{"paper_id":"672c902e-160d-4583-ae32-525777fd51ad","body_text":"RESEARCH Open Access\nMiddle East Fertility\nSociety Journal\n© The Author(s) 2026. Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, \nsharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and \nthe source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this \narticle are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included \nin the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will \nneed to obtain permission directly from the copyright holder. To view a copy of this licence, visit  h t t p  : / /  c r e a  t i  v e c  o m m  o n s .  o r  g / l i c e n s e s / b y / 4 . 0 /.\nShahgholi et al. Middle East Fertility Society Journal           (2026) 31:29 \nhttps://doi.org/10.1186/s43043-026-00310-8\n*Correspondence:\nZahra Noormohammadi\nmarjannm@yahoo.com\n1Department of Biology, SR.C., Islamic Azad University, Tehran, Iran\n2Department of Endocrinology and Female Infertility, Reproductive \nBiomedicine Research Center , Royan Institute for Reproductive \nBiomedicine, ACECR, Tehran, Iran\n3Breast Disease Research Center (BDRC), Tehran University of Medical \nScience, Tehran, Iran\n4Department of Obstetrics and Gynecology, Arash Women’s Hospital, \nTehran University of Medical Sciences, Tehran, Iran\n5Department of Molecular Medicine, Biotechnology Research Center, \nPasteur Institute of Iran, Tehran, Iran\nAbstract\nObjective Endometriosis is a multifactorial inflammatory disease characterized by the growth of endometrial-like \ntissue outside the uterus, frequently causing chronic pelvic pain and infertility. This study aimed to identify conserved \ndifferentially expressed genes (DEGs) and robust hub genes across heterogeneous cohorts to elucidate key molecular \nmechanisms in endometriosis pathogenesis.\nMethods Transcriptomic data from four independent microarray datasets (GSE7305, GSE120103, GSE23339, \nGSE51981) were analyzed using the Limma package in R to identify DEGs. Common DEGs were intersected across \ndatasets, and a protein-protein interaction (PPI) network was constructed using STRING and visualized in Cytoscape \n(version 3.10.4). Hub genes were selected through a multi-metric approach in cytoHubba (MCC, Degree, EPC, DMNC). \nFunctional enrichment (GO and pathway) and GeneMANIA network analysis were performed to explore biological \nroles.\nResults Intersection of DEGs revealed conserved expression signatures despite cohort heterogeneity. A consensus \nset of 20 hub genes was identified, with 12 (60%) functioning as transcription factors. Enrichment analysis highlighted \ntranscriptional regulation, DNA binding, developmental processes (e.g., tissue development, pattern specification, \nembryonic morphogenesis), and epithelial barrier components (e.g., tight junctions). Pathway analysis implicated \nnuclear receptor signaling, developmental biology pathways, and WNT signaling.GeneMANIA analysis confirmed \nstrong co-expression and physical interactions among hub genes, particularly in transcriptional and developmental \nfunctions.\nConclusion The conserved hub genes, enriched in transcription factors, suggest central roles for transcriptional \ndysregulation and developmental pathways in endometriosis across diverse populations. These findings provide \nrobust candidates for further validation as potential biomarkers or therapeutic targets.\nKeywords Biomarker, Endometriosis, Transcription factors, Hub gene, Developmental pathways\nElucidating the role of transcription factors \nin molecular pathways underlying infertility \nin endometriosis: a bioinformatics approach\nNiloofar Shahgholi1, Zahra Noormohammadi1*, Ashraf Moini2,3,4 and Morteza Karimipoor5\n\nPage 2 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nIntroduction\nEndometriosis is a frequently chronic inflammatory con -\ndition in women, characterized by the presence of ecto -\npic endometrial-like tissue outside the uterus. These sites \nare primarily located in the pelvic region, including the \novaries, ligaments, peritoneal surfaces, and the bowel \nand bladder [1]. The primary symptoms of endometriosis \ninclude chronic pelvic pain, severely painful menstrual \nperiods (dysmenorrhea), and infertility [ 2]. The theory of \nretrograde menstruation, also known as Sampson’s the -\nory, suggests that menstrual blood, which contains endo -\nmetrial cells, flows backward through the fallopian tubes \ninto the peritoneal cavity. Once there, these cells have the \npotential to implant and proliferate [1].\nEndometriosis is also a complex inflammatory condi -\ntion that impacts women’s reproductive health globally, \nfrom adolescence to menopause, crossing ethnic and \nsocio-economic boundaries, and leading to significant \nhealth challenges [ 3]. Based on familial studies, the inci -\ndence of endometriosis is influenced by a genetic factor, \naccounting for approximately 50% of the susceptibility to \nthe condition [ 4]. The malfunction of the genes involved \nin pathways such as steroidogenesis, sex hormone recep -\ntors, inflammation, immune response, tissue remodeling, \nangiogenesis, metabolism regulation, and DNA repair \ncould be associated with endometriosis [ 5]. However, the \nprecise genetic and pathophysiological basis of endome -\ntriosis remains unclear [ 6]. Moreover, diagnosis of endo -\nmetriosis is often delayed by 4 to 11 years from symptom \nonset until surgical confirmation. While laparoscopy \nis the gold standard for diagnosis, it is an invasive pro -\ncedure, highlighting the need for improved diagnostic \nmethods [ 7]. Using a biomarker or a set of biomarkers \nthat can be easily measured, typically noninvasive, and \nmay assist the clinician in diagnosing and tracking the \ntreatment response is critical [ 8]. Current biomarkers \nrecommended for endometriosis diagnosis include Can -\ncer Antigen 125 (CA-125), Cancer Antigen 199 (CA-199), \nUrocortin (UCN), and Interleukin-6 (IL-6). However, \nnone of these new markers have yet been approved as an \nexclusive diagnostic biomarker for endometriosis [9].\nThe exact mechanisms by which endometriosis causes \ninfertility are not fully understood. In patients with \nendometriosis, implantation failure occurs in the endo -\nmetrium, contributing to infertility. One of the factors \ninvolved is progesterone resistance in the eutopic endo -\nmetrium, which leads to the abnormal activation of \nthe WNT/β-catenin signaling pathway. This activation \nresults in the overexpression of WNT target genes such \nas Homeobox A10 (HOXA10) and Matrix Metallopro -\nteinases 9 and 2 (MMP-9 and MMP-2). These changes \nmay impair endometrial receptivity during the critical \nwindow of implantation [ 10]. Emerging evidence also \nhighlights the potential of targeting angiogenic pathways, \nsuch as VEGFR-2 signaling, to improve oocyte quality in \nwomen with endometriosis [ 11]. Accordingly, the iden -\ntification of major pathways implicated in endometri -\nosis-related infertility could facilitate the development \nof targeted therapies to improve fertility outcomes in \naffected women.\nIn the present study, we analyzed differentially \nexpressed genes derived from independent endometrio -\nsis cohorts with heterogeneous populations and demo -\ngraphic backgrounds. By intersecting DEGs across these \ndatasets, we aimed to identify conserved gene expression \nsignatures and pathway regulation patterns that are con -\nsistently associated with endometriosis despite popula -\ntion diversity. Furthermore, we sought to characterize \nkey network pathways and hub genes as robust candidate \nbiomarkers in the pathogenesis of endometriosis.\nMethods\nData acquisition\nWe downloaded 4 endometriosis-associated datasets \n(GSE51981, GSE7305, GSE23339, and GSE120103) from \nthe Gene Expression Omnibus (GEO) database. Based \non the GPL570 platform [HG-U133_Plus_2], Affymetrix \nHuman Genome U133 Plus 2.0 Array with 54,675 entries, \nGSE51981 contains the gene expression profiles of 148 \nAmerican women with and without endometriosis, and \nGSE7305 involves 20 endometrium samples of Cauca -\nsian women with ovarian endometriosis and with normal \nconditions. Additionally, GSE23339 was on the GPL6102 \nplatform (Illumina human-6 v2.0 expression bead chip) \nwith 48,702 entries, which included endometrial sam -\nples from American women (10 cases of endometrioma \nand 9 cases of non-endometriosis). GSE120103 was on \nthe GPL6480 platform, Agilent-014850 Whole Human \nGenome Microarray 4 × 44 K G4112F (Probe Name ver -\nsion), with 41,108 entries, including 36 samples from \nIndian women with and without endometriosis.\nDifferentially expressed genes identification\nRaw microarray data from four independent endome -\ntriosis cohorts (GSE7305, GSE51981, GSE23339, and \nGSE120103) were retrieved from GEO using the GEO -\nquery package. Datasets were generated on different plat-\nforms (Affymetrix, Illumina, Agilent), requiring careful \nhandling of cross-platform heterogeneity. Probe identi -\nfiers were mapped to official gene symbols, and multiple \nprobes per gene were collapsed using the median. Gene-\nlevel matrices were merged across studies, and missing \nvalues were imputed with the median.\nBatch effects at the dataset level were corrected using \nlimma’s removeBatchEffect function, while platform-\nlevel correction was not applied due to confounding \nbetween platform and dataset. PCA confirmed effec -\ntive batch correction, preserving biological differences \n\nPage 3 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nbetween endometriosis and control samples (Supplemen-\ntary Figure S1).\nFigure S1. PCA of combined gene expression data \nfrom four microarray datasets after batch effect correc -\ntion. Samples separate clearly by disease status (triangles: \nendometriosis; circles: control) along PC1, rather than by \ndataset (GSE120103: pink, GSE23339: green, GSE51981: \nblue, GSE7305: purple), indicating successful removal of \ntechnical variation and highlighting conserved biological \ndifferences.\nDifferential expression analysis was performed for each \ndataset using limma, retaining platform-specific normal -\nization (e.g., RMA for Affymetrix). Genes with |log₂ fold \nchange| >1.5 and adjusted P < 0.05 were considered sig -\nnificant. Robust DEGs were defined as those consistently \ndetected across all four cohorts, prioritizing reproduc -\nible, disease-associated expression changes over dataset-\nspecific effects.\nConstruction of differential gene core modules and \nscreening of hub genes\nThe STRING database (Version 12.0)  (   h t t p s : / / c n . s t r i n g - d \nb . o r g /     ) was used to provide comprehensive information \non experimental and predicted findings about protein-\nprotein interactions between intersecting genes of the \nfour selected microarray datasets. The network obtained \nfrom STRING was imported into Cytoscape software \n(Version 3.10.4) to visualize the interactions between \nproteins. The cytoHubba plugin is utilized to identify hub \ngenes by calculating topological measurements within \nthe protein network.\nGene ontology enrichment analysis\nEnrichment analysis was conducted to elucidate the bio -\nlogical processes of overlapping differentially expressed \ngenes (DEGs) using the online platform Web-based Gene \nSet Analysis Toolkit (WebGestalt,  h t t p s : / / w w w . w e b g e s t a \nl t . o r g /     ) . The list of gene names was uploaded to WebGe -\nstalt, and the Kyoto Encyclopedia of Genes and Genomes \n(KEGG, https://www.genome.jp/kegg/)  d a t a b a s e was \nselected to identify the biological and functional path -\nways associated with these genes, using a cutoff point of \nFDR < 0.25.\nNetwork construction of hub genes\nThe online tool GeneMANIA ( https://genemania.org/) \nwas used to construct the hub genes’ interaction net -\nwork. The network weighting method was employed \nfor this construction, where the weighting method was \nautomatically selected. The maximum number of resul -\ntant genes and resultant attributes was set to 20 and 10, \nrespectively.\nIdentification of key transcription factors in gene \nregulation networks\nThe human transcription factor dataset [ 12], comprising \n1,639 transcription factors (TF), was utilized to identify \nwhich of the selected hub genes function as transcription \nfactors.\nResults\nDifferential gene identification from microarray datasets\nThis study aimed to identify differentially expressed genes \n(DEGs) from four microarray datasets (GSE120103, \nGSE7305, GSE51981, and GSE23339) (Table  1). Based \non the Limma package, we identified 49 DEGs in \nGSE7305, 396 in GSE23339, 205 in GSE51981, and 70 in \nGSE120103. Each dataset’s upregulated and downregu -\nlated genes were visualized using volcano plots (Fig. 1).\nComparison of overlapped genes among the four GEO \ndatasets\nAmong the up-regulated DEGs, three common genes \nwere identified between GSE120103 and GSE23339, two \ncommon genes between GSE120103 and GSE51981, one \ngene between GSE7305 and GSE23339, and one gene \nbetween GSE120103 and GSE7305. Among the down-\nregulated DEGs, 189 common genes were identified \nbetween GSE23339 and GSE51981, six common genes \nbetween GSE23339, GSE51981, and GSE7305, and one \ngene between GSE23339, GSE51981, and GSE120103 \n(Table S1). Venn diagram analysis identified genes con -\nsistently shared across all four datasets, including 7 com -\nmonly up-regulated and 196 commonly down-regulated \ngenes (Fig. 2).\nTable 1 The information related to the analyzed GEO datasets in this study\nDataset Platform Method Tissue Type Menstrual Phase Normal \n(No.)\nEndome-\ntriosis \n(No.)\nTotal \n(No.)\nPMID\nGSE7305 GPL570 Microarray Eutopic endometrium Secretory phase 10 10 20 17,640,886\nGSE120103 GPL6480 Microarray Eutopic endometrium Proliferative and secretory 18 18 36 30,760,267\nGSE23339 GPL6102 Microarray Eutopic endometrium Not specified 9 10 19 21,436,257\nGSE51981 GPL570 Microarray Eutopic endometrium Proliferative, early secretory, \nmid-secretory\n34 114 148 25,243,856\n\nPage 4 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nIntegrated analysis of differentially expressed genes: \nconstruction and evaluation of PPI networks\nA total of 202 differentially expressed genes (DEGs) \nwere identified across the four datasets. Protein-pro -\ntein interaction (PPI) information for these DEGs was \nobtained from the STRING database (version 12.0) with \na minimum required interaction score of 0.4 (medium \nconfidence). The resulting network was imported into \nCytoscape (version 3.10.4) for visualization and analysis, \ncomprising 142 nodes and 307 edges (Figure S2).\nHub genes were identified using the cytoHubba plu -\ngin. The top 20 genes were independently ranked by four \nalgorithms in cytoHubba: MCC, Degree, Edge Percolated \nComponent (EPC), and Density of Maximum Neigh -\nborhood Component (DMNC). Given the high overlap \namong these rankings, a consensus set of 20 hub genes \nFig. 1 Volcano plots depicting the differentially expressed genes (DEGs) identified from GSE7305, GSE120103, GSE23339, and GSE51981. The criteria used \nfor identifying DEGs were |logFC2| > 1.5 and p-value < 0.05. Upregulated genes are highlighted in red, while downregulated genes are marked in blue\n \n\nPage 5 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nwas selected based on consistent high performance \nacross methods (Fig.  3). These hub genes include ESR1, \nEPCAM , MUC1, MMP9, CLDN7 , MET, RAB25 , PRSS8 , \nPGR, DKK1, FOXA2, MSX1, HOXB4, HOXA9, HOXA10, \nPAX8, GATA2, SOX17, RUNX1, and MAL2.\nEnrichment analysis of endometriosis-related hub genes\nPathway and Gene Ontology (GO) enrichment analyses \nwere performed using WebGestalt to explore the func -\ntional roles of the 20 hub genes, with pathways sourced \nfrom KEGG and Reactome databases. GO analysis was \nconducted across three categories: biological process \n(BP), molecular function (MF), and cellular component \n(CC).\nKEGG pathway analysis indicated enrichment in \ncancer-related pathways, including pathways in cancer, \nproteoglycans in cancer, and tight junction (Fig.  4-A). \nIn contrast, Reactome analysis highlighted significant \ninvolvement in developmental biology pathways, such as \nsignaling by nuclear receptors, RUNX1-regulated tran -\nscription, WNT signaling, gastrulation, formation of \ndefinitive endoderm, estrogen-dependent gene expres -\nsion, and ESR-mediated signaling (Fig. 4-A).\nGO biological process analysis revealed that hub genes \nare predominantly associated with tissue development, \nregionalization, pattern specification processes, epi -\nthelium development, embryonic morphogenesis, and \npositive regulation of transcription/DNA-templated \ntranscription/RNA biosynthetic processes by RNA poly -\nmerase II (Fig.  4-B). Molecular function analysis showed \nenrichment in transcription regulatory region nucleic \nacid/DNA binding, sequence-specific double-stranded \nDNA binding, and DNA-binding transcription acti -\nvator activity (Fig.  4-B). For cellular component, hub \ngenes were mainly localized to transcription regulator \ncomplexes, protein-DNA complexes, chromatin/chro -\nmosomes, tight junctions, and plasma membrane com -\nponents (apical/basal/lateral) (Fig. 4-B).\nThese results suggest that the hub genes contribute \nto endometriosis pathogenesis through transcriptional \nregulation, disruption of epithelial barrier function (e.g., \ntight junctions), and dysregulation of developmental \nprocesses, consistent with the hormonal and structural \nabnormalities observed in the disease.\nThe GeneMANIA database was used to construct a \nfunctional association network for the 20 hub genes, \nillustrating their potential interactions and shared bio -\nlogical roles (Fig.  5). The network incorporates multiple \ntypes of evidence, including co-expression, physical \ninteractions, predicted interactions, genetic interactions, \nco-localization, shared protein domains, and pathway \nassociations, with line colors indicating the type of evi -\ndence (as shown in the legend). Node size and pie chart \ncomposition reflect the relative contribution of each evi -\ndence type to the gene’s connectivity, highlighting genes \nwith stronger functional associations.\nGenes with higher connectivity and involvement in \nmultiple shared functions include FOXA2, SOX17, PAX8, \nHOXA9, HOXA10 , HOXB4 , HOXB7 , HOXB8 , GATA2 , \nRUNX1, and DKK1 (Fig.  5). These genes participate in \nenriched processes such as pattern specification, region -\nalization, embryonic morphogenesis, embryonic organ \ndevelopment, and positive regulation of transcription by \nRNA polymerase II, consistent with their roles in endo -\nmetrial development and transcriptional regulation.\nFig. 2 A Venn diagram illustrating the overlap of DEGs across the datasets. A Upregulated overlapped DEGs. B Downregulated overlapped DEGs\n \n\nPage 6 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nIdentification of hub genes functioning as transcription \nfactors\nAmong the 20 identified hub genes, 12 were found to act \nas transcription factors. A Venn diagram illustrating the \noverlap between the 1,639 human transcription factors \nand the identified hub genes is presented in Fig. 6.\nDiscussion\nEndometriosis is a chronic inflammatory disease that \npredominantly affects women of reproductive age. It is \ncharacterized by symptoms such as irregular menstrua -\ntion, menorrhagia, and infertility [13]. Although the exact \nmechanism of the disease remains unclear, substantial \nevidence supports its multifactorial nature, influenced \nby anatomical, hormonal, immunological, estrogenic, \ngenetic, epigenetic, and environmental factors [ 14]. \nNumerous genes have been identified as playing roles in \nthe pathogenesis of endometriosis, many of which have \nbeen discovered through experimental studies. How -\never, microarray technologies and RNA sequencing have \nemerged as powerful tools for identifying potential bio -\nmarkers in endometriosis research. This allows for a com-\nprehensive analysis of expression profiles by identifying \ndifferentially expressed genes (DEGs) between endome -\ntriosis patients and healthy controls [15].\nIn this study, an integrative analysis of four indepen -\ndent microarray datasets identified 20 hub genes that \nwere consistently altered across cohorts, providing a \nrobust molecular signature of endometriosis. Among \nthese, 12 hub genes function as transcription fac -\ntors (TFs )—including FOXA2, SOX17, PAX8, HOXA9, \nHOXA10, HOXB4, HOXB7, RUNX1, GATA2, ESR1, PGR, \nand MSX1— highlighting a central role of transcriptional \nregulation in endometrial homeostasis and pathology \n[16– 18].\nFunctional enrichment analyses of these hub genes \nrevealed coordinated roles in transcriptional regulation, \ntissue development, and epithelial structure maintenance. \nGene Ontology (GO) molecular function terms indicated \nthat these TFs primarily act as DNA-binding activators, \nregulating downstream genes essential for uterine gland \nformation and endometrial receptivity. Transcription fac-\ntors can lead to abnormal biological outcomes in endo -\nmetriosis, such as increased estrogen levels, immune \nsystem inflammation, and enhanced angiogenesis [ 19]. \nFor instance, FOXA2 acts as a pioneer TF modulating \nFig. 3 Protein-protein interaction (PPI) network of 20 consensus hub genes in endometriosis. Node size and color reflect degree centrality (larger/darker \nnodes indicate higher connectivity); edges represent STRING interaction confidence (thicker/darker edges indicate higher score)\n \n\nPage 7 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nchromatin accessibility and glandular development, while \nSOX17 ensures epithelial identity and proper glandular \nmorphogenesis [ 20]. PAX8 regulates epithelial differen -\ntiation and is essential for the homeostatic regeneration \nand maintenance of both luminal and glandular endo -\nmetrial epithelium [21], and members of the HOX family \n(such as HOXA10 and HOXA11) are pivotal in uterine \ntissue patterning and endometrial development [22].\nEndometriosis is a complex condition that involves \nhormonal, neurological, and immunological factors [ 23]. \nThe imbalance of ovarian steroid hormones, specifically \nProgesterone (P4) and Estrogen (E2), along with the dys -\nregulation of their downstream signaling targets, plays \nFig. 4 Dot plots of enrichment and Gene Ontology (GO) analysis of the top 60 hub genes. A Enriched pathway enrichment analysis conducted through \nKEGG and Reactome. B Gene Ontology (GO) analysis for the top 20 hub genes\n \n\nPage 8 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \na crucial role in the development and persistence of the \ndisorder [ 24]. Hormonal regulation through ESR1 and \nPGR, also identified among hub genes, reinforces the \ncrucial interplay between steroid signaling and transcrip -\ntional control. Dysregulation of these nuclear receptors \nlikely disrupts downstream transcriptional networks, \ncontributing to abnormal proliferation, inflammation, \nand compromised endometrial receptivity—hallmarks of \nendometriosis [25, 26].\nProtein-protein interaction and GeneMANIA analy -\nses further revealed extensive co-expression and physi -\ncal interactions among hub genes, forming a cohesive \nfunctional module enriched for developmental processes \nsuch as pattern specification, regionalization, embryonic \nmorphogenesis, and organ development. These obser -\nvations suggest that transcriptional dysregulation and \naltered developmental signaling may affect adenogenesis \nand epithelial barrier integrity, processes that have been \npreviously implicated in ectopic lesion establishment \nand infertility in endometriosis [ 27– 29]. GO cellular \ncomponent terms related to tight junctions and plasma \nmembrane structures support the relevance of these hub \ngenes in maintaining epithelial integrity, where disrup -\ntion may promote lesion establishment and invasion [30]. \nImportantly, our results underscore the interconnected \nnature of transcriptional, developmental, and hormonal \npathways in endometriosis. Hub genes such as HOXA10, \nHOXA9, HOXB4, FOXA2, SOX17, PAX8, RUNX1, and \nGATA2 appear to act synergistically, orchestrating gene \nnetworks critical for endometrial morphogenesis, gland \nformation, and hormone responsiveness. The consistent \ndifferential expression of these genes across multiple \ndatasets strengthens their candidacy as potential bio -\nmarkers or therapeutic targets.\nWhile literature supports the functional roles of \nFOXA2, SOX17, PAX8, and HOX family members in \nFig. 5 Gene interaction network of the top 20 hub genes visualized using GeneMANIA\n \n\nPage 9 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nuterine development and fertility [ 17, 18]. Our integra -\ntive transcriptomic analysis across multiple independent \ncohorts identified a set of shared differentially expressed \nhub genes relevant to endometriosis. These shared genes \nwere selected based on consistent up- or downregulation \nacross datasets, irrespective of the ancestral background \nor population heterogeneity of individual study partici -\npants. Consequently, this approach emphasizes genes \nwith robust involvement in endometriosis pathogenesis \nrather than population-specific effects. Deeper stratified \nanalyses (e.g., by disease stage, phenotype, or infertility \nstatus) were not feasible due to the lack of harmonized \nclinical metadata across the included GEO datasets. \nNonetheless, limitations include the absence of direct \nfunctional validation, a lack of detailed patient fertility \nstatus data, and the need for proteomic corroboration.\nConclusion\nIn this study, an integrative analysis of four indepen -\ndent transcriptomic datasets identified a robust set of \n20 hub genes that were consistently altered in endome -\ntriosis. Among these, 12 transcription factors—includ -\ning FOXA2, SOX17, PAX8, and members of the HOX \nfamily—play central roles in transcriptional regulation, \nuterine development, and endometrial homeostasis. \nFunctional enrichment and network analyses revealed \ncoordinated roles in developmental processes, gland \nformation, and epithelial integrity, highlighting mecha -\nnisms potentially contributing to infertility and lesion \nestablishment in endometriosis. Importantly, the shared \ndifferential expression of these hub genes across cohorts \nwas observed independently of participant ancestry, \nemphasizing disease-specific molecular signatures rather \nthan population-specific effects. These findings provide a \nfoundation for future functional studies and may inform \nthe development of biomarkers or targeted therapies for \nthe treatment of endometriosis.\nSupplementary Information\nThe online version contains supplementary material available at  h t t p  s : /  / d o i  . o  r \ng /  1 0 .  1 1 8 6  / s  4 3 0 4 3 - 0 2 6 - 0 0 3 1 0 - 8.\nSupplementary Material 1. \nSupplementary Material 2. Table S1. Overlapped DEGs among 4 datasets, \nGSE7305, GSE120103, GSE23339, and GSE51981.\nSupplementary Material 3. Figure S1. PCA of combined gene expression \ndata from four microarray datasets after batch effect correction. Samples \nseparate clearly by disease status (triangles: endometriosis; circles: control) \nalong PC1, rather than by dataset (GSE120103: pink, GSE23339: green, \nGSE51981: blue, GSE7305: purple), indicating successful removal of techni-\ncal variation and highlighting conserved biological differences. Figure S2. \nProtein-protein interaction (PPI) network of the overlapped DEGs among \nthe datasets.\nAuthors’ contributions\nN.S. and Z.N. contributed to the Conceptualization, Methodology, and \nInvestigation. Z.N., A.M., and M.K. provided Supervision and Project \nAdministration. All authors were involved in Writing, reviewing, and Editing, \nand approved the final version of the manuscript for submission.\nFunding\nNo grants were received to support the conduct of this study.\nFig. 6 Venn Diagram showing the overlap between the list of human transcription factors (TFs) and the high-scoring genes involved in the gene interac-\ntion network\n \n\nPage 10 of 10\nShahgholi et al. Middle East Fertility Society Journal            (2026) 31:29 \nData availability\nNo datasets were generated during the current study.\nDeclarations\nEthics approval and consent to participate\nNot applicable.\nConsent for publication\nNot applicable.\nCompeting interests\nThe authors declare no competing interests.\nReceived: 24 September 2025 / Accepted: 23 February 2026\nReferences\n1. Zondervan K, Becker C, Koga K, Viganò P , Endometriosis (2018) Nat Rev Dis \nPrimers 4(1):9.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 3 8  / s  4 1 5 7 2 - 0 1 8 - 0 0 0 8 - 5\n2. Smolarz B, Szyłło K, Romanowicz H (2021) Endometriosis: epidemiology, \nclassification, pathogenesis, treatment and genetics (review of literature). Int \nJ Mol Sci 22(19):10554.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / i  j m s 2 2 1 9 1 0 5 5 4\n3. Roy C, Mondal N (2023) Global risks of endometriosis in women – an \nappraisal. Eur J Clin Exp Med 21(2):405–415.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 5 5 8  4 /  e j c e m . 2 \n0 2 3 . 2 . 2 7\n4. Montgomery GW, Mortlock S, Giudice LC (2020) Should genetics now be \nconsidered the pre-eminent etiologic factor in endometriosis? J Minim \nInvasive Gynecol 27(2):280–286.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 1 6  / j  . j m i g . 2 0 1 9 . 1 1 . 0 1 1\n5. Deiana D, Gessa S, Anardu M, Daniilidis A, Nappi L, D’Alterio MN et al (2019) \nGenetics of endometriosis: a comprehensive review. Gynecol Endocrinol \n35(7):561–570.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 8 0  / 0  9 5 1  3 5 9  0 . 2 0  1 9  . 1 5 8 8 8 7 2\n6. Goulielmos GN, Matalliotakis M, Matalliotaki C, Eliopoulos E, Matalliotakis I, \nZervou MI (2020) Endometriosis research in the -omics era. Gene 741:144545.  \nh t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 1 6  / j  . g e n e . 2 0 2 0 . 1 4 4 5 4 5\n7. Agarwal SK, Chapron C, Giudice LC, Laufer MR, Leyland N, Missmer SA et \nal (2019) Clinical diagnosis of endometriosis: a call to action. Am J Obstet \nGynecol 220(4):354.e1-354.e12.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 1 6  / j  . a j o g . 2 0 1 8 . 1 2 . 0 3 9\n8. Anastasiu CV, Moga MA, Neculau AE, Bălan A, Scârneciu I, Dragomir RM et al \n(2020) Biomarkers for the noninvasive diagnosis of endometriosis: state of the \nart and future perspectives. Int J Mol Sci 21(5):1750.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / i  j \nm s 2 1 0 5 1 7 5 0\n9. Tian Z, Chang XH, Zhao Y, Zhu HL (2020) Current biomarkers for the detection \nof endometriosis. Chin Med J (Engl) 133(19):2346–2352.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 \n9 7  / C  M 9 .  0 0 0  0 0 0 0  0 0  0 0 0 1 0 6 2\n10. Pazhohan A, Amidi F, Akbari-Asbagh F, Seyedrezazadeh E, Farzadi L, Kho-\ndarahmin M et al (2018) The Wnt/β-catenin signaling in endometriosis, the \nexpression of total and active forms of β-catenin, total and inactive forms of \nglycogen synthase kinase-3β, WNT7a and DICKKOPF-1. Eur J Obstet Gynecol \nReprod Biol 220:1–5.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 1 6  / j  . e j  o g r  b . 2 0  1 7  . 1 1 . 0 1 0\n11. Shahgholi N, Noormohammadi Z, Moini A, Karimipoor M (2025) Cabergoline’s \npromise in endometriosis: restoring molecular balance to improve reproduc-\ntive potential. Gynecol Obstet Invest 1–16. Advance online publication.  h t t p  s : \n/  / d o i  . o  r g /  1 0 .  1 1 5 9  / 0  0 0 5 4 6 1 9 8\n12. Lambert SA, Jolma A, Campitelli LF, Das PK, Yin Y, Albu M et al (2018) The \nhuman transcription factors. Cell 172(4):650–665.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 1 6  / j  . c e l \nl . 2 0 1 8 . 0 1 . 0 2 9\n13. Taylor HS, Kotlyar AM, Flores VA (2021) Endometriosis is a chronic systemic \ndisease: clinical challenges and novel innovations. Lancet 397(10276):839–\n852.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 1 6  / S  0 1 4 0 - 6 7 3 6 ( 2 1 ) 0 0 3 8 9 - 9\n14. Laganà AS, Garzon S, Götte M, Viganò P , Franchi M, Ghezzi F et al (2019) The \npathogenesis of endometriosis: molecular and cell biology insights. Int J Mol \nSci 20(22):5615.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / i  j m s 2 0 2 2 5 6 1 5\n15. Cho SB (2023) Molecular mechanisms of endometriosis revealed using \nomics data. Biomedicines 11(8):2210.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / b  i o m  e d i  c i n \ne  s 1  1 0 8 2 2 1 0\n16. Kelleher AM, Behura SK, Burns GW, Young SL, DeMayo FJ, Spencer TE (2019) \nIntegrative analysis of the forkhead box A2 (FOXA2) cistrome for the human \nendometrium. FASEB J 33(7):8543–8554.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 9 6  / f   . 2 0 1 8 0 2 9 0 \n0 R R\n17. Wang X, Li X, Wang T, Wu SP , Jeong JW, Kim TH et al (2018) SOX17 regulates \nuterine epithelial-stromal cross-talk acting via a distal enhancer upstream of \nIhh. Nat Commun 9(1):4421.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 3 8  / s  4 1 4 6 7 - 0 1 8 - 0 6 6 5 2 - 4\n18. Lazim N, Elias MH, Sutaji Z, Abdul Karim AK, Abu MA, Ugusman A et al (2023) \nExpression of HOXA10 gene in women with endometriosis: a systematic \nreview. Int J Mol Sci 24(16):12869.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / i  j m s 2 4 1 6 1 2 8 6 9\n19. Lamceva J, Uljanovs R, Strumfa I (2023) The main theories on the pathogen-\nesis of endometriosis. Int J Mol Sci 24(5):4254.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / i  j m s 2 4 \n0 5 4 2 5 4\n20. Kinnear S, Salamonsen LA, Francois M, Harley V, Evans J (2019) Uterine SOX17: \na key player in human endometrial receptivity and embryo implantation. Sci \nRep 9(1):15495.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 3 8  / s  4 1 5 9 8 - 0 1 9 - 5 1 8 5 0 - 9\n21. Fu DJ, De Micheli AJ, Bidarimath M, Ellenson LH, Cosgrove BD, Flesken-Nikitin \nA et al (2020) Cells expressing PAX8 are the main source of homeostatic \nregeneration of adult mouse endometrial epithelium and give rise to serous \nendometrial carcinoma. Dis Model Mech 13(10):dmm047035.  h t t p  s : /  / d o i  . o  r g /  \n1 0 .  1 2 4 2  / d  m m . 0 4 7 0 3 5\n22. Du H, Taylor HS (2015) The role of hox genes in female reproductive tract \ndevelopment, adult function, and fertility. Cold Spring Harb Perspect Med \n6(1):a023002.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 1 0 1  / c  s h p  e r s  p e c t  . a  0 2 3 0 0 2\n23. Cano-Herrera G, Salmun Nehmad S, Ruiz de Chávez Gascón J, Méndez \nVionet A, van Tienhoven XA, Osorio Martínez MF et al (2024) Endometriosis: \na comprehensive analysis of the pathophysiology, treatment, and nutritional \naspects, and its repercussions on the quality of life of patients. Biomedicines \n12(7):1476.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / b  i o m  e d i  c i n e  s 1  2 0 7 1 4 7 6\n24. Marquardt RM, Kim TH, Shin JH, Jeong JW (2019) Progesterone and estrogen \nsignaling in the endometrium: what goes wrong in endometriosis? Int J Mol \nSci 20(15):3822.  h t t p  s : /  / d o i  . o  r g /  1 0 .  3 3 9 0  / i  j m s 2 0 1 5 3 8 2 2\n25. Yilmaz BD, Bulun SE (2019) Endometriosis and nuclear receptors. Hum \nReprod Update 25(4):473–85.  h t t p  s : /  / d o i  . o  r g /  1 0 .  1 0 9 3  / h  u m u p d / d m z 0 0 5\n26. Zhang P , Wang G (2023) Progesterone resistance in endometriosis: current \nevidence and putative mechanisms. Int J Mol Sci 24(8):6992.  h t t p  s : /  / d o i  . o  r g /  1 \n0 .  3 3 9 0  / i  j m s 2 4 0 8 6 9 9 2\n27. Kelleher AM, DeMayo FJ, Spencer TE (2019) Uterine glands: developmental \nbiology and functional roles in pregnancy. Endocr Rev 40(5):1424–1445.  h t t p  s \n: /  / d o i  . o  r g /  1 0 .  1 2 1 0  / e  r . 2 0 1 8 - 0 0 2 4 3\n28. Grund S, Grümmer R (2018) Direct cell–cell interactions in the endometrium \nand in endometrial pathophysiology. Int J Mol Sci 19(8):2227.  h t t p  s : /  / d o i  . o  r g /  \n1 0 .  3 3 9 0  / i  j m s 1 9 0 8 2 2 2 7\n29. Chen CW, Chavez JB, Kumar R, Go VA, Pant A, Jain A et al (2024) Hypersensi-\ntive intercellular responses of endometrial stromal cells drive invasion in \nendometriosis. eLife 13:e94778.  h t t p  s : /  / d o i  . o  r g /  1 0 .  7 5 5 4  / e  L i f e . 9 4 7 7 8\n30. Classen-Linke I, Buck VU, Sternberg AK, Kohlen M, Izmaylova L, Leube RE \n(2025) Changes in epithelial cell polarity and adhesion guide human endo-\nmetrial receptivity: how in vitro systems help to untangle mechanistic details. \nBiomolecules 15(8):1057. Advance online publication.  h t t p s :   /  / d o  i .  o r  g  /  1 0  . 3 3   9 0  \n/ b i o m 1 5 0 8 1 0 5 7\nPublisher’s Note\nSpringer Nature remains neutral with regard to jurisdictional claims in \npublished maps and institutional affiliations.","source_license":"CC0","license_restricted":false}