Causal Genes and Immune-Epigenetic Mechanisms Underlying Polycystic Ovary Syndrome: A Multi-Omics Mendelian Randomization Study

Causal Genes and Immune-Epigenetic Mechanisms Underlying Polycystic Ovary Syndrome: A Multi-Omics Mendelian Randomization Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Causal Genes and Immune-Epigenetic Mechanisms Underlying Polycystic Ovary Syndrome: A Multi-Omics Mendelian Randomization Study Junxiu Liu, Chengzi Huang, Jun Jiao, Yue Sun, Yingxiu Ma, Yang Yang, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7532908/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Polycystic ovary syndrome (PCOS) is a common endocrine-metabolic disorder, yet its genetic basis remains incompletely understood. This study aimed to identify causal genes and elucidate upstream epigenetic and immune cell–specific regulatory mechanisms using a multi-omics Mendelian randomization (MR) framework. Methods We performed two-sample MR analyses using expression quantitative trait loci (eQTLs) from the eQTLGen consortium and protein QTLs (pQTLs) from the UK Biobank to assess the causal effects of gene expression and protein levels on PCOS risk. Genes significant in both datasets were retained as candidate genes and further evaluated using summary-data-based Mendelian randomization (SMR) with GTEx whole-blood eQTLs to determine colocalized genetic signals. To investigate upstream regulation, we conducted mediation MR analysis using methylation QTLs (mQTLs) from the GoDMC database to identify CpG sites potentially mediating gene expression and PCOS risk. Finally, we performed cell-type–specific MR using single-cell eQTLs (sc-eQTLs) from the OneK1K project across 14 immune cell types. Results MR identified 1,715 eQTL- and 182 pQTL-associated genes, with 60 overlapping candidates. SMR prioritized six causal genes: CRELD1, NSFL1C, ITIH4, IL6R, SNAP29, and PON2. Mediation MR revealed a borderline-significant effect for cg20688791 upstream of IL6R and suggestive mediation at cg00335892 within SNAP29 . sc-eQTL analysis showed that CRELD1 , ITIH4 , PON2 , and SNAP29 had significant causal effects in CD8⁺ T cells, CD4⁺ T cells, monocytes, and NK cells, respectively. Conclusion This integrative analysis identifies multi-omics-supported causal genes for PCOS and reveals epigenetic and immune cell–specific regulatory mechanisms, offering novel insights into pathogenesis and potential therapeutic targets. Polycystic ovary syndrome Mendelian randomization Multi-omics integration DNA methylation Single-cell eQTLs Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Key points Multi-omics Mendelian randomization identified six causal genes for PCOS. SMR analysis confirmed colocalization between genetic variants, gene expression, and PCOS risk. DNA methylation mediation analysis implicated epigenetic regulation of IL6R and SNAP29. Single-cell eQTL analysis revealed immune cell type–specific effects of causal genes. Findings provide novel mechanistic insights and potential therapeutic targets for PCOS. Introduction Polycystic ovary syndrome (PCOS) is a common endocrine disorder affecting approximately 5% to 10% of women of reproductive age worldwide[1]. It is characterized by a heterogeneous clinical presentation, including menstrual irregularities, hyperandrogenism, and polycystic ovarian morphology[2]. Beyond its impact on fertility, PCOS is strongly associated with metabolic disturbances such as insulin resistance, type 2 diabetes, and cardiovascular disease, posing significant threats to both physical and mental health[3]. Although the precise etiology of PCOS remains elusive, accumulating evidence supports a substantial genetic predisposition. The heritability of PCOS has been estimated to be as high as 70%, indicating that genetic factors are instrumental in its development. For instance, twin studies have demonstrated a strong familial correlation for phenotypic features of PCOS, suggesting a polygenic inheritance pattern rather than a single genetic defect[4]. Genetic variants associated with PCOS include polymorphisms in genes involved in insulin signaling and hormone regulation, such as insulin gene polymorphisms and variations in the FSH receptor[5]. With the advancement of genomics and epigenomics technologies, increasing efforts have been directed toward identifying PCOS-related pathogenic genes and their regulatory mechanisms using multi-omics approaches[6, 7], with the goal of providing theoretical foundations and novel targets for clinical intervention. Mendelian randomization (MR) is an analytical framework that leverages naturally occurring genetic variants—typically single nucleotide polymorphisms (SNPs)—as instrumental variables to infer causality between exposures and disease outcomes[8, 9]. By simulating the conditions of a randomized controlled trial, MR can effectively mitigate confounding and reverse causation, which are common limitations in traditional observational studies. This has made MR increasingly valuable in elucidating causal relationships in complex diseases[10]. DNA methylation is a key epigenetic modification that plays a crucial role in regulating gene expression and mediating phenotypic variation. It refers to the addition of a methyl group to the cytosine residue of DNA, predominantly at CpG dinucleotides, leading to the formation of 5-methylcytosine[11]. This modification can modulate gene expression by altering chromatin structure, affecting DNA accessibility, and influencing the binding of transcription factors. Although DNA methylation patterns are often stable, accumulating evidence suggests that they can also exhibit plasticity in response to environmental cues such as nutritional status, psychological stress, or toxic exposures[12, 13]. These dynamic changes in DNA methylation are considered to underlie, at least in part, the molecular mechanisms through which organisms adapt their gene expression programs and phenotypic traits to external environmental conditions[14, 15]. Aberrant methylation patterns have been implicated in a wide range of metabolic and reproductive disorders[16, 17], including PCOS[18, 19]. Mediation analysis of DNA methylation, gene expression, and disease risk can help uncover potential regulatory pathways and identify functionally relevant loci within gene regulatory networks[20-22]. The immune system has also been increasingly recognized as a key contributor to PCOS pathogenesis. Emerging evidence suggests that PCOS is characterized by a state of chronic low-grade inflammation, marked by elevated levels of pro-inflammatory cytokines and immune cell dysregulation[23, 24]. Investigating the expression of causal genes in specific immune cell types at the single-cell level may provide insight into their functional roles within the immune microenvironment and help explain phenotypic heterogeneity in PCOS[25, 26]. However, to date, no study has systematically integrated multi-omics MR, summary-data-based MR (SMR), methylation quantitative trait loci(mQTLs), and single-cell expression quantitative trait loci (sc-eQTLs) approaches to unravel the mechanisms underlying PCOS. In this study, we systematically integrated expression quantitative trait loci (eQTLs), protein QTL (pQTLs) and genome-wide association study (GWAS) to identify key PCOS-associated genes through two-sample MR and SMR analyses. We further explored upstream epigenetic regulation using DNA methylation–based mediation MR, and evaluated the cell-type–specific expression of candidate genes across 14 immune cell subtypes using sc-eQTLs data. By leveraging genetic, epigenetic, and immune-cell–specific regulatory evidence, our study aims to dissect the complex etiology of PCOS and advance precision medicine in reproductive endocrinology. Materials and Methods 1. Data Sources This study integrated multiple authoritative multi-omics datasets to ensure comprehensive and robust analyses. eQTLs data were obtained from the eQTLGen Consortium , which provides genome-wide cis-eQTL summary statistics based on peripheral blood samples from 31,684 individuals of European ancestry (https://www.eqtlgen.org/cis-eqtls.html). pQTLs data were derived from the UK Biobank , in which plasma proteomic profiles of 54,219 participants were measured using the SomaScan platform and linked to genetic variation through association analysis[27]. Summary-level genome-wide association study (GWAS) data for PCOS were obtained from the FinnGen project (release R11) , comprising 1,909 cases and 241,998 controls (https://r11.finngen.fi/). To explore upstream epigenetic regulatory mechanisms, we incorporated mQTLs data from the Genetics of DNA Methylation Consortium (GoDMC; http://mqtldb.godmc.org.uk/downloads), which catalog extensive associations between CpG methylation levels and genetic variants. To further investigate the cell type–specific regulation of candidate genes, we utilized sc-eQTLs summary data released by the OneK1K project (https://onek1k.org/), covering 14 major immune cell subtypes[28]. This dataset represents a cutting-edge resource for dissecting cell-specific genetic effects in human immune traits. An overview of the analytical workflow is illustrated in Figure 1. Instrumental variables used in each MR framework are available in Supplementary Data 1–4. 2. Two-Sample Mendelian Randomization (MR) Analysis In this study, we performed two-sample Mendelian randomization analyses to evaluate the potential causal effects of gene expression and protein levels in peripheral blood on the risk of PCOS. The analytical framework is illustrated in Figure 2. We selected cis-QTL variants significantly associated with exposures (P 10). To ensure independence among instruments, we applied linkage disequilibrium (LD) clumping using PLINK 2.0, setting an LD threshold of r² < 0.1 and a window size of 10,000 kb, retaining only independent lead SNPs. Exposure and outcome datasets were harmonized to align effect alleles before MR estimation[29, 30]. The primary MR analysis was conducted using the inverse variance weighted (IVW) method to estimate causal effects. Robustness and sensitivity were further evaluated using complementary methods, including MR-Egger regression, the weighted median estimator, Cochran’s Q test for heterogeneity, MR-PRESSO for horizontal pleiotropy detection, and leave-one-out analysis. All analyses were performed in R (version 4.3.2), primarily using the TwoSampleMR and MRPRESSO packages, with visualization via ggplot2 and forestplot. To ensure the reliability of causal inference, we applied the following filtering criteria: (1) a nominally significant IVW P-value ( 0.05) to exclude horizontal pleiotropy. Genes meeting both criteria were considered to have stable causal signals and were retained for downstream analysis. Finally, we intersected the results from eQTL-based and pQTL-based MR to derive a set of 60 candidate genes for subsequent validation and functional investigation. 3. Summary-data–based Mendelian Randomization (SMR) Analysis To further evaluate whether the expression or protein levels of the 60 candidate genes mediate PCOS risk through colocalized genetic variants, we applied SMR. This method integrates eQTLs (or pQTLs) and GWAS summary statistics to test whether the same SNP is associated with both gene expression and disease phenotype, thereby inferring whether the gene may functionally mediate disease risk[31]. To ensure the independence of exposure data from the prior MR analyses, we did not reuse the eQTLGen dataset. Instead, we utilized eQTLs data from whole blood provided by the GTEx v8 project (https://gtexportal.org/home/) as the source of gene expression instruments for SMR. The SMR analysis consisted of two primary statistical tests: (1) the SMR test (p_SMR), which assesses the association between gene expression and disease phenotype; and (2) the HEIDI (Heterogeneity in Dependent Instruments) test (p_HEIDI), which evaluates whether the observed association is likely driven by a single causal variant rather than by multiple colocalized but distinct signals. A non-significant p_HEIDI (p > 0.05) supports the existence of a shared causal variant. Genes were retained if they met the following criteria: p_SMR 0.05, and an odds ratio (OR) > 1, indicating that increased gene expression is associated with higher PCOS risk. All analyses were conducted using the SMR software (https://yanglab.westlake.edu.cn/software/smr/#DataResource). Based on these criteria, six genes were prioritized as likely functional mediators of PCOS risk: CRELD1 , NSFL1C , ITIH4 , IL6R , SNAP29 , and PON2 . 4. DNA Methylation Mediation MR Analysis To systematically assess whether DNA methylation mediates the relationship between gene expression and PCOS risk, we conducted mediation MR analyses based on mQTLs data, focusing on the six key genes identified by SMR analysis ( CRELD1 , NSFL1C , ITIH4 , IL6R , SNAP29 , and PON2 ). The overall analytical workflow was as follows: (1) CpG sites related to each candidate gene were identified using the EWAS DataHub provided by the National Genomics Data Center (https://ngdc.cncb.ac.cn/ewas/datahub/exploration); (2) SNPs significantly associated with these CpG sites were extracted from publicly available summary-level mQTL datasets. For each SNP, we recorded the effect allele, effect size (β), standard error (SE), and P-value, and grouped the data by gene to facilitate downstream analysis; (3) To ensure instrument strength and independence, we applied linkage disequilibrium (LD) clumping for each CpG site to retain only genome-wide significant and independent SNPs using the following parameters: P < 5 × 10⁻⁸, clump_kb = 10,000, clump_r² = 0.1, and clump_p = 1. These mQTL SNPs were then used as instrumental variables in two-sample MR models: ① to estimate the causal effect of CpG methylation on PCOS risk; ② to estimate the causal effect of CpG methylation on gene expression. The indirect (mediation) effect of DNA methylation was then calculated by combining the two causal estimates for each CpG site, representing the extent to which methylation influences PCOS risk via gene expression. Confidence intervals and statistical significance for the mediation effect were estimated using the Delta method. In addition, the mediation proportion was computed to evaluate the contribution of the indirect effect relative to the total effect of DNA methylation on PCOS. This approach enabled the identification of specific CpG sites that may function as upstream regulators of key PCOS-related genes through epigenetic mechanisms. 5. Single-cell eQTLs Analysis in Immune Cell Types To investigate the genetic regulatory characteristics of key candidate genes in distinct immune cell types and to evaluate their potential causal relationships with PCOS, we performed cell type–specific MR analyses based on publicly available sc-eQTL summary data. Specifically, we retrieved sc-eQTL summary-level statistics for 14 major immune cell subtypes from the OneK1K project. For each of the six PCOS-associated genes identified in the SMR analysis ( CRELD1 , NSFL1C , ITIH4 , IL6R , SNAP29 , PON2 ), we extracted cell-specific eQTL data across all immune subtypes, generating a series of “gene–cell type” datasets for downstream analysis. To ensure instrument validity and independence, we applied linkage disequilibrium (LD) clumping to each gene–cell pair using the following parameters: significance threshold: P < 0.05；LD window: 100 kb；LD r² threshold: r² < 0.3 (i.e., clump_kb = 100, clump_r² = 0.3, clump_p = 1). This step was performed to select independent SNPs that were significantly associated with gene expression in each immune cell context, enabling robust causal inference. The filtered sc-eQTL datasets for each gene–cell combination were treated as exposure variables, while the PCOS GWAS summary statistics served as the outcome. Two-sample MR analyses were conducted to assess whether the expression of candidate genes in specific immune cell types causally influenced PCOS risk. Causal estimates were derived using the inverse-variance weighted (IVW) method. MR-Egger regression, MR-PRESSO, and other sensitivity tests were applied where appropriate to evaluate robustness and detect potential horizontal pleiotropy. Results 1. Initial MR Screening Identifies PCOS-Associated Genes A total of 1,715 eQTL-associated genes and 182 pQTL-associated genes were identified as significantly associated with PCOS (Supplementary material: Table S1-S2). By intersecting the results from both analyses, we identified 60 genes whose expression and protein levels were both significantly associated with PCOS risk (Figure 3, Table S3). This intersection strategy enhances the reliability of causal inference by prioritizing genes supported by both transcriptomic and proteomic evidence, and reduces false positives potentially introduced by platform-specific effects or model assumptions. These 60 candidate genes were subsequently selected for integrative multi-omics analyses and functional validation. 2. SMR Analysis Identifies Six Causal Genes with Colocalized Genetic Signals Building on the two-sample MR results, we conducted SMR analyses on the 60 candidate genes to assess whether their expression or protein levels mediate PCOS risk via colocalized genetic variants (Table S4). Six genes— CRELD1 , NSFL1C , ITIH4 , IL6R , SNAP29 , and PON2 —met the predefined criteria for SMR significance (p_SMR 0.05), and positive association (OR > 1), suggesting that they may exert functional causal effects through genetically regulated expression (Table1). Among them, CRELD1 showed the most significant association in the SMR model (OR = 2.13, 95% CI: 1.34–3.38, p_SMR = 1.47 × 10⁻³), indicating that increased expression of CRELD1 may substantially elevate PCOS risk. Both NSFL1C and IL6R also displayed consistent positive associations (OR = 1.81 and 1.90, respectively). As illustrated in Figure 4, we visualized the effect sizes and directions of these six genes across eQTL-, pQTL-, and SMR-based MR analyses using an odds ratio (OR) forest plot. All six genes demonstrated concordant risk-enhancing effects across multiple omics layers, further supporting their potential roles in the genetic architecture of PCOS. Collectively, these results highlight six robust candidate genes with functional genetic support, laying a solid foundation for downstream investigations into epigenetic regulation and immune cell–specific mechanisms. 3. DNA Methylation Mediation MR Reveals Upstream Regulatory Mechanisms To explore whether the causal relationship between gene expression and PCOS is mediated by epigenetic regulation, we conducted DNA methylation–based mediation MR analyses using mQTL summary data for the six genes identified by SMR. The results revealed potential mediation effects for upstream CpG sites of SNAP29 and IL6R only (Tabel2, Figure 5A, Tabel S5), suggesting that specific DNA methylation signals may influence PCOS risk through the regulation of gene expression. For SNAP29 , two CpG sites (cg00335892 and cg20180721) were included in the analysis, both showing negative mediation effects (β = –0.0471 and –0.0376, respectively). However, the associations did not reach statistical significance (P = 0.46 and 0.41), indicating a weak regulatory trend that requires validation in larger cohorts. In contrast, four CpG sites associated with IL6R were analyzed, among which cg20688791 exhibited a borderline significant mediation effect (β = –0.1850, SE = 0.0944, P = 0.0501), as shown in Figure 5B. This provides preliminary evidence that methylation at this site may modulate IL6R expression and thereby contribute to PCOS susceptibility. Although most tested CpG sites did not show statistically significant mediation, the observed directional effects suggest that epigenetic regulation may represent an important upstream layer in the genetic architecture of PCOS. These findings warrant further validation using large-scale, high-resolution epigenomic datasets to better characterize the regulatory mechanisms involved. 4. Single-cell eQTL Analysis Reveals Immune Cell–Specific Causal Effects To elucidate the cell type–specific regulatory patterns of key PCOS-associated genes within the immune system, we conducted causal inference analyses based on sc-eQTL data across 14 major immune cell subtypes (TableS6). As shown in Figure 6, several genes demonstrated statistically significant causal effects in specific immune cell, suggesting that their functions may be mediated through immune microenvironment–dependent mechanisms. Specifically, CRELD1 exhibited negative causal effects in CD8⁺ T cells and dendritic cells (OR range: 0.36–0.63; all P < 0.05), suggesting a potential protective role within the cellular immune landscape. ITIH4 showed a significant positive association in CD4⁺ T cells (OR = 1.27, P = 9.6 × 10⁻³), possibly reflecting its involvement in inflammatory regulation within helper T cells. PON2 displayed a robust positive effect in monocytes (OR = 1.53, P = 8.9 × 10⁻⁵), reinforcing its potential role in oxidative stress and proinflammatory pathways. Additionally, SNAP29 showed a borderline-significant positive effect in natural killer (NK) cells (OR = 1.14, P = 0.031), implying a possible role in modulating innate immune responses relevant to PCOS pathophysiology. In contrast, neither IL6R nor NSFL1C demonstrated statistically significant associations in any single immune cell type, which may suggest broader systemic effects or limited statistical power due to cell-specific sample size constraints. These results are visualized in a forest plot (Figure 6), highlighting the direction and magnitude of causal effects for each gene across different immune cell types and underscoring their potential cell-environment–dependent regulatory roles in PCOS. Discussion PCOS is a common endocrine and metabolic disorder with complex etiologies and multifactorial mechanisms. While GWAS have identified numerous genetic loci associated with PCOS risk, a major challenge remains in pinpointing the functional causal genes underlying these associations and elucidating their biological mechanisms[ 32 , 33 ]. Bridging this gap is essential for translating genetic discoveries into biomedical insights. To address this challenge, our study leveraged a multi-omics integrative framework that combined two-sample MR, SMR, DNA methylation–based mediation analysis, and single-cell eQTL profiling. These complementary approaches enabled us to dissect the genetic regulation of PCOS from three critical dimensions: causal inference, upstream regulatory mechanisms, and immune cell–specific expression patterns. Through large-scale two-sample MR analyses using eQTL and pQTL data, we identified 60 candidate genes with significant causal associations with PCOS. Subsequent SMR analysis refined this list to six robust causal genes— CRELD1 , NSFL1C , ITIH4 , IL6R , SNAP29 , and PON2 —supported by colocalized genetic evidence. Notably, several of these genes have previously been implicated in chronic inflammation and metabolic dysregulation. IL6R encodes the interleukin-6 receptor, a key mediator of inflammatory signaling pathways, and has been widely reported to be involved in obesity, type 2 diabetes, and other metabolic diseases[ 34 – 36 ]. ITIH4 encodes inter-alpha-trypsin inhibitor heavy chain 4, an acute-phase protein primarily involved in extracellular matrix stabilization and inflammatory responses. Recent studies have shown that ITIH4 is implicated in various metabolic and inflammatory conditions[ 37 , 38 ]. PON2, a member of the paraoxonase family, is known for its antioxidative properties and regulatory roles in lipid metabolism[ 39 , 40 ]. These prior findings lend strong biological plausibility to our genetically supported identification of IL6R, ITIH4 and PON2 as potential contributors to PCOS pathophysiology. To explore the upstream regulatory mechanisms of the key causal genes, we further integrated mQTL data and conducted mediation MR analysis. Our results suggested a marginally significant mediation effect of the upstream CpG site cg20688791 on IL6R (P = 0.0501), indicating that this CpG site may be involved in the development of PCOS by indirectly regulating IL6R expression. Previous studies have shown that miR-520h can inhibit the viability of KGN cells and promote apoptosis by modulating IL6R and its downstream JAK/STAT pathway, thereby playing a critical role in the pathogenesis of PCOS[ 41 ]. However, no studies to date have investigated the epigenetic regulation of IL6R in PCOS, particularly regarding DNA methylation. Notably, emerging evidence from other chronic inflammatory diseases has revealed methylation-related features of IL6R. For instance, in patients with periodontitis, the promoter region of the IL6R gene exhibited hypermethylation during the early stages of disease and hypomethylation in advanced stages in peripheral immune cells[ 42 ]. These findings suggest that DNA methylation of IL6R may participate in the onset and progression of inflammatory diseases. In light of our mediation MR results, we propose that the DNA methylation regulation of IL6R in PCOS warrants further investigation, which may offer novel insights into the epigenetic mechanisms underlying PCOS. Recent studies have integrated sc-eQTL data from 14 major peripheral immune cell types with MR analyses to systematically identify cell type-specific causal genes associated with complex diseases such as atherosclerotic cardiovascular disease and pulmonary arterial hypertension[ 43 , 44 ]. However, these studies primarily focused on performing transcriptome-wide screening within each immune cell type to discover novel disease-associated genes and potential cell-specific pathogenic mechanisms. In contrast, our study employed a hypothesis-driven approach to specifically validate the causal effects of six key genes, which were prioritized in our previous multi-omics Mendelian randomization analyses, across different immune cell types. By leveraging sc-eQTL data, we conducted targeted cell type-specific MR analyses to evaluate the associations between the expression of these six genes and PCOS risk. This validation-oriented analytical strategy not only complements previous genome-wide screening efforts but also provides new insights into the immune cell-specific regulatory mechanisms of key causal genes in PCOS. Despite the systematic advantages of our study in integrating multi-omics data and inferring causal mechanisms, several limitations should be acknowledged. First, the GWAS, QTL, and mQTL datasets used in this study were primarily derived from European populations, which may introduce population stratification bias. Further validation in multi-ethnic cohorts is needed to assess the generalizability of our findings. Second, both SMR and MR analyses rely on the quality of instrumental variables (SNPs) and LD structures. Although we performed LD clumping and applied multiple sensitivity analyses to ensure robustness, the possibility of residual horizontal pleiotropy and reverse causality cannot be completely excluded. Third, the mediation MR analysis was conducted using summary-level mQTL data, without access to individual-level joint models, which may limit statistical power and introduce bias in the estimation of mediation proportions. Additionally, although single-cell eQTL analyses improved cell type-specific resolution, the limited sample sizes for certain immune cell subtypes may have led to underestimation of gene effects in these cells. Finally, all findings from this study warrant further experimental validation and confirmation in clinical cohorts. Future studies should further advance this research in several key directions. First, large-scale, multi-ethnic GWAS and QTL analyses are warranted to enhance the generalizability and robustness of causal inferences. Second, the integration of single-cell multi-omics data, such as single-cell methylation QTL (sc-mQTL) and single-cell chromatin accessibility (scATAC-seq) datasets, may help delineate the hierarchical structure of gene regulatory networks and identify key regulatory axes. Third, functional studies using PCOS-relevant cellular or animal models are essential to experimentally validate the regulatory pathways of IL6R, as well as their specific effects on ovarian function and inflammatory responses. Fourth, future investigations may incorporate PCOS-related phenotypes—such as insulin resistance, ovarian morphology, and hormone levels—into causal analyses to construct multilayered regulatory network maps. Fifth, coupling these findings with drug target databases could help evaluate the druggability of the identified genes and expand potential avenues for precision therapeutics. Conclusion In this study, we systematically integrated multi-omics data and identified six key candidate genes (CRELD1, NSFL1C, ITIH4, IL6R, SNAP29, and PON2) with robust causal associations with PCOS. Further analyses revealed that some of these genes may be regulated by specific DNA methylation sites and exhibit cell type-specific pathogenic effects within the immune microenvironment. These findings provide novel insights into the genetic regulatory mechanisms underlying PCOS and offer potential biomarkers for future targeted therapies and early screening strategies. Declarations Ethical Approval This study exclusively utilized publicly available, de-identified summary statistics from previously published GWAS, eQTL, pQTL, mQTL, and sc-eQTL datasets. No individual-level data were used. Therefore, ethical approval or informed consent was not required. Funding This work was supported by the National Natural Science Foundation of China (grant numbers 82071620 and 82371695). Availability of data and materials This study integrated multiple authoritative multi-omics datasets to ensure comprehensive and robust analyses. eQTLs data were obtained from the eQTLGen Consortium, pQTLs data were derived from the UK Biobank. Summary-level genome-wide association study (GWAS) data for PCOS were obtained from the FinnGen project (release R11). mQTLs data from the Genetics of DNA Methylation Consortium (GoDMC; http://mqtldb.godmc.org.uk/downloads). sc-eQTLs summary data released by the OneK1K project (https://onek1k.org/). Competing interests The authors declare no competing interests. Author contributions Junxiu Liu : Methodology, Visualization. Chengzi Huang : Investigation. Jun Jiao : Data analysis. Yue Sun : Data reduction. Yingxiu Ma : Methodology. Yang Yang : Review. Lan Chao : Review and editing. References Carson SA, Kallen AN: Diagnosis and Management of Infertility: A Review . Jama 2021, 326 (1):65-76. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS) . Human reproduction (Oxford, England) 2004, 19 (1):41-47. Chen W, Pang Y: Metabolic Syndrome and PCOS: Pathogenesis and the Role of Metabolites . Metabolites 2021, 11 (12). Vink JM, Sadrzadeh S, Lambalk CB, Boomsma DI: Heritability of Polycystic Ovary Syndrome in a Dutch Twin-Family Study . The Journal of Clinical Endocrinology & Metabolism 2006, 91 (6):2100-2104. Fu L, Zhang Z, Zhang A, Xu J, Huang X, Zheng Q, Cao Y, Wang L, Du J: Association Study Between FSHR Ala307Thr and Ser680Asn Variants and Polycystic Ovary Syndrome (PCOS) in Northern Chinese Han Women . Journal of assisted reproduction and genetics 2013, 30 (5):717-721. Li X, Li Y, Zhang B, Wang J, Yang Y, Du Y: Plasma cytokine levels and PCOS risk: Mendelian randomization analysis reveals IL6R as a preventive factor . Journal of ovarian research 2025, 18 (1):69. Qu J, Li B, Qiu M, Wang J, Chen Z, Li K, Teng X: Discovery of immune-related diagnostic biomarkers and construction of diagnostic model in varies polycystic ovary syndrome . Archives of gynecology and obstetrics 2022, 306 (5):1607-1615. Burgess S, Smith GD, Davies NM, Dudbridge F, Gill D, Glymour MM, Hartwig FP, Holmes MV, Minelli C, Relton CL et al : Guidelines for Performing Mendelian Randomization Investigations . Wellcome Open Research 2020, 4 :186. Ziegler A, Mwambi H, König IR: Mendelian Randomization Versus Path Models: Making Causal Inferences in Genetic Epidemiology . Human Heredity 2015, 79 (3-4):194-204. Slob EAW, Burgess S: A Comparison of Robust Mendelian Randomization Methods Using Summary Data . Genetic epidemiology 2020, 44 (4):313-329. Mattei AL, Bailly N, Meissner A: DNA methylation: a historical perspective . Trends in genetics : TIG 2022, 38 (7):676-707. Nishiyama A, Nakanishi M: Navigating the DNA methylation landscape of cancer . Trends in genetics : TIG 2021, 37 (11):1012-1027. Galbraith K, Snuderl M: DNA methylation as a diagnostic tool . Acta neuropathologica communications 2022, 10 (1):71. Clark-Boucher D, Zhou X, Du J, Liu Y, Needham BL, Smith JA, Mukherjee B: Methods for mediation analysis with high-dimensional DNA methylation data: Possible choices and comparisons . PLoS genetics 2023, 19 (11):e1011022. Lv Y, Jia Z, Wang Y, Huang Y, Li C, Chen X, Xia W, Liu H, Xu S, Li Y: Prenatal EDC exposure, DNA Methylation, and early childhood growth: A prospective birth cohort study . Environment international 2024, 190 :108872. Kim AY, Park YJ, Pan X, Shin KC, Kwak S-H, Bassas AF, Sallam RM, Park KS, Alfadda AA, Xu A et al : Obesity-Induced DNA Hypermethylation of the Adiponectin Gene Mediates Insulin Resistance . Nature communications 2015, 6 (1). Castellano‐Castillo D, Moreno‐Indias I, Sánchez‐Alcoholado L, Ramos‐Molina B, Alcaide‐Torres J, Morcillo S, Ocaña-Wilhelmi L, Tinahones FJ, Queipo‐Ortuño MI, Cardona F: Altered Adipose Tissue DNA Methylation Status in Metabolic Syndrome: Relationships Between Global DNA Methylation and Specific Methylation at Adipogenic, Lipid Metabolism and Inflammatory Candidate Genes and Metabolic Variables . Journal of clinical medicine 2019, 8 (1):87. Vatier C, Christin-Maitre S: Epigenetic/circadian clocks and PCOS . Human reproduction (Oxford, England) 2024, 39 (6):1167-1175. Mimouni NEH, Paiva I, Barbotin AL, Timzoura FE, Plassard D, Le Gras S, Ternier G, Pigny P, Catteau-Jonard S, Simon V et al : Polycystic ovary syndrome is transmitted via a transgenerational epigenetic process . Cell metabolism 2021, 33 (3):513-530.e518. Fujii R, Sato S, Tsuboi Y, Cardenas A, Suzuki K: DNA methylation as a mediator of associations between the environment and chronic diseases: A scoping review on application of mediation analysis . Epigenetics 2022, 17 (7):759-785. Zhang K, Chen G, He J, Chen Z, Pan M, Tong J, Liu F, Xiang H: DNA methylation mediates the effects of PM(2.5) and O(3) on ceramide metabolism: A novel mechanistic link between air pollution and insulin resistance . Journal of hazardous materials 2024, 469 :133864. Wang J, Zhu QW, Mai JH, Zhang S, Wang Y, Liang J, Zhou JY: A multi-omics study of brain tissue transcription and DNA methylation revealing the genetic pathogenesis of ADHD . Briefings in bioinformatics 2024, 25 (6). Luan YY, Zhang L, Peng YQ, Li YY, Liu RX, Yin CH: Immune regulation in polycystic ovary syndrome . Clinica chimica acta; international journal of clinical chemistry 2022, 531 :265-272. Mancini A, Bruno C, Vergani E, d'Abate C, Giacchi E, Silvestrini A: Oxidative Stress and Low-Grade Inflammation in Polycystic Ovary Syndrome: Controversies and New Insights . International journal of molecular sciences 2021, 22 (4). Oelen R, Vries DHd, Brugge H, Gordon GM, Vochteloo M, Ye C, Westra H-J, Franke L, Wijst MGPvd: Single-Cell RNA-sequencing Reveals Widespread Personalized, Context-Specific Gene Expression Regulation in Immune Cells . 2021. Ma T, Li H, Zhang X: Discovering Single-Cell eQTLs From scRNA-seq Data Only . 2021. Sun BB, Chiou J, Traylor M, Benner C, Hsu YH, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG et al : Plasma proteomic associations with genetics and health in the UK Biobank . Nature 2023, 622 (7982):329-338. Yazar S, Alquicira-Hernandez J, Wing K, Senabouth A, Gordon MG, Andersen S, Lu Q, Rowson A, Taylor TRP, Clarke L et al : Single-cell eQTL mapping identifies cell type-specific genetic control of autoimmune disease . Science (New York, NY) 2022, 376 (6589):eabf3041. Gkatzionis A, Burgess S, Newcombe PJ: Statistical methods for cis-Mendelian randomization with two-sample summary-level data . Genetic epidemiology 2023, 47 (1):3-25. Pierce BL, Ahsan H, Vanderweele TJ: Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants . International journal of epidemiology 2011, 40 (3):740-752. Zhu Z, Zhang F, Hu H, Bakshi A, Robinson MR, Powell JE, Montgomery GW, Goddard ME, Wray NR, Visscher PM et al : Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets . Nature genetics 2016, 48 (5):481-487. Zhu T, Goodarzi MO: Causes and Consequences of Polycystic Ovary Syndrome: Insights From Mendelian Randomization . The Journal of clinical endocrinology and metabolism 2022, 107 (3):e899-e911. Wang C, Wu W, Yang H, Ye Z, Zhao Y, Liu J, Mu L: Mendelian randomization analyses for PCOS: evidence, opportunities, and challenges . Trends in genetics : TIG 2022, 38 (5):468-482. Cupido AJ, Asselbergs FW, Natarajan P, Ridker PM, Hovingh GK, Schmidt AF: Dissecting the IL-6 pathway in cardiometabolic disease: A Mendelian randomization study on both IL6 and IL6R . British journal of clinical pharmacology 2022, 88 (6):2875-2884. Hu A, She Y, Wang Y, Cao X, Li R, Wang J, Yu L, Jiang H, Zhao Y: Effects of IL-6R Expression on IL-6/STAT3/HIF-1 α Signaling Pathway in a Mouse Model of Parkinson's Disease with Type 2 Diabetes Co-Morbidity . Discovery medicine 2024, 36 (186):1386-1397. Bowker N, Shah RL, Sharp SJ, Luan J, Stewart ID, Wheeler E, Ferreira MAR, Baras A, Wareham NJ, Langenberg C et al : Meta-analysis investigating the role of interleukin-6 mediated inflammation in type 2 diabetes . EBioMedicine 2020, 61 :103062. Zhang Y, Mei H, Xu K, Li C, Chang R, Qi H, Zhang Y, Zhang J: Associations between KCNQ1 and ITIH4 gene polymorphisms and infant weight gain in early life . Pediatric research 2022, 91 (5):1290-1295. Voisin S, Almén MS, Zheleznyakova GY, Lundberg L, Zarei S, Castillo S, Eriksson FE, Nilsson EK, Blüher M, Böttcher Y et al : Many obesity-associated SNPs strongly associate with DNA methylation changes at proximal promoters and enhancers . Genome medicine 2015, 7 :103. Van Dijck E, Diels S, Fransen E, Cremers TC, Verrijken A, Dirinck E, Hoischen A, Vandeweyer G, Vanden Berghe W, Van Gaal L et al : A Case-Control Study Supports Genetic Contribution of the PON Gene Family in Obesity and Metabolic Dysfunction Associated Steatotic Liver Disease . Antioxidants (Basel, Switzerland) 2024, 13 (9). Shin GC, Lee HM, Kim N, Yoo SK, Park HS, Choi LS, Kim KP, Lee AR, Seo SU, Kim KH: Paraoxonase-2 contributes to promoting lipid metabolism and mitochondrial function via autophagy activation . Scientific reports 2022, 12 (1):21483. Huang YH, Dong LP, Cui YG, Lu HY: MiR-520h inhibits viability and facilitates apoptosis of KGN cells through modulating IL6R and the JAK/STAT pathway . Reproductive biology 2022, 22 (1):100607. Cárdenas AM, Ardila LJ, Vernal R, Melgar-Rodríguez S, Hernández HG: Biomarkers of Periodontitis and Its Differential DNA Methylation and Gene Expression in Immune Cells: A Systematic Review . International journal of molecular sciences 2022, 23 (19). Chen M, Pang L, Huang T, Liang W, Xian Q, Guo D, Zhong M, Song L, Huang Z, Liu Y et al : Integrating single-cell RNA-seq, bulk RNA-seq, and Mendelian randomization to elucidate the role of HLA-DPA1 expression levels and non-classical monocytes in the pathogenesis of idiopathic pulmonary arterial hypertension . International journal of biological macromolecules 2025, 319 (Pt 3):145284. Ray A, Alabarse P, Malik R, Sargurupremraj M, Bernhagen J, Dichgans M, Baumeister SE, Georgakis MK: Single-cell transcriptome-wide Mendelian randomization and colocalization analyses uncover cell-specific mechanisms in atherosclerotic cardiovascular disease . American journal of human genetics 2025. Tables Table 1. Summary of SMR Analysis for Causal Genes Associated with Polycystic Ovary Syndrome . Gene topSNP A1 A2 Freq p_SMR p_HEIDI OR（95%CI） IL6R rs4553185 T C 0.54672 0.029226 0.8751668 1.8976 (1.0669-3.3751) CRELD1 rs2270894 G C 0.215706 0.001474 0.235592 2.1253 (1.3353–3.3827) ITIH4 rs12496077 A G 0.264414 0.009342 0.5924614 1.1828 (1.0422–1.3425) PON2 rs43038 A G 0.868787 0.049671 0.6078683 1.2981 (1.0004–1.6845) NSFL1C rs6033861 C T 0.287276 0.005293 0.4884989 1.8133 (1.1934–2.7550) SNAP29 rs2072514 G A 0.400596 0.039226 0.4003079 1.3472 (1.0148–1.7885) Legend: topSNP: Lead SNP for gene expression (eQTL); A1 / A2: Effect allele (A1) and reference allele (A2); Freq: Frequency of the effect allele (A1); p_SMR: P value for the SMR test assessing the causal effect of gene expression on PCOS; p_HEIDI: P value for the HEIDI test evaluating potential heterogeneity (p > 0.05 suggests no heterogeneity); OR (95% CI): Odds ratio and 95% confidence interval for the causal effect. Table 2 . Mediation Mendelian Randomization Results for CpG Sites Regulating IL6R and SNAP29 Gene CpG_ID beta1 beta2 beta_all beta12 beta12_p P IL6R cg09257526 -0.2958 0.4595 -0.2994 -0.1359 0.4541 0.1366 IL6R cg13549904 -0.2388 0.4595 -0.1556 -0.1097 0.7052 0.2288 IL6R cg20688791 -0.4025 0.4595 -0.3941 -0.185 0.4693 0.0501 IL6R cg21262032 -0.1097 0.4595 -0.1364 -0.0504 0.3694 0.5775 SNAP29 cg00335892 -0.4893 0.0963 -0.1026 -0.0471 0.4596 0.6352 SNAP29 cg20180721 -0.3901 0.0963 -0.0911 -0.0376 0.4125 0.6008 Legend: beta1: Effect size representing the causal effect of DNA methylation on gene expression (methylation → gene expression); beta2: Effect size representing the causal effect of gene expression on PCOS risk (gene expression → PCOS); beta_all: Total causal effect of DNA methylation on PCOS risk, including both direct and indirect effects; beta12 : Estimated mediation effect representing the indirect effect of DNA methylation on PCOS risk through gene expression (methylation → gene expression → PCOS); Mediation proportion (beta12_p) : Proportion of the total effect mediated by gene expression, calculated as the ratio of β₁₂ to β_all. Additional Declarations No competing interests reported. Supplementary Files supplementarytable.xlsx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7532908","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":521822528,"identity":"f9776a0f-0ab8-4ec9-af63-4eda970899b5","order_by":0,"name":"Junxiu Liu","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Junxiu","middleName":"","lastName":"Liu","suffix":""},{"id":521822529,"identity":"2de36a24-d723-49db-9674-d2d06dc42953","order_by":1,"name":"Chengzi Huang","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Chengzi","middleName":"","lastName":"Huang","suffix":""},{"id":521822530,"identity":"e5ded20e-1064-4208-a810-5814fcc6edf4","order_by":2,"name":"Jun Jiao","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Jiao","suffix":""},{"id":521822531,"identity":"c20450d4-d597-4aa6-bdaf-83e6e35d54ed","order_by":3,"name":"Yue Sun","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Sun","suffix":""},{"id":521822532,"identity":"96729da9-2c97-4b17-a303-70603a79731b","order_by":4,"name":"Yingxiu Ma","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Yingxiu","middleName":"","lastName":"Ma","suffix":""},{"id":521822533,"identity":"55572ccd-a5b7-4627-9e6d-f4de3b645e0d","order_by":5,"name":"Yang Yang","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Yang","middleName":"","lastName":"Yang","suffix":""},{"id":521822534,"identity":"9a688ec4-826d-4421-a4c4-90b34f9be84c","order_by":6,"name":"Lan Chao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+UlEQVRIiWNgGAWjYDACCSB+wMAgB2IfADIYG6CC+LUkMDAYg7UkkKIlEaSSgSgt/LObHz5IqLmT3i92+CHQljrZDQeYD97mYbDLw2nJnWPGBgnHnuXOnJ1mANRy2HjDAbZkax6G5GJcWgwkEswkEtgO5264nQDSciBxwwEeM2keIKMBp5b0bxIJ/w6n299O/wByGFAL/zcCWnLMJBLbDicYSOeAbGEG2cKGV4vEjZxig8S+w4YzbucUHEgwOGw88zCbseUcg2ScWvhnpG988OHbYXn+2embP3yoqJPtO9788MabCjucWtDdCcTMMMYoGAWjYBSMArIBAK+FXUkQC04BAAAAAElFTkSuQmCC","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":true,"prefix":"","firstName":"Lan","middleName":"","lastName":"Chao","suffix":""}],"badges":[],"createdAt":"2025-09-04 06:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7532908/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7532908/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92511127,"identity":"4e016137-9bac-48c2-a73a-a833153218ae","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2068741,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/3cd6994573ba880e77b9fa0f.docx"},{"id":92511125,"identity":"633539b2-af9c-4716-a49d-109a7be15d92","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8785,"visible":true,"origin":"","legend":"","description":"","filename":"22516c02bf1a40a3bb7497ad882b19a4.json","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/e32e39191f461ea87b38e267.json"},{"id":92511147,"identity":"47fef092-19a4-4c9b-9125-41d871ec814c","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2467876,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarytable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/efc097a0e4d7072ee543e307.xlsx"},{"id":92511375,"identity":"175550e8-062b-489f-8b17-d5525acac137","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134400,"visible":true,"origin":"","legend":"","description":"","filename":"22516c02bf1a40a3bb7497ad882b19a41enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/2a2a8571f63551529f97a8bb.xml"},{"id":92512656,"identity":"461c902c-c179-44ff-a7c6-c18ab3057731","added_by":"auto","created_at":"2025-09-30 13:42:06","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103598,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/b252859ce92130d90f4a3d7c.png"},{"id":92511129,"identity":"8414a5cd-8fbd-415f-901c-4a7576c664df","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42679,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/adf3d7b1557ba0f5ed65c75f.png"},{"id":92511136,"identity":"2c63765f-663d-4c2f-87f4-5eb010a276bd","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104388,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/4b0bdc0b316f853c2cd451c1.jpeg"},{"id":92511379,"identity":"4b685ae6-5e56-46c8-a5ab-294f545f5646","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":531480,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/e593d4c84567d679ed088985.jpeg"},{"id":92511137,"identity":"1af4e4d7-6b3a-4647-978e-f96d7b5e8d10","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":113486,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/ee089388e46c5869304e3a28.jpeg"},{"id":92511378,"identity":"be0cf0cd-c51e-4b08-8467-c90ba35cec98","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1063116,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/9921e89644e71e200b7d3fa5.jpeg"},{"id":92511139,"identity":"fed7be33-b70b-4807-bb83-f2cbaef35c8f","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30427,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/c6ecbb11874d74571a7824d9.png"},{"id":92511143,"identity":"86f27a2f-5848-4044-91ea-b4647d4757b6","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17272,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/577f341ff984d3302f9b7af6.png"},{"id":92511135,"identity":"e812641c-d5f0-4013-8cb9-bf9312567d99","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15253,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/843a5c21c8b644913da6811f.png"},{"id":92511380,"identity":"2e739412-2a55-4c70-acd5-bdfc539953b7","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84068,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/4558b83d7a83a57a80041cb0.png"},{"id":92511377,"identity":"1a6ff80b-cc9f-4585-8eba-8db016b09540","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22715,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/4f6d4b3fc17fb48b21129fcb.png"},{"id":92511141,"identity":"2962dbf2-a006-462d-8b39-b68bb9906588","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14365,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/2d42f26a7229a90a4e22e4f1.png"},{"id":92511146,"identity":"45d793c3-69dd-4752-8c26-bef4584f60d1","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134363,"visible":true,"origin":"","legend":"","description":"","filename":"22516c02bf1a40a3bb7497ad882b19a41structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/3b4631d29e9c2a8df29a630e.xml"},{"id":92511145,"identity":"59d8d078-acac-458f-9e15-731f1bd2b656","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":145430,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/837fbc9aef72b7a1c61b5b05.html"},{"id":92511123,"identity":"1e609165-39c8-4f2d-8e08-721e61e278fd","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":103598,"visible":true,"origin":"","legend":"\u003cp\u003eWorkflow of integrative multi-omics analysis to identify functional causal genes for PCOS.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/3042593193db1bea9572bab2.png"},{"id":92511373,"identity":"1bb25355-e6af-4d29-bc0e-f2840ad7e847","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":42679,"visible":true,"origin":"","legend":"\u003cp\u003eConceptual framework of two-sample Mendelian randomization (MR) analysis.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/e0969b10cdb51a05d6608a7a.png"},{"id":92511126,"identity":"731508b5-eb58-4f94-b8a7-218da32395c6","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":51431,"visible":true,"origin":"","legend":"\u003cp\u003eVenn diagram showing the overlap of PCOS-associated causal genes\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/a01f294155212025b6164f6a.png"},{"id":92511133,"identity":"a10f0224-45d4-49cf-a64e-d3b7d8b0706d","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":326397,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot of ORs for candidate genes associated with PCOS based on eQTL-, pQTL-, and SMR-based MR analyses.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/5004d730730e8e3357903602.png"},{"id":92511376,"identity":"5a0bf17c-5732-49a6-bcea-1aac41f0055e","added_by":"auto","created_at":"2025-09-30 13:34:06","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":82600,"visible":true,"origin":"","legend":"\u003cp\u003eDNA methylation mediation analysis of PCOS-associated genes\u003c/p\u003e\n\u003cp\u003e(A) Heatmap showing the estimated mediation effects (β coefficients) of CpG sites located upstream of IL6R and SNAP29 on PCOS risk via gene expression.\u003c/p\u003e\n\u003cp\u003e(B) Methylation at cg20688791 partially mediates the effect of IL6R expression on PCOS risk, with a mediation proportion of 46%.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/9a12839d88a39709a957f2f1.png"},{"id":92511128,"identity":"8d339df8-5b13-4de4-81ca-83786056b981","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":29333,"visible":true,"origin":"","legend":"\u003cp\u003eCell type-specific causal effects of PCOS-associated genes identified through single-cell eQTL-based MR analysis.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/520a83a0dd082f63ba675b1f.png"},{"id":101761492,"identity":"91eeb731-c809-4b8e-8d0f-9a5000dd9a1e","added_by":"auto","created_at":"2026-02-03 11:18:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3345851,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/bec0b49d-ff66-4736-a444-d7c233c555a9.pdf"},{"id":92511132,"identity":"4b574b07-2b16-47d1-b4ae-73cbfc06267e","added_by":"auto","created_at":"2025-09-30 13:26:06","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":2467876,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarytable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7532908/v1/8fd96be9e61043938a13ea2d.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Causal Genes and Immune-Epigenetic Mechanisms Underlying Polycystic Ovary Syndrome: A Multi-Omics Mendelian Randomization Study","fulltext":[{"header":"Key points","content":"\u003col\u003e\n \u003cli\u003eMulti-omics Mendelian randomization identified six causal genes for PCOS.\u003c/li\u003e\n \u003cli\u003eSMR analysis confirmed colocalization between genetic variants, gene expression, and PCOS risk.\u003c/li\u003e\n \u003cli\u003eDNA methylation mediation analysis implicated epigenetic regulation of IL6R and SNAP29.\u003c/li\u003e\n \u003cli\u003eSingle-cell eQTL analysis revealed immune cell type\u0026ndash;specific effects of causal genes.\u003c/li\u003e\n \u003cli\u003eFindings provide novel mechanistic insights and potential therapeutic targets for PCOS.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Introduction","content":"\u003cp\u003ePolycystic ovary syndrome (PCOS) is a common endocrine disorder affecting approximately 5% to 10% of women of reproductive age worldwide[1]. It is characterized by a heterogeneous clinical presentation, including menstrual irregularities, hyperandrogenism, and polycystic ovarian morphology[2]. Beyond its impact on fertility, PCOS is strongly associated with metabolic disturbances such as insulin resistance, type 2 diabetes, and cardiovascular disease, posing significant threats to both physical and mental health[3]. Although the precise etiology of PCOS remains elusive, accumulating evidence supports a substantial genetic predisposition. The heritability of PCOS has been estimated to be as high as 70%, indicating that genetic factors are instrumental in its development. For instance, twin studies have demonstrated a strong familial correlation for phenotypic features of PCOS, suggesting a polygenic inheritance pattern rather than a single genetic defect[4]. Genetic variants associated with PCOS include polymorphisms in genes involved in insulin signaling and hormone regulation, such as insulin gene polymorphisms and variations in the FSH receptor[5]. With the advancement of genomics and epigenomics technologies, increasing efforts have been directed toward identifying PCOS-related pathogenic genes and their regulatory mechanisms using multi-omics approaches[6, 7], with the goal of providing theoretical foundations and novel targets for clinical intervention.\u003c/p\u003e\n\u003cp\u003eMendelian randomization (MR) is an analytical framework that leverages naturally occurring genetic variants\u0026mdash;typically single nucleotide polymorphisms (SNPs)\u0026mdash;as instrumental variables to infer causality between exposures and disease outcomes[8, 9]. By simulating the conditions of a randomized controlled trial, MR can effectively mitigate confounding and reverse causation, which are common limitations in traditional observational studies. This has made MR increasingly valuable in elucidating causal relationships in complex diseases[10].\u003c/p\u003e\n\u003cp\u003eDNA methylation is a key epigenetic modification that plays a crucial role in regulating gene expression and mediating phenotypic variation. It refers to the addition of a methyl group to the cytosine residue of DNA, predominantly at CpG dinucleotides, leading to the formation of 5-methylcytosine[11]. This modification can modulate gene expression by altering chromatin structure, affecting DNA accessibility, and influencing the binding of transcription factors. Although DNA methylation patterns are often stable, accumulating evidence suggests that they can also exhibit plasticity in response to environmental cues such as nutritional status, psychological stress, or toxic exposures[12, 13]. These dynamic changes in DNA methylation are considered to underlie, at least in part, the molecular mechanisms through which organisms adapt their gene expression programs and phenotypic traits to external environmental conditions[14, 15]. Aberrant methylation patterns have been implicated in a wide range of metabolic and reproductive disorders[16, 17], including PCOS[18, 19]. Mediation analysis of DNA methylation, gene expression, and disease risk can help uncover potential regulatory pathways and identify functionally relevant loci within gene regulatory networks[20-22].\u003c/p\u003e\n\u003cp\u003eThe immune system has also been increasingly recognized as a key contributor to PCOS pathogenesis. Emerging evidence suggests that PCOS is characterized by a state of chronic low-grade inflammation, marked by elevated levels of pro-inflammatory cytokines and immune cell dysregulation[23, 24]. Investigating the expression of causal genes in specific immune cell types at the single-cell level may provide insight into their functional roles within the immune microenvironment and help explain phenotypic heterogeneity in PCOS[25, 26]. However, to date, no study has systematically integrated multi-omics MR, summary-data-based MR (SMR), methylation quantitative trait loci(mQTLs), and single-cell expression quantitative trait loci (sc-eQTLs) approaches to unravel the mechanisms underlying PCOS.\u003c/p\u003e\n\u003cp\u003eIn this study, we systematically integrated expression quantitative trait loci (eQTLs), protein QTL (pQTLs) and genome-wide association study (GWAS) to identify key PCOS-associated genes through two-sample MR and SMR analyses. We further explored upstream epigenetic regulation using DNA methylation\u0026ndash;based mediation MR, and evaluated the cell-type\u0026ndash;specific expression of candidate genes across 14 immune cell subtypes using sc-eQTLs data. By leveraging genetic, epigenetic, and immune-cell\u0026ndash;specific regulatory evidence, our study aims to dissect the complex etiology of PCOS and advance precision medicine in reproductive endocrinology.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003e1. Data Sources\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study integrated multiple authoritative multi-omics datasets to ensure comprehensive and robust analyses. eQTLs data were obtained from the \u003cem\u003eeQTLGen Consortium\u003c/em\u003e, which provides genome-wide cis-eQTL summary statistics based on peripheral blood samples from 31,684 individuals of European ancestry (https://www.eqtlgen.org/cis-eqtls.html). pQTLs data were derived from the \u003cem\u003eUK Biobank\u003c/em\u003e, in which plasma proteomic profiles of 54,219 participants were measured using the SomaScan platform and linked to genetic variation through association analysis[27]. Summary-level genome-wide association study (GWAS) data for PCOS were obtained from the \u003cem\u003eFinnGen project (release R11)\u003c/em\u003e, comprising 1,909 cases and 241,998 controls (https://r11.finngen.fi/). To explore upstream epigenetic regulatory mechanisms, we incorporated mQTLs data from the \u003cem\u003eGenetics of DNA Methylation Consortium \u003c/em\u003e(GoDMC; http://mqtldb.godmc.org.uk/downloads), which catalog extensive associations between CpG methylation levels and genetic variants. To further investigate the cell type\u0026ndash;specific regulation of candidate genes, we utilized sc-eQTLs summary data released by the \u003cem\u003eOneK1K project \u003c/em\u003e(https://onek1k.org/), covering 14 major immune cell subtypes[28]. This dataset represents a cutting-edge resource for dissecting cell-specific genetic effects in human immune traits. An overview of the analytical workflow is illustrated in Figure 1. Instrumental variables used in each MR framework are available in Supplementary Data 1\u0026ndash;4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. Two-Sample Mendelian Randomization (MR) Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, we performed two-sample Mendelian randomization analyses to evaluate the potential causal effects of gene expression and protein levels in peripheral blood on the risk of PCOS. The analytical framework is illustrated in Figure 2. We selected cis-QTL variants significantly associated with exposures (P \u0026lt; 5 \u0026times; 10⁻⁸) and with sufficient instrument strength (F-statistic\u0026gt;10). To ensure independence among instruments, we applied linkage disequilibrium (LD) clumping using PLINK 2.0, setting an LD threshold of r\u0026sup2; \u0026lt; 0.1 and a window size of 10,000 kb, retaining only independent lead SNPs. Exposure and outcome datasets were harmonized to align effect alleles before MR estimation[29, 30].\u003c/p\u003e\n\u003cp\u003eThe primary MR analysis was conducted using the inverse variance weighted (IVW) method to estimate causal effects. Robustness and sensitivity were further evaluated using complementary methods, including MR-Egger regression, the weighted median estimator, Cochran\u0026rsquo;s Q test for heterogeneity, MR-PRESSO for horizontal pleiotropy detection, and leave-one-out analysis. All analyses were performed in R (version 4.3.2), primarily using the TwoSampleMR and MRPRESSO packages, with visualization via ggplot2 and forestplot. To ensure the reliability of causal inference, we applied the following filtering criteria: (1) a nominally significant IVW P-value (\u0026lt; 0.05), (2) a non-significant MR-Egger intercept (P \u0026gt; 0.05) to exclude horizontal pleiotropy. Genes meeting both criteria were considered to have stable causal signals and were retained for downstream analysis. Finally, we intersected the results from eQTL-based and pQTL-based MR to derive a set of 60 candidate genes for subsequent validation and functional investigation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Summary-data\u0026ndash;based Mendelian Randomization (SMR) Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further evaluate whether the expression or protein levels of the 60 candidate genes mediate PCOS risk through colocalized genetic variants, we applied SMR. This method integrates eQTLs (or pQTLs) and GWAS summary statistics to test whether the same SNP is associated with both gene expression and disease phenotype, thereby inferring whether the gene may functionally mediate disease risk[31].\u003c/p\u003e\n\u003cp\u003eTo ensure the independence of exposure data from the prior MR analyses, we did not reuse the eQTLGen dataset. Instead, we utilized eQTLs data from whole blood provided by the GTEx v8 project (https://gtexportal.org/home/) as the source of gene expression instruments for SMR. The SMR analysis consisted of two primary statistical tests: (1) the SMR test (p_SMR), which assesses the association between gene expression and disease phenotype; and (2) the HEIDI (Heterogeneity in Dependent Instruments) test (p_HEIDI), which evaluates whether the observed association is likely driven by a single causal variant rather than by multiple colocalized but distinct signals. A non-significant p_HEIDI (p \u0026gt; 0.05) supports the existence of a shared causal variant.\u003c/p\u003e\n\u003cp\u003eGenes were retained if they met the following criteria: p_SMR \u0026lt; 0.05, p_HEIDI \u0026gt; 0.05, and an odds ratio (OR) \u0026gt; 1, indicating that increased gene expression is associated with higher PCOS risk. All analyses were conducted using the SMR software (https://yanglab.westlake.edu.cn/software/smr/#DataResource). Based on these criteria, six genes were prioritized as likely functional mediators of PCOS risk: \u003cem\u003eCRELD1\u003c/em\u003e, \u003cem\u003eNSFL1C\u003c/em\u003e, \u003cem\u003eITIH4\u003c/em\u003e, \u003cem\u003eIL6R\u003c/em\u003e, \u003cem\u003eSNAP29\u003c/em\u003e, and \u003cem\u003ePON2\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4. DNA Methylation Mediation MR Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo systematically assess whether DNA methylation mediates the relationship between gene expression and PCOS risk, we conducted mediation MR analyses based on mQTLs data, focusing on the six key genes identified by SMR analysis (\u003cem\u003eCRELD1\u003c/em\u003e, \u003cem\u003eNSFL1C\u003c/em\u003e, \u003cem\u003eITIH4\u003c/em\u003e, \u003cem\u003eIL6R\u003c/em\u003e, \u003cem\u003eSNAP29\u003c/em\u003e, and \u003cem\u003ePON2\u003c/em\u003e). The overall analytical workflow was as follows: (1) CpG sites related to each candidate gene were identified using the EWAS DataHub provided by the National Genomics Data Center (https://ngdc.cncb.ac.cn/ewas/datahub/exploration); (2) SNPs significantly associated with these CpG sites were extracted from publicly available summary-level mQTL datasets. For each SNP, we recorded the effect allele, effect size (\u0026beta;), standard error (SE), and P-value, and grouped the data by gene to facilitate downstream analysis; (3) To ensure instrument strength and independence, we applied linkage disequilibrium (LD) clumping for each CpG site to retain only genome-wide significant and independent SNPs using the following parameters: P \u0026lt; 5 \u0026times; 10⁻⁸, clump_kb = 10,000, clump_r\u0026sup2; = 0.1, and clump_p = 1. \u003c/p\u003e\n\u003cp\u003eThese mQTL SNPs were then used as instrumental variables in two-sample MR models: ① to estimate the causal effect of CpG methylation on PCOS risk; ② to estimate the causal effect of CpG methylation on gene expression. The indirect (mediation) effect of DNA methylation was then calculated by combining the two causal estimates for each CpG site, representing the extent to which methylation influences PCOS risk via gene expression. Confidence intervals and statistical significance for the mediation effect were estimated using the Delta method. In addition, the mediation proportion was computed to evaluate the contribution of the indirect effect relative to the total effect of DNA methylation on PCOS. This approach enabled the identification of specific CpG sites that may function as upstream regulators of key PCOS-related genes through epigenetic mechanisms.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5. Single-cell eQTLs Analysis in Immune Cell Types\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate the genetic regulatory characteristics of key candidate genes in distinct immune cell types and to evaluate their potential causal relationships with PCOS, we performed cell type\u0026ndash;specific MR analyses based on publicly available sc-eQTL summary data. Specifically, we retrieved sc-eQTL summary-level statistics for 14 major immune cell subtypes from the OneK1K project. For each of the six PCOS-associated genes identified in the SMR analysis (\u003cem\u003eCRELD1\u003c/em\u003e, \u003cem\u003eNSFL1C\u003c/em\u003e, \u003cem\u003eITIH4\u003c/em\u003e, \u003cem\u003eIL6R\u003c/em\u003e, \u003cem\u003eSNAP29\u003c/em\u003e, \u003cem\u003ePON2\u003c/em\u003e), we extracted cell-specific eQTL data across all immune subtypes, generating a series of \u0026ldquo;gene\u0026ndash;cell type\u0026rdquo; datasets for downstream analysis.\u003c/p\u003e\n\u003cp\u003eTo ensure instrument validity and independence, we applied linkage disequilibrium (LD) clumping to each gene\u0026ndash;cell pair using the following parameters: significance threshold: P \u0026lt; 0.05；LD window: 100 kb；LD r\u0026sup2; threshold: r\u0026sup2; \u0026lt; 0.3 (i.e., clump_kb = 100, clump_r\u0026sup2; = 0.3, clump_p = 1). This step was performed to select independent SNPs that were significantly associated with gene expression in each immune cell context, enabling robust causal inference. The filtered sc-eQTL datasets for each gene\u0026ndash;cell combination were treated as exposure variables, while the PCOS GWAS summary statistics served as the outcome. Two-sample MR analyses were conducted to assess whether the expression of candidate genes in specific immune cell types causally influenced PCOS risk. Causal estimates were derived using the inverse-variance weighted (IVW) method. MR-Egger regression, MR-PRESSO, and other sensitivity tests were applied where appropriate to evaluate robustness and detect potential horizontal pleiotropy.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003e1. Initial MR Screening Identifies PCOS-Associated Genes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 1,715 eQTL-associated genes and 182 pQTL-associated genes were identified as significantly associated with PCOS (Supplementary material: Table S1-S2). By intersecting the results from both analyses, we identified 60 genes whose expression and protein levels were both significantly associated with PCOS risk (Figure 3, Table S3). This intersection strategy enhances the reliability of causal inference by prioritizing genes supported by both transcriptomic and proteomic evidence, and reduces false positives potentially introduced by platform-specific effects or model assumptions. These 60 candidate genes were subsequently selected for integrative multi-omics analyses and functional validation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. SMR Analysis Identifies Six Causal Genes with Colocalized Genetic Signals\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBuilding on the two-sample MR results, we conducted SMR analyses on the 60 candidate genes to assess whether their expression or protein levels mediate PCOS risk via colocalized genetic variants (Table S4). Six genes\u0026mdash;\u003cem\u003eCRELD1\u003c/em\u003e, \u003cem\u003eNSFL1C\u003c/em\u003e, \u003cem\u003eITIH4\u003c/em\u003e, \u003cem\u003eIL6R\u003c/em\u003e, \u003cem\u003eSNAP29\u003c/em\u003e, and \u003cem\u003ePON2\u003c/em\u003e\u0026mdash;met the predefined criteria for SMR significance (p_SMR \u0026lt; 0.05), absence of heterogeneity (p_HEIDI \u0026gt; 0.05), and positive association (OR \u0026gt; 1), suggesting that they may exert functional causal effects through genetically regulated expression (Table1).\u003c/p\u003e\n\u003cp\u003eAmong them, \u003cem\u003eCRELD1\u003c/em\u003e showed the most significant association in the SMR model (OR = 2.13, 95% CI: 1.34\u0026ndash;3.38, p_SMR = 1.47 \u0026times; 10⁻\u0026sup3;), indicating that increased expression of \u003cem\u003eCRELD1\u003c/em\u003e may substantially elevate PCOS risk. Both \u003cem\u003eNSFL1C\u003c/em\u003e and \u003cem\u003eIL6R\u003c/em\u003e also displayed consistent positive associations (OR = 1.81 and 1.90, respectively). As illustrated in Figure 4, we visualized the effect sizes and directions of these six genes across eQTL-, pQTL-, and SMR-based MR analyses using an odds ratio (OR) forest plot. All six genes demonstrated concordant risk-enhancing effects across multiple omics layers, further supporting their potential roles in the genetic architecture of PCOS. Collectively, these results highlight six robust candidate genes with functional genetic support, laying a solid foundation for downstream investigations into epigenetic regulation and immune cell\u0026ndash;specific mechanisms.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. DNA Methylation Mediation MR Reveals Upstream Regulatory Mechanisms\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo explore whether the causal relationship between gene expression and PCOS is mediated by epigenetic regulation, we conducted DNA methylation\u0026ndash;based mediation MR analyses using mQTL summary data for the six genes identified by SMR. The results revealed potential mediation effects for upstream CpG sites of \u003cem\u003eSNAP29\u003c/em\u003e and \u003cem\u003eIL6R\u003c/em\u003e only (Tabel2, Figure 5A, Tabel S5), suggesting that specific DNA methylation signals may influence PCOS risk through the regulation of gene expression.\u003c/p\u003e\n\u003cp\u003eFor \u003cem\u003eSNAP29\u003c/em\u003e, two CpG sites (cg00335892 and cg20180721) were included in the analysis, both showing negative mediation effects (\u0026beta; = \u0026ndash;0.0471 and \u0026ndash;0.0376, respectively). However, the associations did not reach statistical significance (P = 0.46 and 0.41), indicating a weak regulatory trend that requires validation in larger cohorts. In contrast, four CpG sites associated with \u003cem\u003eIL6R\u003c/em\u003e were analyzed, among which cg20688791 exhibited a borderline significant mediation effect (\u0026beta; = \u0026ndash;0.1850, SE = 0.0944, P = 0.0501), as shown in Figure 5B. This provides preliminary evidence that methylation at this site may modulate \u003cem\u003eIL6R\u003c/em\u003e expression and thereby contribute to PCOS susceptibility. Although most tested CpG sites did not show statistically significant mediation, the observed directional effects suggest that epigenetic regulation may represent an important upstream layer in the genetic architecture of PCOS. These findings warrant further validation using large-scale, high-resolution epigenomic datasets to better characterize the regulatory mechanisms involved.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4. Single-cell eQTL Analysis Reveals Immune Cell\u0026ndash;Specific Causal Effects\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo elucidate the cell type\u0026ndash;specific regulatory patterns of key PCOS-associated genes within the immune system, we conducted causal inference analyses based on sc-eQTL data across 14 major immune cell subtypes (TableS6). As shown in Figure 6, several genes demonstrated statistically significant causal effects in specific immune cell, suggesting that their functions may be mediated through immune microenvironment\u0026ndash;dependent mechanisms.\u003c/p\u003e\n\u003cp\u003eSpecifically, \u003cem\u003eCRELD1\u003c/em\u003e exhibited negative causal effects in CD8⁺ T cells and dendritic cells (OR range: 0.36\u0026ndash;0.63; all P \u0026lt; 0.05), suggesting a potential protective role within the cellular immune landscape. \u003cem\u003eITIH4\u003c/em\u003e showed a significant positive association in CD4⁺ T cells (OR = 1.27, P = 9.6 \u0026times; 10⁻\u0026sup3;), possibly reflecting its involvement in inflammatory regulation within helper T cells. \u003cem\u003ePON2\u003c/em\u003e displayed a robust positive effect in monocytes (OR = 1.53, P = 8.9 \u0026times; 10⁻⁵), reinforcing its potential role in oxidative stress and proinflammatory pathways. Additionally, \u003cem\u003eSNAP29\u003c/em\u003e showed a borderline-significant positive effect in natural killer (NK) cells (OR = 1.14, P = 0.031), implying a possible role in modulating innate immune responses relevant to PCOS pathophysiology.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn contrast, neither \u003cem\u003eIL6R\u003c/em\u003e nor \u003cem\u003eNSFL1C\u003c/em\u003e demonstrated statistically significant associations in any single immune cell type, which may suggest broader systemic effects or limited statistical power due to cell-specific sample size constraints. These results are visualized in a forest plot (Figure 6), highlighting the direction and magnitude of causal effects for each gene across different immune cell types and underscoring their potential cell-environment\u0026ndash;dependent regulatory roles in PCOS.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003ePCOS is a common endocrine and metabolic disorder with complex etiologies and multifactorial mechanisms. While GWAS have identified numerous genetic loci associated with PCOS risk, a major challenge remains in pinpointing the functional causal genes underlying these associations and elucidating their biological mechanisms[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Bridging this gap is essential for translating genetic discoveries into biomedical insights. To address this challenge, our study leveraged a multi-omics integrative framework that combined two-sample MR, SMR, DNA methylation\u0026ndash;based mediation analysis, and single-cell eQTL profiling. These complementary approaches enabled us to dissect the genetic regulation of PCOS from three critical dimensions: causal inference, upstream regulatory mechanisms, and immune cell\u0026ndash;specific expression patterns.\u003c/p\u003e\u003cp\u003eThrough large-scale two-sample MR analyses using eQTL and pQTL data, we identified 60 candidate genes with significant causal associations with PCOS. Subsequent SMR analysis refined this list to six robust causal genes\u0026mdash;\u003cem\u003eCRELD1\u003c/em\u003e, \u003cem\u003eNSFL1C\u003c/em\u003e, \u003cem\u003eITIH4\u003c/em\u003e, \u003cem\u003eIL6R\u003c/em\u003e, \u003cem\u003eSNAP29\u003c/em\u003e, and \u003cem\u003ePON2\u003c/em\u003e\u0026mdash;supported by colocalized genetic evidence. Notably, several of these genes have previously been implicated in chronic inflammation and metabolic dysregulation. IL6R encodes the interleukin-6 receptor, a key mediator of inflammatory signaling pathways, and has been widely reported to be involved in obesity, type 2 diabetes, and other metabolic diseases[\u003cspan additionalcitationids=\"CR35\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. ITIH4 encodes inter-alpha-trypsin inhibitor heavy chain 4, an acute-phase protein primarily involved in extracellular matrix stabilization and inflammatory responses. Recent studies have shown that ITIH4 is implicated in various metabolic and inflammatory conditions[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. PON2, a member of the paraoxonase family, is known for its antioxidative properties and regulatory roles in lipid metabolism[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. These prior findings lend strong biological plausibility to our genetically supported identification of IL6R, ITIH4 and PON2 as potential contributors to PCOS pathophysiology.\u003c/p\u003e\u003cp\u003eTo explore the upstream regulatory mechanisms of the key causal genes, we further integrated mQTL data and conducted mediation MR analysis. Our results suggested a marginally significant mediation effect of the upstream CpG site cg20688791 on IL6R (P\u0026thinsp;=\u0026thinsp;0.0501), indicating that this CpG site may be involved in the development of PCOS by indirectly regulating IL6R expression. Previous studies have shown that miR-520h can inhibit the viability of KGN cells and promote apoptosis by modulating IL6R and its downstream JAK/STAT pathway, thereby playing a critical role in the pathogenesis of PCOS[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. However, no studies to date have investigated the epigenetic regulation of IL6R in PCOS, particularly regarding DNA methylation. Notably, emerging evidence from other chronic inflammatory diseases has revealed methylation-related features of IL6R. For instance, in patients with periodontitis, the promoter region of the IL6R gene exhibited hypermethylation during the early stages of disease and hypomethylation in advanced stages in peripheral immune cells[\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. These findings suggest that DNA methylation of IL6R may participate in the onset and progression of inflammatory diseases. In light of our mediation MR results, we propose that the DNA methylation regulation of IL6R in PCOS warrants further investigation, which may offer novel insights into the epigenetic mechanisms underlying PCOS.\u003c/p\u003e\u003cp\u003eRecent studies have integrated sc-eQTL data from 14 major peripheral immune cell types with MR analyses to systematically identify cell type-specific causal genes associated with complex diseases such as atherosclerotic cardiovascular disease and pulmonary arterial hypertension[\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. However, these studies primarily focused on performing transcriptome-wide screening within each immune cell type to discover novel disease-associated genes and potential cell-specific pathogenic mechanisms. In contrast, our study employed a hypothesis-driven approach to specifically validate the causal effects of six key genes, which were prioritized in our previous multi-omics Mendelian randomization analyses, across different immune cell types. By leveraging sc-eQTL data, we conducted targeted cell type-specific MR analyses to evaluate the associations between the expression of these six genes and PCOS risk. This validation-oriented analytical strategy not only complements previous genome-wide screening efforts but also provides new insights into the immune cell-specific regulatory mechanisms of key causal genes in PCOS.\u003c/p\u003e\u003cp\u003eDespite the systematic advantages of our study in integrating multi-omics data and inferring causal mechanisms, several limitations should be acknowledged. First, the GWAS, QTL, and mQTL datasets used in this study were primarily derived from European populations, which may introduce population stratification bias. Further validation in multi-ethnic cohorts is needed to assess the generalizability of our findings. Second, both SMR and MR analyses rely on the quality of instrumental variables (SNPs) and LD structures. Although we performed LD clumping and applied multiple sensitivity analyses to ensure robustness, the possibility of residual horizontal pleiotropy and reverse causality cannot be completely excluded. Third, the mediation MR analysis was conducted using summary-level mQTL data, without access to individual-level joint models, which may limit statistical power and introduce bias in the estimation of mediation proportions. Additionally, although single-cell eQTL analyses improved cell type-specific resolution, the limited sample sizes for certain immune cell subtypes may have led to underestimation of gene effects in these cells. Finally, all findings from this study warrant further experimental validation and confirmation in clinical cohorts.\u003c/p\u003e\u003cp\u003eFuture studies should further advance this research in several key directions. First, large-scale, multi-ethnic GWAS and QTL analyses are warranted to enhance the generalizability and robustness of causal inferences. Second, the integration of single-cell multi-omics data, such as single-cell methylation QTL (sc-mQTL) and single-cell chromatin accessibility (scATAC-seq) datasets, may help delineate the hierarchical structure of gene regulatory networks and identify key regulatory axes. Third, functional studies using PCOS-relevant cellular or animal models are essential to experimentally validate the regulatory pathways of IL6R, as well as their specific effects on ovarian function and inflammatory responses. Fourth, future investigations may incorporate PCOS-related phenotypes\u0026mdash;such as insulin resistance, ovarian morphology, and hormone levels\u0026mdash;into causal analyses to construct multilayered regulatory network maps. Fifth, coupling these findings with drug target databases could help evaluate the druggability of the identified genes and expand potential avenues for precision therapeutics.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn this study, we systematically integrated multi-omics data and identified six key candidate genes (CRELD1, NSFL1C, ITIH4, IL6R, SNAP29, and PON2) with robust causal associations with PCOS. Further analyses revealed that some of these genes may be regulated by specific DNA methylation sites and exhibit cell type-specific pathogenic effects within the immune microenvironment. These findings provide novel insights into the genetic regulatory mechanisms underlying PCOS and offer potential biomarkers for future targeted therapies and early screening strategies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthical Approval\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eThis study exclusively utilized publicly available, de-identified summary statistics from previously published GWAS, eQTL, pQTL, mQTL, and sc-eQTL datasets. No individual-level data were used. Therefore, ethical approval or informed consent was not required.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the National Natural Science Foundation of China\u003c/p\u003e\n\u003cp\u003e(grant numbers 82071620 and 82371695).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study integrated multiple authoritative multi-omics datasets to ensure comprehensive and robust analyses. eQTLs data were obtained from the eQTLGen Consortium, pQTLs data were derived from the UK Biobank. Summary-level genome-wide association study (GWAS) data for PCOS were obtained from the FinnGen project (release R11). mQTLs data from the Genetics of DNA Methylation Consortium (GoDMC; http://mqtldb.godmc.org.uk/downloads). sc-eQTLs summary data released by the OneK1K project (https://onek1k.org/).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJunxiu Liu\u003c/strong\u003e: Methodology, Visualization. \u003cstrong\u003eChengzi Huang\u003c/strong\u003e: Investigation. \u003cstrong\u003eJun Jiao\u003c/strong\u003e: Data analysis. \u003cstrong\u003eYue Sun\u003c/strong\u003e: Data reduction. \u003cstrong\u003eYingxiu Ma\u003c/strong\u003e: Methodology. \u003cstrong\u003eYang Yang\u003c/strong\u003e: Review. \u003cstrong\u003eLan Chao\u003c/strong\u003e: Review and editing.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCarson SA, Kallen AN: \u003cstrong\u003eDiagnosis and Management of Infertility: A Review\u003c/strong\u003e. \u003cem\u003eJama \u003c/em\u003e2021, \u003cstrong\u003e326\u003c/strong\u003e(1):65-76.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eRevised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS)\u003c/strong\u003e. \u003cem\u003eHuman reproduction (Oxford, England) \u003c/em\u003e2004, \u003cstrong\u003e19\u003c/strong\u003e(1):41-47.\u003c/li\u003e\n\u003cli\u003eChen W, Pang Y: \u003cstrong\u003eMetabolic Syndrome and PCOS: Pathogenesis and the Role of Metabolites\u003c/strong\u003e. \u003cem\u003eMetabolites \u003c/em\u003e2021, \u003cstrong\u003e11\u003c/strong\u003e(12).\u003c/li\u003e\n\u003cli\u003eVink JM, Sadrzadeh S, Lambalk CB, Boomsma DI: \u003cstrong\u003eHeritability of Polycystic Ovary Syndrome in a Dutch Twin-Family Study\u003c/strong\u003e. \u003cem\u003eThe Journal of Clinical Endocrinology \u0026amp; Metabolism \u003c/em\u003e2006, \u003cstrong\u003e91\u003c/strong\u003e(6):2100-2104.\u003c/li\u003e\n\u003cli\u003eFu L, Zhang Z, Zhang A, Xu J, Huang X, Zheng Q, Cao Y, Wang L, Du J: \u003cstrong\u003eAssociation Study Between FSHR Ala307Thr and Ser680Asn Variants and Polycystic Ovary Syndrome (PCOS) in Northern Chinese Han Women\u003c/strong\u003e. \u003cem\u003eJournal of assisted reproduction and genetics \u003c/em\u003e2013, \u003cstrong\u003e30\u003c/strong\u003e(5):717-721.\u003c/li\u003e\n\u003cli\u003eLi X, Li Y, Zhang B, Wang J, Yang Y, Du Y: \u003cstrong\u003ePlasma cytokine levels and PCOS risk: Mendelian randomization analysis reveals IL6R as a preventive factor\u003c/strong\u003e. \u003cem\u003eJournal of ovarian research \u003c/em\u003e2025, \u003cstrong\u003e18\u003c/strong\u003e(1):69.\u003c/li\u003e\n\u003cli\u003eQu J, Li B, Qiu M, Wang J, Chen Z, Li K, Teng X: \u003cstrong\u003eDiscovery of immune-related diagnostic biomarkers and construction of diagnostic model in varies polycystic ovary syndrome\u003c/strong\u003e. \u003cem\u003eArchives of gynecology and obstetrics \u003c/em\u003e2022, \u003cstrong\u003e306\u003c/strong\u003e(5):1607-1615.\u003c/li\u003e\n\u003cli\u003eBurgess S, Smith GD, Davies NM, Dudbridge F, Gill D, Glymour MM, Hartwig FP, Holmes MV, Minelli C, Relton CL\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eGuidelines for Performing Mendelian Randomization Investigations\u003c/strong\u003e. \u003cem\u003eWellcome Open Research \u003c/em\u003e2020, \u003cstrong\u003e4\u003c/strong\u003e:186.\u003c/li\u003e\n\u003cli\u003eZiegler A, Mwambi H, K\u0026ouml;nig IR: \u003cstrong\u003eMendelian Randomization Versus Path Models: Making Causal Inferences in Genetic Epidemiology\u003c/strong\u003e. \u003cem\u003eHuman Heredity \u003c/em\u003e2015, \u003cstrong\u003e79\u003c/strong\u003e(3-4):194-204.\u003c/li\u003e\n\u003cli\u003eSlob EAW, Burgess S: \u003cstrong\u003eA Comparison of Robust Mendelian Randomization Methods Using Summary Data\u003c/strong\u003e. \u003cem\u003eGenetic epidemiology \u003c/em\u003e2020, \u003cstrong\u003e44\u003c/strong\u003e(4):313-329.\u003c/li\u003e\n\u003cli\u003eMattei AL, Bailly N, Meissner A: \u003cstrong\u003eDNA methylation: a historical perspective\u003c/strong\u003e. \u003cem\u003eTrends in genetics : TIG \u003c/em\u003e2022, \u003cstrong\u003e38\u003c/strong\u003e(7):676-707.\u003c/li\u003e\n\u003cli\u003eNishiyama A, Nakanishi M: \u003cstrong\u003eNavigating the DNA methylation landscape of cancer\u003c/strong\u003e. \u003cem\u003eTrends in genetics : TIG \u003c/em\u003e2021, \u003cstrong\u003e37\u003c/strong\u003e(11):1012-1027.\u003c/li\u003e\n\u003cli\u003eGalbraith K, Snuderl M: \u003cstrong\u003eDNA methylation as a diagnostic tool\u003c/strong\u003e. \u003cem\u003eActa neuropathologica communications \u003c/em\u003e2022, \u003cstrong\u003e10\u003c/strong\u003e(1):71.\u003c/li\u003e\n\u003cli\u003eClark-Boucher D, Zhou X, Du J, Liu Y, Needham BL, Smith JA, Mukherjee B: \u003cstrong\u003eMethods for mediation analysis with high-dimensional DNA methylation data: Possible choices and comparisons\u003c/strong\u003e. \u003cem\u003ePLoS genetics \u003c/em\u003e2023, \u003cstrong\u003e19\u003c/strong\u003e(11):e1011022.\u003c/li\u003e\n\u003cli\u003eLv Y, Jia Z, Wang Y, Huang Y, Li C, Chen X, Xia W, Liu H, Xu S, Li Y: \u003cstrong\u003ePrenatal EDC exposure, DNA Methylation, and early childhood growth: A prospective birth cohort study\u003c/strong\u003e. \u003cem\u003eEnvironment international \u003c/em\u003e2024, \u003cstrong\u003e190\u003c/strong\u003e:108872.\u003c/li\u003e\n\u003cli\u003eKim AY, Park YJ, Pan X, Shin KC, Kwak S-H, Bassas AF, Sallam RM, Park KS, Alfadda AA, Xu A\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eObesity-Induced DNA Hypermethylation of the Adiponectin Gene Mediates Insulin Resistance\u003c/strong\u003e. \u003cem\u003eNature communications \u003c/em\u003e2015, \u003cstrong\u003e6\u003c/strong\u003e(1).\u003c/li\u003e\n\u003cli\u003eCastellano‐Castillo D, Moreno‐Indias I, S\u0026aacute;nchez‐Alcoholado L, Ramos‐Molina B, Alcaide‐Torres J, Morcillo S, Oca\u0026ntilde;a-Wilhelmi L, Tinahones FJ, Queipo‐Ortu\u0026ntilde;o MI, Cardona F: \u003cstrong\u003eAltered Adipose Tissue DNA Methylation Status in Metabolic Syndrome: Relationships Between Global DNA Methylation and Specific Methylation at Adipogenic, Lipid Metabolism and Inflammatory Candidate Genes and Metabolic Variables\u003c/strong\u003e. \u003cem\u003eJournal of clinical medicine \u003c/em\u003e2019, \u003cstrong\u003e8\u003c/strong\u003e(1):87.\u003c/li\u003e\n\u003cli\u003eVatier C, Christin-Maitre S: \u003cstrong\u003eEpigenetic/circadian clocks and PCOS\u003c/strong\u003e. \u003cem\u003eHuman reproduction (Oxford, England) \u003c/em\u003e2024, \u003cstrong\u003e39\u003c/strong\u003e(6):1167-1175.\u003c/li\u003e\n\u003cli\u003eMimouni NEH, Paiva I, Barbotin AL, Timzoura FE, Plassard D, Le Gras S, Ternier G, Pigny P, Catteau-Jonard S, Simon V\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003ePolycystic ovary syndrome is transmitted via a transgenerational epigenetic process\u003c/strong\u003e. \u003cem\u003eCell metabolism \u003c/em\u003e2021, \u003cstrong\u003e33\u003c/strong\u003e(3):513-530.e518.\u003c/li\u003e\n\u003cli\u003eFujii R, Sato S, Tsuboi Y, Cardenas A, Suzuki K: \u003cstrong\u003eDNA methylation as a mediator of associations between the environment and chronic diseases: A scoping review on application of mediation analysis\u003c/strong\u003e. \u003cem\u003eEpigenetics \u003c/em\u003e2022, \u003cstrong\u003e17\u003c/strong\u003e(7):759-785.\u003c/li\u003e\n\u003cli\u003eZhang K, Chen G, He J, Chen Z, Pan M, Tong J, Liu F, Xiang H: \u003cstrong\u003eDNA methylation mediates the effects of PM(2.5) and O(3) on ceramide metabolism: A novel mechanistic link between air pollution and insulin resistance\u003c/strong\u003e. \u003cem\u003eJournal of hazardous materials \u003c/em\u003e2024, \u003cstrong\u003e469\u003c/strong\u003e:133864.\u003c/li\u003e\n\u003cli\u003eWang J, Zhu QW, Mai JH, Zhang S, Wang Y, Liang J, Zhou JY: \u003cstrong\u003eA multi-omics study of brain tissue transcription and DNA methylation revealing the genetic pathogenesis of ADHD\u003c/strong\u003e. \u003cem\u003eBriefings in bioinformatics \u003c/em\u003e2024, \u003cstrong\u003e25\u003c/strong\u003e(6).\u003c/li\u003e\n\u003cli\u003eLuan YY, Zhang L, Peng YQ, Li YY, Liu RX, Yin CH: \u003cstrong\u003eImmune regulation in polycystic ovary syndrome\u003c/strong\u003e. \u003cem\u003eClinica chimica acta; international journal of clinical chemistry \u003c/em\u003e2022, \u003cstrong\u003e531\u003c/strong\u003e:265-272.\u003c/li\u003e\n\u003cli\u003eMancini A, Bruno C, Vergani E, d\u0026apos;Abate C, Giacchi E, Silvestrini A: \u003cstrong\u003eOxidative Stress and Low-Grade Inflammation in Polycystic Ovary Syndrome: Controversies and New Insights\u003c/strong\u003e. \u003cem\u003eInternational journal of molecular sciences \u003c/em\u003e2021, \u003cstrong\u003e22\u003c/strong\u003e(4).\u003c/li\u003e\n\u003cli\u003eOelen R, Vries DHd, Brugge H, Gordon GM, Vochteloo M, Ye C, Westra H-J, Franke L, Wijst MGPvd: \u003cstrong\u003eSingle-Cell RNA-sequencing Reveals Widespread Personalized, Context-Specific Gene Expression Regulation in Immune Cells\u003c/strong\u003e. 2021.\u003c/li\u003e\n\u003cli\u003eMa T, Li H, Zhang X: \u003cstrong\u003eDiscovering Single-Cell eQTLs From scRNA-seq Data Only\u003c/strong\u003e. 2021.\u003c/li\u003e\n\u003cli\u003eSun BB, Chiou J, Traylor M, Benner C, Hsu YH, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003ePlasma proteomic associations with genetics and health in the UK Biobank\u003c/strong\u003e. \u003cem\u003eNature \u003c/em\u003e2023, \u003cstrong\u003e622\u003c/strong\u003e(7982):329-338.\u003c/li\u003e\n\u003cli\u003eYazar S, Alquicira-Hernandez J, Wing K, Senabouth A, Gordon MG, Andersen S, Lu Q, Rowson A, Taylor TRP, Clarke L\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eSingle-cell eQTL mapping identifies cell type-specific genetic control of autoimmune disease\u003c/strong\u003e. \u003cem\u003eScience (New York, NY) \u003c/em\u003e2022, \u003cstrong\u003e376\u003c/strong\u003e(6589):eabf3041.\u003c/li\u003e\n\u003cli\u003eGkatzionis A, Burgess S, Newcombe PJ: \u003cstrong\u003eStatistical methods for cis-Mendelian randomization with two-sample summary-level data\u003c/strong\u003e. \u003cem\u003eGenetic epidemiology \u003c/em\u003e2023, \u003cstrong\u003e47\u003c/strong\u003e(1):3-25.\u003c/li\u003e\n\u003cli\u003ePierce BL, Ahsan H, Vanderweele TJ: \u003cstrong\u003ePower and instrument strength requirements for Mendelian randomization studies using multiple genetic variants\u003c/strong\u003e. \u003cem\u003eInternational journal of epidemiology \u003c/em\u003e2011, \u003cstrong\u003e40\u003c/strong\u003e(3):740-752.\u003c/li\u003e\n\u003cli\u003eZhu Z, Zhang F, Hu H, Bakshi A, Robinson MR, Powell JE, Montgomery GW, Goddard ME, Wray NR, Visscher PM\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eIntegration of summary data from GWAS and eQTL studies predicts complex trait gene targets\u003c/strong\u003e. \u003cem\u003eNature genetics \u003c/em\u003e2016, \u003cstrong\u003e48\u003c/strong\u003e(5):481-487.\u003c/li\u003e\n\u003cli\u003eZhu T, Goodarzi MO: \u003cstrong\u003eCauses and Consequences of Polycystic Ovary Syndrome: Insights From Mendelian Randomization\u003c/strong\u003e. \u003cem\u003eThe Journal of clinical endocrinology and metabolism \u003c/em\u003e2022, \u003cstrong\u003e107\u003c/strong\u003e(3):e899-e911.\u003c/li\u003e\n\u003cli\u003eWang C, Wu W, Yang H, Ye Z, Zhao Y, Liu J, Mu L: \u003cstrong\u003eMendelian randomization analyses for PCOS: evidence, opportunities, and challenges\u003c/strong\u003e. \u003cem\u003eTrends in genetics : TIG \u003c/em\u003e2022, \u003cstrong\u003e38\u003c/strong\u003e(5):468-482.\u003c/li\u003e\n\u003cli\u003eCupido AJ, Asselbergs FW, Natarajan P, Ridker PM, Hovingh GK, Schmidt AF: \u003cstrong\u003eDissecting the IL-6 pathway in cardiometabolic disease: A Mendelian randomization study on both IL6 and IL6R\u003c/strong\u003e. \u003cem\u003eBritish journal of clinical pharmacology \u003c/em\u003e2022, \u003cstrong\u003e88\u003c/strong\u003e(6):2875-2884.\u003c/li\u003e\n\u003cli\u003eHu A, She Y, Wang Y, Cao X, Li R, Wang J, Yu L, Jiang H, Zhao Y: \u003cstrong\u003eEffects of IL-6R Expression on IL-6/STAT3/HIF-1\u003c/strong\u003e\u003cstrong\u003e\u0026alpha; Signaling Pathway in a Mouse Model of Parkinson\u0026apos;s Disease with Type 2 Diabetes Co-Morbidity\u003c/strong\u003e. \u003cem\u003eDiscovery medicine \u003c/em\u003e2024, \u003cstrong\u003e36\u003c/strong\u003e(186):1386-1397.\u003c/li\u003e\n\u003cli\u003eBowker N, Shah RL, Sharp SJ, Luan J, Stewart ID, Wheeler E, Ferreira MAR, Baras A, Wareham NJ, Langenberg C\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eMeta-analysis investigating the role of interleukin-6 mediated inflammation in type 2 diabetes\u003c/strong\u003e. \u003cem\u003eEBioMedicine \u003c/em\u003e2020, \u003cstrong\u003e61\u003c/strong\u003e:103062.\u003c/li\u003e\n\u003cli\u003eZhang Y, Mei H, Xu K, Li C, Chang R, Qi H, Zhang Y, Zhang J: \u003cstrong\u003eAssociations between KCNQ1 and ITIH4 gene polymorphisms and infant weight gain in early life\u003c/strong\u003e. \u003cem\u003ePediatric research \u003c/em\u003e2022, \u003cstrong\u003e91\u003c/strong\u003e(5):1290-1295.\u003c/li\u003e\n\u003cli\u003eVoisin S, Alm\u0026eacute;n MS, Zheleznyakova GY, Lundberg L, Zarei S, Castillo S, Eriksson FE, Nilsson EK, Bl\u0026uuml;her M, B\u0026ouml;ttcher Y\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eMany obesity-associated SNPs strongly associate with DNA methylation changes at proximal promoters and enhancers\u003c/strong\u003e. \u003cem\u003eGenome medicine \u003c/em\u003e2015, \u003cstrong\u003e7\u003c/strong\u003e:103.\u003c/li\u003e\n\u003cli\u003eVan Dijck E, Diels S, Fransen E, Cremers TC, Verrijken A, Dirinck E, Hoischen A, Vandeweyer G, Vanden Berghe W, Van Gaal L\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eA Case-Control Study Supports Genetic Contribution of the PON Gene Family in Obesity and Metabolic Dysfunction Associated Steatotic Liver Disease\u003c/strong\u003e. \u003cem\u003eAntioxidants (Basel, Switzerland) \u003c/em\u003e2024, \u003cstrong\u003e13\u003c/strong\u003e(9).\u003c/li\u003e\n\u003cli\u003eShin GC, Lee HM, Kim N, Yoo SK, Park HS, Choi LS, Kim KP, Lee AR, Seo SU, Kim KH: \u003cstrong\u003eParaoxonase-2 contributes to promoting lipid metabolism and mitochondrial function via autophagy activation\u003c/strong\u003e. \u003cem\u003eScientific reports \u003c/em\u003e2022, \u003cstrong\u003e12\u003c/strong\u003e(1):21483.\u003c/li\u003e\n\u003cli\u003eHuang YH, Dong LP, Cui YG, Lu HY: \u003cstrong\u003eMiR-520h inhibits viability and facilitates apoptosis of KGN cells through modulating IL6R and the JAK/STAT pathway\u003c/strong\u003e. \u003cem\u003eReproductive biology \u003c/em\u003e2022, \u003cstrong\u003e22\u003c/strong\u003e(1):100607.\u003c/li\u003e\n\u003cli\u003eC\u0026aacute;rdenas AM, Ardila LJ, Vernal R, Melgar-Rodr\u0026iacute;guez S, Hern\u0026aacute;ndez HG: \u003cstrong\u003eBiomarkers of Periodontitis and Its Differential DNA Methylation and Gene Expression in Immune Cells: A Systematic Review\u003c/strong\u003e. \u003cem\u003eInternational journal of molecular sciences \u003c/em\u003e2022, \u003cstrong\u003e23\u003c/strong\u003e(19).\u003c/li\u003e\n\u003cli\u003eChen M, Pang L, Huang T, Liang W, Xian Q, Guo D, Zhong M, Song L, Huang Z, Liu Y\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eIntegrating single-cell RNA-seq, bulk RNA-seq, and Mendelian randomization to elucidate the role of HLA-DPA1 expression levels and non-classical monocytes in the pathogenesis of idiopathic pulmonary arterial hypertension\u003c/strong\u003e. \u003cem\u003eInternational journal of biological macromolecules \u003c/em\u003e2025, \u003cstrong\u003e319\u003c/strong\u003e(Pt 3):145284.\u003c/li\u003e\n\u003cli\u003eRay A, Alabarse P, Malik R, Sargurupremraj M, Bernhagen J, Dichgans M, Baumeister SE, Georgakis MK: \u003cstrong\u003eSingle-cell transcriptome-wide Mendelian randomization and colocalization analyses uncover cell-specific mechanisms in atherosclerotic cardiovascular disease\u003c/strong\u003e. \u003cem\u003eAmerican journal of human genetics \u003c/em\u003e2025.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Summary of SMR Analysis for Causal Genes Associated with Polycystic Ovary Syndrome\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"595\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGene\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003etopSNP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eA1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eA2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFreq\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ep_SMR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ep_HEIDI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOR（95%CI）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eIL6R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers4553185\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.54672\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.029226\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.8751668\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.8976 (1.0669-3.3751)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCRELD1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2270894\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.215706\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.001474\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.235592\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.1253 (1.3353\u0026ndash;3.3827)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eITIH4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12496077\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.264414\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.009342\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.5924614\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.1828 (1.0422\u0026ndash;1.3425)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePON2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers43038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.868787\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.049671\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.6078683\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.2981 (1.0004\u0026ndash;1.6845)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNSFL1C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers6033861\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.287276\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.005293\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.4884989\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.8133 (1.1934\u0026ndash;2.7550)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSNAP29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2072514\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.400596\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.039226\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.4003079\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.3472 (1.0148\u0026ndash;1.7885)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eLegend:\u0026nbsp;\u003cstrong\u003etopSNP:\u003c/strong\u003e Lead SNP for gene expression (eQTL); \u003cstrong\u003eA1 / A2:\u003c/strong\u003e Effect allele (A1) and reference allele (A2); \u003cstrong\u003eFreq:\u003c/strong\u003e Frequency of the effect allele (A1); \u003cstrong\u003ep_SMR:\u003c/strong\u003e P value for the SMR test assessing the causal effect of gene expression on PCOS; \u003cstrong\u003ep_HEIDI:\u003c/strong\u003e P value for the HEIDI test evaluating potential heterogeneity (p \u0026gt; 0.05 suggests no heterogeneity); \u003cstrong\u003eOR (95% CI):\u003c/strong\u003e Odds ratio and 95% confidence interval for the causal effect.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003cstrong\u003e. Mediation Mendelian Randomization Results for CpG Sites Regulating IL6R and SNAP29\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eGene\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eCpG_ID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003ebeta1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003ebeta2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003ebeta_all\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003ebeta12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 74px;\"\u003e\n \u003cp\u003ebeta12_p\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eP\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eIL6R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003ecg09257526\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e-0.2958\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.4595\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.2994\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.1359\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0.4541\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e0.1366\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eIL6R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003ecg13549904\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e-0.2388\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.4595\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.1556\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.1097\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0.7052\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e0.2288\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eIL6R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003ecg20688791\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e-0.4025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.4595\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.3941\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.185\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0.4693\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e0.0501\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eIL6R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003ecg21262032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e-0.1097\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.4595\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.1364\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.0504\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0.3694\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e0.5775\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eSNAP29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003ecg00335892\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e-0.4893\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.0963\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.1026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.0471\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0.4596\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e0.6352\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 72px;\"\u003e\n \u003cp\u003eSNAP29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003ecg20180721\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e-0.3901\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.0963\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.0911\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e-0.0376\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0.4125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e0.6008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eLegend:\u003cstrong\u003e\u0026nbsp;beta1:\u003c/strong\u003e Effect size representing the causal effect of DNA methylation on gene expression (methylation \u0026rarr; gene expression); \u003cstrong\u003ebeta2:\u0026nbsp;\u003c/strong\u003eEffect size representing the causal effect of gene expression on PCOS risk (gene expression \u0026rarr; PCOS); \u003cstrong\u003ebeta_all:\u0026nbsp;\u003c/strong\u003eTotal causal effect of DNA methylation on PCOS risk, including both direct and indirect effects;\u0026nbsp;\u003cstrong\u003ebeta12\u003c/strong\u003e: Estimated mediation effect representing the indirect effect of DNA methylation on PCOS risk through gene expression (methylation \u0026rarr; gene expression \u0026rarr; PCOS);\u0026nbsp;\u003cstrong\u003eMediation proportion (beta12_p)\u003c/strong\u003e: Proportion of the total effect mediated by gene expression, calculated as the ratio of \u0026beta;₁₂ to \u0026beta;_all.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Polycystic ovary syndrome, Mendelian randomization, Multi-omics integration, DNA methylation, Single-cell eQTLs","lastPublishedDoi":"10.21203/rs.3.rs-7532908/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7532908/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003ePolycystic ovary syndrome (PCOS) is a common endocrine-metabolic disorder, yet its genetic basis remains incompletely understood. This study aimed to identify causal genes and elucidate upstream epigenetic and immune cell\u0026ndash;specific regulatory mechanisms using a multi-omics Mendelian randomization (MR) framework.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eWe performed two-sample MR analyses using expression quantitative trait loci (eQTLs) from the eQTLGen consortium and protein QTLs (pQTLs) from the UK Biobank to assess the causal effects of gene expression and protein levels on PCOS risk. Genes significant in both datasets were retained as candidate genes and further evaluated using summary-data-based Mendelian randomization (SMR) with GTEx whole-blood eQTLs to determine colocalized genetic signals. To investigate upstream regulation, we conducted mediation MR analysis using methylation QTLs (mQTLs) from the GoDMC database to identify CpG sites potentially mediating gene expression and PCOS risk. Finally, we performed cell-type\u0026ndash;specific MR using single-cell eQTLs (sc-eQTLs) from the OneK1K project across 14 immune cell types.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eMR identified 1,715 eQTL- and 182 pQTL-associated genes, with 60 overlapping candidates. SMR prioritized six causal genes: CRELD1, NSFL1C, ITIH4, IL6R, SNAP29, and PON2. Mediation MR revealed a borderline-significant effect for cg20688791 upstream of \u003cem\u003eIL6R\u003c/em\u003e and suggestive mediation at cg00335892 within \u003cem\u003eSNAP29\u003c/em\u003e. sc-eQTL analysis showed that \u003cem\u003eCRELD1\u003c/em\u003e, \u003cem\u003eITIH4\u003c/em\u003e, \u003cem\u003ePON2\u003c/em\u003e, and \u003cem\u003eSNAP29\u003c/em\u003e had significant causal effects in CD8⁺ T cells, CD4⁺ T cells, monocytes, and NK cells, respectively.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003eThis integrative analysis identifies multi-omics-supported causal genes for PCOS and reveals epigenetic and immune cell\u0026ndash;specific regulatory mechanisms, offering novel insights into pathogenesis and potential therapeutic targets.\u003c/p\u003e","manuscriptTitle":"Causal Genes and Immune-Epigenetic Mechanisms Underlying Polycystic Ovary Syndrome: A Multi-Omics Mendelian Randomization Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-30 13:26:01","doi":"10.21203/rs.3.rs-7532908/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8d6de1b3-ff99-4704-b38c-32506034618a","owner":[],"postedDate":"September 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-03T11:13:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-30 13:26:01","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7532908","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7532908","identity":"rs-7532908","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}