108(5):602-16. 27. Singh P, Song QQ, Singh RK, Li HB, Solanki MK, Yang LT, et al. Physiological and molecular analysis of sugarcane (Va- rietiesF134 and NCo310) during Sporisorium scitamineum interaction. Sugar Tech. 2019;21(4):631-44. 28. Erdayani E, Nagarajan R, Grant NP, Gill KS. Genome-wide analysis of the HSP101/CLPB gene family for heat toler- ance in hexaploid wheat. Scientific reports. 2020;10(1):1- 17. 29. Liu Y, Liu Y, Spetz C, Li L, Wang X. Comparative transcrip- tome analysis in Triticum aestivum infecting wheat dwarf virus reveals the effects of viral infection on phytohormone and photosynthesis metabolism pathways. Phytopathology Research. 2020;2(1):1-13. 30. Lata C, Prasad P, Gangwar O, Adhikari S, Thakur RK, Savadi S, et al. Temporal behavior of wheat–Puccinia striiformis interaction prompted defense-responsive genes. Journal of Plant Interactions. 2022;17(1):674-84. 31. Vishwakarma H, Junaid A, Manjhi J, Singh GP, Gaikwad K, Padaria JC. Heat stress transcripts, differential expres- sion, and profiling of heat stress tolerant gene TaHsp90 in Indian wheat (Triticum aestivum L.) cv C306. PloS one. 2018;13(6):e0198293. 32. Al-Attala MN. The role of Some MYB genes in defense response of wheat against stripe rust pathogen. Egyptian Journal of Desert Research. 2019;69(3):113-29. 33. Liu Y, Yu TF, Li YT, Zheng L, Lu ZW, Zhou YB, et al. Mitogen-activated protein kinase TaMPK3 suppresses ABA response by destabilising TaPYL4 receptor in wheat. New Phytologist. 2022;236(1):114-31. 34. Sorahinobar M, Niknam V, Jahedi A, Ebrahimzadeh H, Moradi B, Behmanesh M, et al. Application of sodium salicylate up-regulates defense response against Fusar- ium graminearum in wheat spikes. Biologia plantarum. 2019;63:690-8. 35. Kang G, Li G, Yang W, Han Q, Ma H, Wang Y, et al. Tran- scriptional profile of the spring freeze response in the leaves of bread wheat (Triticum aestivum L.). Acta physiologiae plantarum. 2013;35(2):575-87. 36. Ekom DCT, Benchekroun MN, Udupa SM, Iraqi D, Ennaji MM, et al. Wheat Genetic Transformation as Efficient Tools to Fight against Fungal Diseases. Journal of Agricultural Science and Technology A. 2015;5(3):153-61. 37. He ZH, Li HW, Shen Y, Li ZS, Mi H. Comparative analysis of the chloroplast proteomes of a wheat (Triticum aestivum L.) single seed descent line and its parents. Journal of plant physiology. 2013;170(13):1139-47. 38. Sun W, Zhou Y, Movahedi A, Wei H, Zhuge Q. Thaumatin- like protein(Pe-TLP)acts as a positive factor in transgenic poplars enhanced resistance to spots disease. Physiological and Molecular Plant Pathology. 2020;112:101512. 39. Patel P, Patil T, Maiti S, Paul D, Amaresan N. Screening of osmotic stress-tolerant bacteria for plant growth promo- tion in wheat (Triticum aestivum L.) and brinjal (Solanum melongena L.) under drought conditions. Letters in Applied Microbiology. 2022;75(5):1286-92. 40. Hegedus G, Kutasy B, Kiniczky M, Decsi K, Juhász Á, Nagy Á, et al.. Liposomal Formulation of Botanical Ex- tracts may Enhance Yield Triggering PR Genes and Phenyl- propanoid Pathway in Barley (Hordeum vulgare). Plants 2022, 11, 2969. s Note: MDPI stays neutral with regard to jurisdictional claims in published ; 2022. Highlights in BioScience Page 11 of 11 December 2022|Volume 5 http://bioscience.highlightsin.org/ Abstract Introduction Methodology Data acquisition Differential gene expression analysis Results and Discussion Identification of DEGs PPI network construction Machine learning