Genome-wide association analysis of endometriosis in Indian women supports shared genetic susceptibility and suggestive population specific signals

Endometriosis is an estrogen-dependent, chronic gynaecological disorder affecting approximately 10% of women of reproductive age and is associated with pelvic pain, infertility, and reduced quality of life. Genetic studies indicate substantial heritability, yet genome-wide association studies (GWAS) have predominantly focused on European populations, with limited evidence from South Asian groups. We conducted a GWAS in 2523 participants from a nationwide, multicentre endometriosis case–control study in India. Genetic sus - ceptibility in Indian women showed both shared features with other populations and evidence of potential population-specific signals. Twenty-one loci demonstrated suggestive associations ( p < 1 × 10⁻5). The strongest signal mapped to chromosome 13 within a long non-coding RNA and in regulatory elements upstream of SHISA2. Colocalization analysis showed shared genetic effects with SHISA2 expression in thyroid tissue and methylation sites in the endometrium. A polygenic risk score derived from European populations demonstrated comparable predictive performance in the Indian cohort. Meta-analysis with a Japanese GWAS replicated two known genome-wide significant associations near WNT4 and CDKN2B-AS1, supporting shared genetic suscep - tibility across populations. No variant reached genome-wide significance in the discovery cohort alone. This study highlights the importance of expanding genetic studies to diverse populations to improve generalisable genetic risk estimates, biological understanding and equitable precision medicine in women’s health. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9

Endometriosis · Genome-wide association study · South Asian population · Polygenic risk score · Meta-analysis · India

Endometriosis is an estrogen-dependent, chronic gynecological disorder characterized by ectopic growth of endometrium-like tissue. It significantly impacts a woman’s overall well-being and quality of life, and is associ- ated with chronic pain, menstrual disturbances, infertility, and psychological distress. 1,2 Endometriosis affects about 10% of women of reproductive age globally; India accounts for an estimated 50 million cases, representing nearly 18% of the global burden.3–5 Increasingly, endometriosis is recognised as a condition with impacts beyond reproductive health, affecting physical, psychological, and social well-being.6,7 Endometriosis arises from a complex interplay of hormonal, environmental, and genetic factors. Several theo- ries have been proposed to explain the disease pathogenesis, including retrograde menstruation, coelomic meta - plasia, lymphatics and vascular metastasis, autoimmunity, inflammation, and genetic and epigenetic factors, as well as exposure to a multitude of dietary, environmental, and lifestyle factors.1 Despite decades of research, the underlying causes and associated risk factors remain incompletely understood, contributing to diagnostic delays and treatment inadequacies. Genetic studies provide strong evidence for a substantial heritable component to the disease, supported by familial clustering, twin studies, and SNP-based heritability estimates. 8–11 GWAS have identified over 40 loci in European and East Asian populations (including Japanese, Taiwanese, and Chinese), implicating pathways related to hormonal regulation, cell adhesion, inflammation, and pain.12–15 However, these studies explain only a small proportion of disease variance, and the predominance of European cohorts limits understanding of ances - try-specific or shared genetic architecture. Differences in allele-frequency and linkage disequilibrium (LD) patterns across populations have important implications for the discovery and interpretation of genetic associations. Variants that are rare or absent in Euro- pean populations may be more common in South Asian groups, while ancestry-specific LD patterns can improve fine-mapping resolution and help distinguish causal variants. These differences contribute to reduced replication of association signals across populations and limit the transferability of PRS derived from European datasets. 16 Inclusion of diverse populations is therefore essential to improve locus discovery, refine causal variants, and enable equitable risk prediction. Despite accounting for a substantial proportion of the global burden of endometriosis, Indian populations remain underrepresented in genetic studies. The Endometriosis Clinical and Genetic Research in India (ECGRI) study was designed to address this gap by investigating the genetic susceptibility in Indian women. 17 In this nationwide case–control study, we conducted a genome-wide association analysis of surgically confirmed endo- metriosis in 1288 cases and 1235 ancestry-matched controls. We further performed fine-mapping and functional annotation analyses, assessed the replication of previously reported loci, evaluated the transferability of a Euro - pean-derived polygenic risk score, and conducted a cross-ancestry meta-analysis using Japanese GWAS data. These analyses aim to characterise shared and putative population-specific genetic risk and to improve under - standing of the genetic architecture of endometriosis in an underrepresented population.

Participant characteristics A total of 2771 participants (1383 cases and 1388 controls) were recruited. The mean age was significantly higher in cases than in controls (32·69 ± 6·77 vs 31·58 ± 6·68; p = 1·37 × 10–5). Participants were recruited across five Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 geographic regions with representation from Indo-European, Dravidian and Tibeto-Burman ancestry groups. The distribution of cases and controls was broadly comparable across ancestry groups (Table 1 and Supplementary Table S1). Quality control and imputation Following genotyping, quality control retained 2,523 individuals (1288 cases and 1235 controls) and 416,262 variants were retained for analysis. Imputation using the SEAD reference resulted in 9,647,516 variants after filtering, including 8,073,999 bi-allelic SNPs. Concordance between SEAD and TOPMed imputation panels was high (> 0·97), supporting the robustness of imputation. Population structure Principal component analysis of genome-wide genotype data showed that study participants clustered with the South Asian population in the 1000 Genomes reference panel, with limited spread extending towards the East Asian clusters and PCA of our cohort alone showed that clustering by ethnicity was consistent with ancestry. No evidence of clustering by genotyping batch was observed (Supplementary Figure S2). Table 1 Demographics of the ECGRI study participants. Characteristics Case Control Total Participants (recruited) 1383 1388 2771 Mean Age ± SD* (years) 32·69 ± 6·77 31·58 ± 6·68 NA Zone Central 391 (28·27) 391 (28·17) 782 (28·22) North 337 (24·37) 354 (25·5) 691 (24·94) East/North East 297 (21·48) 287 (20·68) 584 (21·08) South 138 (9·98) 138 (9·94) 276 (9·96) West 220 (15·91) 218 (15·71) 438 (15·81) Ethnicity Indo-European 1072 (77·51) 1080 (77·81) 2152 (77·66) Dravidian 165 (11·93) 176 (12·68) 341 (12·31) Tibeto-Burmese 146 (10·56) 132 (9·51) 278 (10·03) Endometriosis stage** I/II 201 (14·64) NA NA III/IV 1172 (85·36) NA NA Genotyped participants Case Control Total Participants 1288 1235 2523 Mean age ± SD 32·8 ± 6·75 31·7 ± 6·67 NA Zone Central 364 (28·26) 345 (27·94) 709 (28·1) North 318 (24·69) 320 (25·91) 638 (25·29) East/North East 268 (20·81) 249 (20·16) 517 (20·49) South 133 (10·33) 126 (10·2) 259 (10·27) West 205 (15·92) 195 (15·79) 400 (15·85) Ethnicity Indo-European 995 (77·25) 957 (77·49) 1952 (77·37) Dravidian 158 (12·27) 160 (12·96) 318 (12·6) Tibeto-Burmese 135 (10·48) 118 (9·55) 253 (10·03) Endometriosis stage*** I/II 185 (14·46) NA NA III/IV 1094 (85·54) NA NA *Age missing for 3 cases **Stage data missing for 10 cases; proportion calculated as % out of 1373 ***Stage data missing for 9 cases, proportion calculated as % out of 1279 Proportion calculated as % of cases/controls/total respectively NA, not applicable. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Genetic associations with endometriosis Genome-wide association analysis identified 88 variants associated with endometriosis at a suggestive signifi - cance threshold (p < 1 × 1 0–5), corresponding to 21 genomic risk loci (Fig. 1, Supplementary Table S2). No variant reached genome-wide significance ( p < 1 × 1 0–8). The test statistic showed minimal inflation (λ = 1·03) and the quantile–quantile plots indicated no evidence of systematic bias, though mild deflation of test statistics noted was likely due to small sample size (Supplementary Figure S3). Using the suggestive p-value threshold, FUMA identified 21 genomic risk loci with 24 independent significant lead SNPs (Table 2). All potential 21 risk loci were novel and had not been identified in previous endometriosis GWAS. Analysis with the linear mixed model in GCTA gave consistent results, despite the differences in the statistical models and variance assumptions. There was additional deflation of the test statistics with GCTA due to inclusion of the GRM in the model, indicating that the results are robust to population stratification (Supplementary Figure S4). Replication of risk loci across different ancestral populations Of the 21 genomic risk loci identified, 3 showed allele frequency differences > 0·1 across major Indian ancestry groups, and 16 showed differences across global populations (Supplementary Figure S5; Supplementary Table S3—S4). Of the 46 autosomal risk SNPs identified in the European GWAS dataset 13, 38 were present in our dataset and proxy SNPs in strong LD (R 2 > 0·75) were identified for the remaining eight using LD proxy. 18 Of these eight variants, rs2040445 was not polymorphic in the South Asian population. A total of seven variants rep- licated with nominal significance (p < 0·05), with similar direction of effect in our dataset (Supplementary Table S5). Replicated variants included the variants at the WNT4 and DNM3 loci on Chr 1, the KDR locus on Chr 4, the 7p15·2 locus on Chr 7, the CDKN2B-AS1 locus on Chr 9, the RNLS locus on Chr 10, and the VEZT locus on Chr 12. Figure 2 shows the differences in the allele frequencies of the 45 risk SNPs, including the proxy SNPs, between the European and the Indian population. Differences in allele frequency of > 0·1 were observed in 19 out of 45 variants (Supplementary Table S6). We replicated one lead SNP from the Taiwanese dataset at WNT4 locus on Chr 1 with consistent direction of effect and p < 0·05 (Supplementary Table S5). Out of the 65 significantly associated SNPs from the multiancestry meta-anlyses 19, five replicated in our dataset with p < 0.05. One SNP Fig. 1 Manhattan plot for genome-wide associations for endometriosis in the Indian population. Manhattan plot shows the genome wide association results for endometriosis using 1288 cases and 1235 controls in the ECGRI cohort. The x-axis shows the chromosomal position and y-axis shows the -log 10Pvalue. Colours denote alternative chromosomes. The red hori - zontal line represents the genome-wide significant threshold (p < 5 × 10–8) and the blue horizontal line represents the nominal genome-wide significance (p < 1 × 1 0–5). The lead variant was rs35236165 on chromosome 13 (p = 8·43 × 10–7). Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Table 2 Summary of the GWAS results for 21 genomic risk loci for endometriosis identified by FUMA. Chr SNP Position (bp) RA RAF OR LOG(OR)SE P-value Nearest gene/cytoband 1 rs872595 38,705,915 C 0·26 1·37 0·07 3·46E-06 LINC01343 1 rs145379000 84,911,878 C 0·94 1·75 0·13 8·67E-06 RPF1, DNASE2B 1 rs114511502 231,326,592 C 0·96 2·13 0·16 1·18E-06 TRIM67 1 rs61829563 238,400,886 A 0·91 1·67 0·11 1·63E-06 RP11-136B18·1 2 rs6710746 10,414,723 G 0·44 1·33 0·06 1·32E-06 RN7SL66P 2 rs1861768 15,625,186 C 0·50 1·29 0·06 9·86E-06 NBAS 2 rs2679859 105,927,611 C 0·25 1·35 0·07 5·85E-06 TGFBRAP1 2 rs7563305 113,673,470 C 0·87 1·49 0·09 8·74E-06 IL37 3 rs12631821 65,033,653 C 0·85 1·45 0·08 5·99E-06 ADAMTS9 3 rs1005940 74,221,493 A 0·73 1·37 0·07 3·16E-06 Y-RNA, CNTN3 5 rs58794955 10,571,893 T 0·23 1·38 0·07 4·69E-06 ANKRD33B 6 rs9401193 120,284,989 A 0·94 1·72 0·12 9·65E-06 MIR3144, RNU6 7 rs56991723 73,735,335 G 0·85 1·47 0·09 8·87E-06 CLIP2 10 rs11017546 132,686,453 T 0·12 1·58 0·09 1·19E-06 MIR378C 11 rs1599469 98,783,039 A 0·57 1·30 0·06 9·60E-06 RN7SKP53 12 rs76975203 132,854,297 C 0·95 1·92 0·14 3·12E-06 GALNT9 13 rs35236165 26,627,563 A 0·41 1·34 0·06 8·43E-07 LINC00415 13 rs12867495 29,842,523 T 0·05 1·95 0·15 5·50E-06 MTUS2 15 rs28698234 88,268,228 T 0·94 1·79 0·13 5·87E-06 NTRK3, LINC00052 16 rs140109794 23,168,249 A 0·07 1·72 0·12 7·37E-06 USP31 21 rs2838083 43,063,044 A 0·63 1·30 0·06 7·46E-06 LINC00111, SNOU13 Chr, chromosome; SNP, single nucleotide polymorphism; RA, risk allele; RAF, risk allele frequency; OR, odds ratio with respect to the risk allele; SE, standard error of log odds ratio. Genomic positions are based on GRCh37 (hg19) Fig. 2 Allele frequencies of the 45 known endometriosis risk variants. The graph shows the risk allele frequency of the 45 previously established risk variants for endometriosis in the European population (blue) and in the Indian dataset (ECGRI) (red). The x-axis shows the lead variants annotated with chromosome and the closest gene, and the y-axis shows the risk allele frequency. Dots joined by a grey line represent allele frequencies of the same variant. The variants on the x-axis in red colour represent the proxy SNPs. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 (rs2779747 at CDKN2B-AS1/DMRTA1 locus, p = 5.56 × 10⁻4) passed the Bonferroni-corrected multiple testing thresholds of p < 7.69 × 10–4 (0.05/65) (Supplementary Table S5). These results indicate partial overlap in genetic risk across different populations, highlighting both shared and potentially population specific genetic architecture. Evidence for potential causal variants at chr13q12·13/LINC00415 The strongest signal was located on chromosome 13 (rs35236165; p = 8·43 × 10–7). Functional annotation analy- sis in FUMA showed the lead variant rs35236165 lies within an exon of the long non-coding RNA LINC00415 and is in strong LD with an enhancer variant upstream of the gene SHISA2 (Supplementary Figure S6). A 95% credible set of 65 SNPs was identified within 1 Mb region surrounding rs35236165 (Supplementary Table S7). The lead SNP from our GWAS data showed the highest posterior inclusion probability (PIP) score of 0·386 indicating moderate causality. Combined with the PIP score, high LD in this region (68% of the SNP pairs had r2 > 0·8) suggests a strong but poorly localized association signal. Potential target genes Mapping credible-set variants to epigenetic marks identified 23 variants overlapping enhancer-associated histone marks and one variant (rs3088377) overlapping a promoter-associated histone mark at SHISA2 (Additional Infor- mation 1, Supplementary Figure S7), indicating that a considerable fraction of the credible set lies within or close to regulatory elements, supporting regulatory mechanisms at this locus. Colocalisation (COLOC) and summary-data-based Mendelian randomisation (SMR) analyses were used to evaluate whether endometriosis GWAS loci and molecular quantitative trait loci (QTLs) were driven by shared variants, providing evidence for potential functional genes and regulatory mechanisms underlying disease risk. Colocalisation analyses indicated a shared causal signal between endometriosis risk and SHISA2 expression in thyroid (PPH4 = 0·64 > PPH3 = 0·26), while blood expression QTL (eQTL) data suggested independent causal variants (PPH3 = 0·73) (Table 3). Strong evidence of colocalisation was also observed between endometriosis and DNA methylation at cg19014360 (PPH4 = 0·92) and cg07984168 (PPH4 = 0·91) in endometrium and at cg07984168 in blood (PPH4 = 0·94) (Fig. 3). Both CpG sites are located within the non-coding RNA LINC00415, with cg07984168 overlapping a predicted enhancer. Table 3 Colocalization of GWAS signals with eQTLs and mQTLs. QTL Dataset Gene/ CpG site nsnps PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf Blood_eQTL RNF6 1366 2·08E-07 5·97E-07 2·54E-01 7·30E-01 1·64E-02 Blood_eQTL SHISA2 1366 4·32E-64 1·24E-63 2·55E-01 7·35E-01 9·84E-03 Thyroid_eQTL SHISA2 1284 8·57E-16 2·45E-15 9·26E-02 2·64E-01 6·44E-01 Blood_mQTL cg07984168 634 5·92E-119 1·64E-118 1·64E-02 4·44E-02 9·39E-01 Blood_mQTL cg27012920 565 5·50E-153 8·33E-154 2·77E-01 4·12E-02 6·82E-01 Endometrium_mQTL cg01600817 161 3·37E-07 1·75E-08 3·72E-01 1·87E-02 6·09E-01 Endometrium_mQTL cg05581851 212 5·51E-15 2·08E-15 3·25E-01 1·22E-01 5·53E-01 Endometrium_mQTL cg07984168 369 8·99E-58 2·45E-57 2·53E-02 6·80E-02 9·07E-01 Endometrium_mQTL cg15020125 395 5·10E-51 1·40E-50 3·68E-02 1·00E-01 8·63E-01 Endometrium_mQTL cg19014360 360 1·07E-19 2·95E-19 2·17E-02 5·88E-02 9·20E-01 Endometrium_mQTL cg27012920 256 5·56E-19 3·50E-20 3·69E-01 2·27E-02 6·08E-01 nsnps—number of SNPs in region used in analysis PP.H0.abf—posterior probability of hypothesis 0: neither trait has a genetic association in the region PP.H1.abf—posterior probability of hypothesis 1: only trait 1 has a genetic association in the region PP.H2.abf—posterior probability of hypothesis 2: only trait 2 has a genetic association in the region PP.H3.abf—posterior probability of hypothesis 3: both traits are associated, but with different causal variants PP.H4.abf—posterior probability of hypothesis 4: both traits are associated and share a single causal variant Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Fig. 3 Integrated locus plot at rs35236165 showing GWAS, eQTL and mQTL signals. Overlap of asso- ciation peaks is observed across (A) GWAS SNP plot with variants colored according to LD (r2) with the lead vari- ant rs35236165, (B) eQTL45 associa- tion signal for SHISA2 in thyroid and mQTL46,47 association signal for CpG site cg07984168 in blood (C) and endometrium (D). The x-axis shows the genomic coordinates in hg38 build and y-axis shows the -log10(P) values for associations. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Although no associations reached significance after multiple-testing correction in SMR analyses of gene expression, SHISA2 was among the top signals in thyroid (p_SMR = 3·78 × 10⁻4; p_HEIDI = 0·068). Similarly, SMR analysis using DNA methylation QTL (mQTL) data, the association between rs7321426 and rs4770916 at this locus and methylation at cg07984168 in LINC00415 was the most significant signal, in endometrium (p_ SMR = 7·27 × 10–5, p_HEIDI = 5·81 × 10–2) and in blood (p_SMR = 6·26 × 10–5, p_HEIDI = 0·21), respectively (Supplementary Table S8). All three variants were part of the credible set. Genetic liability is shared across populations To evaluate the extent to which genetic susceptibility identified in European populations is shared in our cohort, we calculated PRSs and tested their association with endometriosis status and disease severity. The mean PRS was higher in cases than controls (p < 2·2 × 10–16) reflected by a shift in the distribution of PRS in cases towards the right in the density plot (Fig. 4a). The AUC of 0·61 was comparable to that estimated in the UK Biobank (0·61) and in the Estonian Biobank (0·60) 20. Individuals in the tenth PRS decile had almost 4 times higher risk of endometriosis compared to the individuals in the first decile (Fig. 4b). The mean PRS was significantly higher in individuals with stage III and stage IV endometriosis compared to individuals with stage I and stage II endo - metriosis (p = 3·99 × 10–4) (Fig. 4c). Meta-analyses To increase power, reduce reliance on European LD structure and improve the identification of loci robust across ancestries, we meta-analysed of our dataset with the Japanese GWAS 21 and identified 31 genome wide signifi - cant variants ( p < 1 × 1 0–8) across two loci on chromosomes 1 and 9 (Supplementary Table S9 & Fig. 5). Test statistics showed no inflation (λ = 1·007; Supplementary Figure S8). The chromosome 1 lead SNP (rs7412010, p = 3·38 × 10–12) was in high LD (r 2 = 0·98) with rs10917151 reported in the European GWAS and the chromo - some 9 lead SNP (rs7041895, p = 1·19 × 10–8) was in high LD (r 2 = 0·91) with rs10122243 reported in the Euro - pean population. Different lead variants emerged in the Japanese and Indian datasets at both loci, yet all were in strong LD with the known European signals, demonstrating that the meta-analysis signal originates from a com- mon association peak across populations rather than one cohort dominating the result. Functional annotation using FUMA identified rs7412010 within the enhancer region upstream of WNT4 and CDC42 with WNT4 being the nearest gene. rs7041895 mapped to the lncRNA ENSG00000298449, located upstream of the CDKN2A/B locus, with CDKN2B-AS1 as the closest gene. Fig. 4 Endometriosis Polygenic Risk Scores. (A) Density plot of PRS in 1288 cases and 1235 controls in the Indian popula- tion. (B) PRS decile plot showing OR (95% CI, y-axis) of endometriosis, by endometriosis PRS decile (x-axis) compared to the first decile. (C) Box plot of mean PRS in the endometriosis cases in Stage I/II and Stage III/IV. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Colocalization analyses showed both shared and distinct causal variants across the chromosome 1 and chro - mosome 9 loci (Supplementary Table S10). At the chromosome 1 locus, endometriosis signals colocalised with CDC42 expression in blood (PPH4 = 0·78) and thyroid (PPH4 = 0·61), and were independent (PPH3 = 0·92) in endometrium. For WNT4, high posterior support for distinct signals was observed in thyroid and endometrium (PPH3 = 0·99). At the chromosome 9 locus, colocalization results supported independent causal variants for endometriosis and expression of both CDKN2B and CDKN2A across blood and pituitary tissues (PPH3 ≥ 0·95). Strong shared causal signals were also observed between endometriosis and methylation at several CpG sites near WNT4, CDC42, in blood and endometrium and CDKN2A in endometrium (PPH4 ≥ 0·95). SMR analyses further identified three expression-mediated and four methylation-mediated associations, including a significant effect of CDC42 expression in blood and multiple methylation sites at the chromosome 1 locus that remained significant after multiple-testing correction (Supplementary Table S11). Several of these methylation associations were replicated across both endometrial and blood mQTL datasets, located in proximal enhancers upstream of WNT4, highlighting consistent regulatory mechanisms at these loci.

This nationwide study provides evidence that genetic susceptibility to endometriosis in Indian women shares key features with other populations, while also suggesting potential population-specific signals. Although no variant reached genome-wide significance in the discovery cohort alone, we detected 21 loci with suggestive evidence of association and replicated seven previously reported susceptibility regions. The novel suggestive variant with the strongest evidence of association was located near SHISA2, a strong candidate target gene. A cross-ancestry meta- analysis using a Japanese dataset identified genome-wide significant signals at established loci on chromosomes 1 (near WNT4) and 9 (near CDKN2B-AS1), reinforcing shared genetic architecture across Asian populations. In addition, a European-derived polygenic risk score demonstrated comparable predictive performance in the Indian cohort and was associated with disease severity. The absence of genome-wide significant findings in the discovery analysis likely reflects limited statistical power compared to large European datasets. Nevertheless, nominal replication of established loci, including Fig. 5 Manhattan plot for genome wide associations for endometriosis in the Asian meta-analysis. Manhattan plot shows the genome wide association results for endometriosis in 3074 cases and 82,210 controls included in the meta-analysis. The x-axis shows the chromosomal position, and y-axis shows the association -log10Pvalue. The red horizontal line represents the genome-wide significant threshold (p < 5 × 10–8) and the blue horizontal line represents the suggestive genome-wide signifi- cance (p < 1 × 1 0–5). The lead variant at Chr 1 was rs7412010, (p = 3·38 × 10–12) and at Chr 9 was rs7041895 (p = 1·19 × 10–8). Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 regions near WNT4 and CDKN2B-AS1, supports cross-population consistency of key susceptibility signals. Inter- estingly, the PRS derived from the European population demonstrated cross-ancestry transferability, with com - parable AUCs in Indian and European cohorts. PRS increased progressively with disease severity, reflecting patterns observed in European studies 20 and supporting shared genetic architecture across ancestries. However, only seven previously reported risk loci were replicated in our dataset, likely reflecting limited statistical power and differences in allele frequencies across ancestries. Similarly, the 21 potential risk loci identified in this study have not been observed in previous GWAS; while some may represent false positives due to limited power, oth- ers may reflect true ancestry-specific, or previously undetected signals influenced by allele frequency differences and cohort characteristics. Notably, the ECGRI study screened controls to exclude individuals with suspected endometriosis or a family history of the condition prior to enrolment, reducing the likelihood of misclassification. The lead variant on chromosome 13 rs35236165 was located within an exon of the long non-coding RNA LINC00415. This was in high LD with variants in enhancer and promoter regions upstream to SHISA2. SHISA2 is reported to negatively regulate WNT signaling and Fibroblast Growth Factor receptor pathways. 22–24 Both these pathways are reported to be biologically relevant to endometrial function and disease.25–28 SHISA2 has also been previously shown to be nominally upregulated during the mid secretory phase in women with endometriosis compared to controls.29 The allele frequency of rs35236165 was different between the European and South Asian population, which together with differences in control ascertainment may explain why this locus has not previ - ously been identified in earlier GWAS studies that were predominantly based on European populations. Using COLOC we identified evidence of shared genetic effects between endometriosis risk at this locus and SHISA2 expression in thyroid derived from eQTL data. We also found strong evidence of a shared causal variant between endometriosis risk and methylation at two CpG sites (cg19014360 and cg07984168) within LINC00415 in endo- metrial mQTL datasets, with cg07984168 also showing colocalization in blood and mapping to an enhancer region. Although the SMR analyses integrating GWAS with eQTL and mQTL datasets did not remain significant after correction for multiple testing, SHISA2 and cg07984168 consistently appeared among the top signals in the thyroid eQTL dataset and in endometrial and blood mQTL datasets, driven by variants in high LD with the lead GWAS variants (r2 > 0·77). Our data suggests SHISA2 as a candidate gene supported by regulatory annotation, colocalisation, and methylation evidence and warrants further detailed investigation as an endometriosis suscep- tibility gene. Clinical studies have previously established association between thyroid dysfunction and endometriosis. 30–32 However, our study provides the first genetic evidence that may implicate thyroid mediated regulatory mecha - nisms in endometriosis susceptibility. Although South Asian populations are genetically distinct from both European and East Asian groups, we con- ducted a non-European meta-analysis to minimise the influence of European LD structure and enhance the iden- tification and fine-mapping of ancestry-robust loci. Differences in LD patterns between South and East Asians can further improve the resolution of association signals. Our cross ancestry meta-analysis identified 31 genome wide significant variants mapping to two loci on Chr 1 and Chr 9, both having been previously reported in Japa- nese and European GWAS studies, further strengthening cross-ancestry support for these associations. The lead variant rs7412010 at Chr 1 lies in an enhancer region upstream of WNT4 and CDC42. The lead variant rs7041895 at Chr 9 lies within a long non-coding RNA upstream of CDKN2A. Variants within the chromosome 1 locus co-localized with eQTL signals for CDC42 in blood and thyroid tissues, with distinct but correlated regulatory signals noted for WNT4 in thyroid and endometrial eQTL datasets. COLOC analysis using endometrium and blood derived mQTL datasets provided strong evidence for shared causal variants at methylation sites mapping to WNT4 and CDC42, with additional support for CDKN2A in endometrial mQTL dataset. Corroborating these findings, SMR analysis identified statistically significant association for CDC42 in blood eQTL data (driven by rs2473290), which showed moderate LD with the lead variant (r 2 > 0·49) across European, East Asian and South Asian populations. We also observed a suggestive signal for LINC00339 in both blood and endometrium, driven by rs11801382, a variant with high D’ (> 0·9) but low r2 (< 0·1) with the lead SNP, reflecting its low allele frequency, particularly in Asian populations and the use of European eQTL datasets where the variant is more Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 common. In addition, SMR supported shared causal variants (r 2 > 0·8 with lead variant rs7412010) for multiple methylation sites across the WNT4 locus in both endometrial and blood tissues. Both WNT4 and CDKN2A loci are well established endometriosis risk loci in prior GWAS and functional studies, supporting roles in hormonal sig- nalling, cellular proliferation, and endometrial tissue regulation. 33 Together, these results reinforce the presence of shared, functionally relevant genetic mechanisms across South Asian, East Asian, and European populations, and suggest that cross-ancestry risk at these loci may converge on endocrine and regulatory pathways central to endometriosis biology. This study has several limitations. The sample size, although substantial for a single-country study, is mod - est compared with large international GWAS, limiting the power to detect variants with small effect sizes. Dif - ferences in allele frequencies and LD structure across ancestries reduce direct comparability with previously conducted GWAS studies, and the limited availability of South Asian-specific reference panels necessitated addi- tional caution and rigorous quality control. To address these challenges, we used multiple approaches, includ - ing multiple imputation panels and different GWAS models, to ensure the robustness of our findings. Another important limitation is the absence of a large, independent South Asian replication cohort for endometriosis, which restricts our ability to validate the observed association signals in an ancestry-matched population. There- fore, the 21 suggestive associations identified in the Indian discovery GWAS should be considered hypothesis- generating and require independent replication before being regarded as confirmed endometriosis susceptibility loci. Although meta-analysis with the Japanese cohort increased power to detect shared ancestral risk loci, it is important to recognise the inherent limitations arising from differences in study design and phenotype defini - tions between cohorts. Variability in case-ascertainment methods, diagnostic criteria, and control selection, or sample size, can introduce heterogeneity and may influence the strength or detectability of association signals. Thus, while the combined analysis provides strong support for cross-ancestry consistency at two established risk loci, population-specific effects may still exist but remain difficult to resolve with the current sample sizes and heterogeneous clinical datasets. Despite these limitations, the study addresses a major gap in genomic research by including a high-burden but underrepresented South Asian population within a well-characterised multicenter clinical cohort. Rigorous quality control, comprehensive analytical pipelines, and multiple complementary approaches are employed to ensure that association signals are robust and biologically interpretable. The integration of a non-European meta- analysis strengthened statistical power and reinforced support for known risk loci across ancestries. Importantly, the potential discovery of a biologically plausible, putative population-specific locus highlights the importance of well-designed studies in diverse populations, revealing mechanisms that may be absent or obscured in larger but ancestrally homogeneous cohorts. In future, establishing large ancestry-matched cohorts for replication and validation of associations, along with developing more representative reference panels, will be crucial. Detailed functional characterization of the newly identified locus, including experimental validation of its regulatory effects and downstream biological con- sequences, would further strengthen these findings. Targeted studies examining the endocrine reproductive inter- action in endometriosis may offer deeper insights into thyroid-related regulatory mechanisms in endometriosis. In conclusion, this study reports the first GWAS of endometriosis in an Indian population and highlights the importance of incorporating diverse ancestries in genetic research to fully capture the spectrum of risk fac - tors. We identified both shared and putative population-specific risk signals, with evidence of possible involve - ment of regulatory mechanisms involving WNT signalling, endocrine pathways and epigenetic regulation. The population specific signals, while requiring independent replication and functional validation, add to the current understanding of the genetics of endometriosis, also forms a basis for future ancestry-informed and mechanistic studies. Strengthening genomic research infrastructure in South Asia will be crucial to advancing equitable preci- sion medicine approaches for endometriosis and improving outcomes for affected women in the region. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9

Ethics statement The study was approved by the Institutional Ethics Committee of the Indian Council of Medical Research National Institute for Research on Women’s Health (ICMR-NIRWoH, formerly ICMR-NIRRCH, approval no. D/ICEC/ Sci-87/91/2019), the ethics committees of all participating study sites, and the Health Ministry Screening Com - mittee (HMSC), Government of India. The study was conducted in accordance with the Declaration of Helsinki, and the National Ethical Guidelines for Biomedical and Health Research Involving Human Participants issued by the Indian Council of Medical Research. All participants provided written informed consent prior to enrol - ment. Genetic data were de-identified before analysis, and data sharing was in accordance with national ethical and legal standards. Study design and participants This investigation was conducted as a part of the Endometriosis Clinical and Genetic Research in India (ECGRI) study, a nationwide, multicenter case–control study. The detailed study design has been published previously. 17 Participants were recruited from multiple public and private tertiary care centres across five geographic regions of India, representing Indo-European, Dravidian, and Tibeto-Burmese ancestry groups. Cases and controls were enrolled contemporaneously at each site to minimise recruitment bias, and their distribution was broadly compa- rable across centres and inferred ancestry clusters. Women aged 18—50 years, with a laparoscopically confirmed endometriosis, were included as cases, with histo-pathological confirmation wherever available. Controls were unrelated women in the same age range recruited from the same sites, without a personal or family history of endometriosis. Clinical data were collected using the modified World Endometriosis Research Foundation Endometriosis Phenome and Biobanking Harmon- isation Project (WERF-EPHect) Patient Questionnaire, and surgical data were recorded using the WERF-EPHect Standard Surgical Form. Peripheral blood samples were collected from all participants for genomic analysis. DNA extraction and genotyping Genomic DNA was extracted from peripheral blood using QIAmp DNA Blood Kit (Qiagen, Hilden, Germany) as per the manufacturer’s instructions. DNA concentration and purity were assessed by spectrophotometry, and sam- ples with yields > 50 ng/µl, and acceptable purity ratios (A260/280: 1·8–2·0; A260/230: 1·8 -2·3) were included. Genotyping was performed using the Infinium Global Screening Assay version 3·0 with multi-disease panel (Illumina Inc., San Diego, CA, USA). Raw intensity data were processed using GenomeStudio version 2·0 (Illu- mina Inc.), and genotype clustering was performed using an Indian population-specific cluster file. 34 The com- plete analysis pipeline is summarized in Supplementary Figure S1. Quality control and Imputation Standard quality control procedures were applied to remove low-quality samples and variants. SNPs with low genotyping quality, minor allele frequency (MAF) < 1%, call rate < 95% or deviation from Hardy–Weinberg equi- librium (p < 1 × 1 0–6) were excluded. Genome Studio filters of GenTrain Score < 0·6, cluster separation < 0·3, and AB ratio < 0·3 were used to remove low-quality SNPs. PLINK (v2) was used to exclude SNPs with MAF < 1% (maf 0·01), call rates < 95% (geno 0·05), and deviating from Hardy–Weinberg equilibrium (p 0·2), and high heterozygosity (Mean F -3SD) were excluded. Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Population structure was assessed to detect population stratification, ethnicity, and batch effects using principal component analysis after merging study genotypes with the 1000 Genomes Project reference panel. Pre-Imputation quality control was performed using the HRC-1000G-check-bim-v4·3 tool, and SNPs were aligned to the GenomeAsia100K reference panel, ensuring that only high-quality, properly aligned SNPs were used for the imputation step. Genotypes were remapped to hg38 reference genome using LiftOver. Phasing, using haplotypes from the South Asian (SAS) population, and imputation were performed using the Westlake Imputation Server with a South and East Asian reference Database reference panel (SEAD) that pro - vides high-resolution coverage for the Asian population. 35 Genotypes were also phased using EAGLE V2 and imputed to the TOPMED reference panel using the TOPMed Imputation Server.36–38 Imputed SNPs were largely consistent between the two imputation approaches; however, the SEAD reference panel enabled the recovery of a greater number of low-frequency variants and was therefore used for subsequent analyses. Post-imputation filtering excluded variants with an imputation quality score (Rsq) < 0·5 and MAF < 0·01. After quality control, a total of 9,647,516 SNPs and 2,523 individuals (1288 cases and 1235 controls) were retained for analysis. Genome-wide association analysis Genome-wide association analysis was performed using logistic regression under an additive genetic model using PLINK version 2·0. Ten principal components generated in PLINK following LD pruning (indep-pairwise 50 5 0·2) were included in the model as covariates to account for population stratification. The model was also adjusted for age, genotyping batch, and homozygosity (F coefficient). A genomic inflation factor (λ) was calcu - lated in R (version 4·5·0), and a QQ plot was generated using the ‘qqman’ package to assess potential inflation of test statistics. Association analysis was also conducted using GCTA 39 to account for population stratification through a genetic relationship matrix (GRM). Functional annotation and fine mapping Functional Mapping and Annotation of Genome-Wide Association Studies (FUMA GWAS) platform40 was used to identify lead independent variants ( p < 1 × 1 0–5; r2 < 0·1). The 1000 Genome Project Phase 3 SAS population was used as the reference panel. Lead SNPs within 250 kb were merged to construct genomic risk loci, and can- didate genes were mapped. Preliminary functional annotation with eQTL datasets was also performed. Fine-mapping analysis was performed on the most significant independent risk loci identified by FUMA. A 500 kb genomic region around the lead SNP was fine-mapped using Sum of Single Effects (SuSiE). 41 Follow- ing alignment of genotypes, phenotypes, and covariates across samples, SuSiE was applied to GWAS summary statistics aligned to raw genotypes rather than LD reference panels. Assuming a single causal variant per locus (L1), 95% credible sets were obtained for each locus. This credible set was further mapped to regulatory sites across tissues using haploreg v4·2 42, phasing with 1000 Genome Phase 1 Asian reference panel. LD thresholds were not applied as the variants were derived from a credible set. The variants showing regulatory marks, includ- ing promoters and enhancers, were further annotated using ENCODEPROJECT 43 to identify nearby candidate cis-regulatory elements (cCRE), their classification, and the closest gene. Colocalisation and Mendelian randomisation analyses COLOC and SMR analyses were performed using cis-eQTL datasets from tissues relevant to endometriosis, including endometrium29, blood44, uterus, ovary, thyroid, and pituitary 45 and mQTL dataset from blood 46 and endometrium47. GWAS summary statistics within ± 250 kb of the lead variant (rs35236165, chr13) were har - monised with corresponding eQTL and mQTL summary data. COLOC was applied to estimate posterior prob - abilities for competing hypotheses. Evidence for colocalisation was defined as a higher probability of a shared Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 causal variant (PPH4 > PPH3). SMR analyses were conducted independently for each molecular dataset, with HEIDI testing used to differentiate between shared genetic effects and linkage-driven associations. Genes passing multiple-testing correction (pSMR 0·01) were considered to share a common genetic signal. Replication and allele frequency analysis Allele frequencies of the lead variants were compared between the three ethnic groups in our cohort (Indo- European, Dravidian, and Tibeto-Burmese) and across European (EUR), American (AMR), African (AFR), East Asian (EAS), and South Asian (SAS) ancestral populations by using population specific allele frequency esti - mates from LDpop tool48, which utilizes the 1000 Genome Project Phase 3 reference panel. Replication of previously published endometriosis risk variants was also tested. We assessed association of 49 lead variants identified in the largest European GWAS13, one genome wide significant variant reported in the Japanese GWAS12, 16 significant variants from the Taiwanese GWAS14, and 65 significant variants from meta- analysis of a large multi-ancestry GWAS of endometriosis.19. Polygenic risk score analysis The PRS for endometriosis derived from a European population20 was applied to our dataset to evaluate the pre- dictive utility of the European-derived polygenic scores for endometriosis in the Indian population. SNP weights generated by McGrath et al20 were used to generate PRSs in our dataset using the ‘score’ function in PLINK. The difference in the mean PRS between cases and controls was tested using a t-test in R. Participants were assigned to deciles in order of ascending PRS, and the OR was estimated between each decile and decile 1. To evaluate the predictive accuracy of the PRS AUC was calculated using the pROC package in R. Endometriosis phenotype was classified as per the revised American Society for Reproductive Medicine (r-ASRM) classification 49, with Stage I, II, III, and IV representing minimum, mild, moderate, and severe disease conditions. To assess if higher polygenic burden was associated with more advanced disease stage, we further examined the relation between PRS and endometriosis stages. This was done by comparing the PRS distribution across milder stages (I and II) against advanced stages (III and IV). Meta-analysis We performed an Asian meta-analysis for endometriosis combining our GWAS data with a Japanese GWAS study21 to have a total sample size of 85,284 (3,074 cases and 82,210 controls). The Japanese GWAS summary statistics were publicly available in the GWAS catalogue (GCST90018619). Variants with allele frequency < 1% were filtered out. Summary statistics from both the studies were harmonized to ensure consistent SNP identifiers and effect alleles. Meta-analysis was performed in METAL using an inverse-variance weighted fixed-effects model.50 Genomic control correction was applied to account for any inflation due to population stratification. Effective sample sizes were calculated to weight each of the summary statistics as Neff = 4/(1/cases + 1/controls). Cochran’s Q statistics and I2 were used to assess heterogeneity of effect estimates across both datasets. The meta- analyzed data was used to identify genome-wide significant association signals and further downstream analysis was conducted using FUMA to identify independent signals, lead SNPs and genomic risk loci. Genome-wide sig- nificant signals from the meta-analysis were integrated with the eQTL and mQTL data using SMR and COLOC to prioritize candidate genes as described above. Supplementary Information The online version contains supplementary material available at https://doi.org/10.1038/ s41598-026-69642-9.

This work was supported by the DBT Wellcome India Alliance Fellowship [Grant no. IA/ Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 CPHI/18/1/503933] awarded to Rahul K Gajbhiye. We acknowledge support from the DHR-ICMR Long Term Fellowship Abroad [No. 2024-00250] awarded to Sandhya Anand; Australian National Endometriosis Clinical and Scientific Trials Fel- lowship [UNSW-RG223907, awarded to Sally Mortlock. Grant Montgomery was supported by grants from the Australian National Health and Medical Research Council (GNT1177194) and the Novo Nordisk Foundation (NNF24OC0090764). Gita Mishra was supported by the Australian National Health and Medical Research Council Leadership Fellow (Grant No. APP2009577). We thank Arindam Maitra, BRIC-National Institute of Biomedical Genomics (NIBMG), Kalyani, for technical support in genotyping and sharing the Indian population-specific cluster files. Data collection was facilitated by and conducted in compliance with the World Endometriosis Research Foundation Endometriosis Phenome and Biobanking Harmonization Project (WERF EPHect). We acknowledge the contributions of the ECGRI consortium team and ECGRI study participants. Author contributions RKG conceptualized the study, designed the study protocol, coordinated with all site investigators of the ECGRI consortium, obtained funding, obtained ethics approvals, and revised the manuscript. GWM and SM contributed to the design of the final study protocol, provided domain knowledge expertise in data analysis, and contributed to drafting and revision of the manuscript. SA executed the study, trained site teams for sample collection and processing, managed the biorepository, performed DNA extractions and quality controls towards genotyping sample preparation, and oversaw sample handling, coding, and the genotyping workflow. SA and SM contributed to genomic data quality controls, data analysis, data interpretation, and drafted the manuscript. HM supervised data collection and the research team’s training, coordinated with the team at the study sites, and revised the manuscript draft. SS, CS, PLS, NKB, PDM, PP, BJ, MS, DK, KK, SaM, NM, AH, NS, RM, MVP, KP, AG, SK, participated in data and sample collection. ShK, SG, NB, and TK performed the DNA extraction and quantification for the study. SDM and GS provided technical and administrative support for the study. GDM provided technical support and supervised the study. All authors revised and approved the final version of the manuscript. Funding DBT Wellcome India Alliance Clinical and Public Health Fellowship [Grant no. IA/CPHI/18/1/503933, awarded to RKG], Australian National Endometriosis Clinical and Scientific Trials Fellowship [UNSW-RG223907, awarded to SM] and Department of Health Research-Indian Council of Medical Research Long Term International Fellowship grant [No. 2024–00250, awarded to SA]. GWM was supported by grants from the Australian National Health and Medical Research Council (GNT1177194) and the Novo Nordisk Foundation (NNF24OC0090764). G.D.M. is an Australian National Health and Medical Research Council Leadership Fellow (Grant No. APP2009577). Data availability Genome-wide association summary statistics generated in this study have been deposited in the GWAS Catalog under accession number GCST90994020. Individual-level genotype and phenotype data are not publicly available because they contain sensitive human genetic and clinical information and are subject to participant consent, institutional ethics approval, and national regulatory requirements in India. Requests for access to individual-level data may be submitted to the corresponding author (RKG) and will be considered subject to approval by the relevant institutional ethics commit - tee and completion of an appropriate data transfer/data use agreement. Processed association results supporting the findings of this study are provided in the Supplementary Information and Additional Information files. Summary statistics from the Japanese GWAS used in the meta-analysis were obtained from the GWAS Catalog under accession number GCST90018619. Code availability Publicly available softwares were used for all the analyses. Appropriate references have been cited in methods. Declarations Competing interests The authors have no competing interests to declare. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 Interna- tional License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Cre- ative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit h t t p : / / c r e a t i v e c o m m o n s . o r g / l i c e n s e s / b y - n c - n d / 4 . 0 / . Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9

1. Zondervan, K. T. et al. Endometriosis. Nat. Rev. Dis. Primer 4, 9 (2018). 2. As-Sanie, S. et al. Endometriosis: A review. JAMA 334, 64–78 (2025). 3. Munshi, H., Piyadigama, I., Senanayake, H. & Gajbhiye, R. K. Endometriosis and adenomyosis research priorities in India and Sri Lanka: A call for regional collaboration. Lancet Reg. Health Southeast Asia 34, 100547 (2025). 4. Feng, J., Zhang, S., Chen, J., Zhu, J. & Yang, J. Global burden of endometriosis in 204 countries and territories from 1990 to 2019. Clin. Exp. Obstet. Gynecol. 49, 235 (2022). 5. Munshi, H. & Gajbhiye, R. K. Redefining endometriosis as a priority in global health. The Lancet Obstetrics, Gynaecology, & Women’ s Health 1, e164–e165 (2025). 6. World Health Organization. Endometriosis; 2025 Oct 15 https://www.who.int/news-room/fact-sheets/detail/endometriosis 7. Munshi, H., Khan, T., Anand, S. & Gajbhiye, R. Expanding the women’s health paradigm beyond reproductive boundaries. Indian J. Med. Res. 161, 38–42 (2025). 8. Treloar, S. A., O’Connor, D. T., O’Connor, V . M. & Martin, N. G. Genetic influences on endometriosis in an Australian twin sample. Fertil. Steril. 71, 701–710 (1999). 9. Matalliotakis, I. M. et al. Familial aggregation of endometriosis in the Yale Series. Arch. Gynecol. Obstet. 278, 507–511 (2008). 10. Lee, S. H. et al. Estimation and partitioning of polygenic variation captured by common SNPs for Alzheimer’s disease, multiple sclerosis and endometriosis. Hum. Mol. Genet. 22, 832–841 (2013). 11. Gajbhiye, R., Fung, J. N. & Montgomery, G. W. Complex genetics of female fertility. Npj Genomic Med. 3, 29 (2018). 12. Uno, S. et al. A genome-wide association study identifies genetic variants in the CDKN2BAS locus associated with endometriosis in Japanese. Nat. Genet. 42, 707–710 (2010). 13. Rahmioglu, N. et al. The genetic basis of endometriosis and comorbidity with other pain and inflammatory conditions. Nat. Genet. 55, 423–436 (2023). 14. Sheu, J.J.-C. et al. Ethnic-specific genetic susceptibility loci for endometriosis in Taiwanese-Han population: a genome- wide association study. J. Hum. Genet. 69, 573–583 (2024). 15. Wang, W. et al. Pooling-based genome-wide association study identifies risk loci in the pathogenesis of ovarian endometrioma in Chinese Han women. Reprod. Sci. 24, 400–406 (2017). 16. Gurdasani, D., Barroso, I., Zeggini, E. & Sandhu, M. S. Genomics of disease risk in globally diverse populations. Nat. Rev. Genet. 20, 520–535 (2019). 17. Gajbhiye, R. K. et al. Protocol for a case-control study investigating the clinical phenotypes and genetic regulation of endometriosis in Indian women: The ECGRI study. BMJ Open 11, e050844 (2021). 18. Machiela, M. J. & Chanock, S. J. LDlink: a web-based application for exploring population-specific haplotype structure and linking correlated alleles of possible functional variants. Bioinformatics 31, 3555–3557 (2015). 19. Guare, L.A. et al. Expanding the genetic landscape of endometriosis: integrative -omics analyses implicate key genes and pathways in a multi-ancestry study of over one million women. Res Sq. [Preprint] h t t p s : / / d o i . o r g / 1 0 . 2 1 2 0 3 / r s . 3 . r s - 8 1 1 6 6 0 2 / v 1 (2025). 20. McGrath, I. M. et al. Interaction between genetic risk and comorbid conditions in endometriosis. HGG Adv. 6, 100456 (2025). 21. Sakaue, S. et al. A cross-population atlas of genetic associations for 220 human phenotypes. Nat. Genet. 53, 1415–1424 (2021). 22. Hedge, T. A. & Mason, I. Expression of Shisa2, a modulator of both Wnt and Fgf signaling, in the chick embryo. Int. J. Dev. Biol. 52, 81–85 (2008). 23. Yamamoto, A., Nagano, T., Takehara, S., Hibi, M. & Aizawa, S. Shisa promotes head formation through the inhibition of receptor protein maturation for the caudalizing factors, Wnt and FGF. Cell 120, 223–235 (2005). 24. Hu, L., Chen, W., Qian, A. & Li, Y .-P. Wnt/β-catenin signaling components and mechanisms in bone formation, homeostasis, and disease. Bone Res. 12, 39 (2024). 25. Rahmioglu, N. et al. Genome-wide enrichment analysis between endometriosis and obesity-related traits reveals novel susceptibility loci. Hum. Mol. Genet. 24, 1185–1199 (2015). 26. Matsuzaki, S. & Darcha, C. Involvement of the Wnt/β-catenin signaling pathway in the cellular and molecular mechanisms of fibrosis in endometriosis. PLoS ONE 8, e76808 (2013). 27. Xu, Y ., Gao, F., Zhang, J., Cai, P. & Xu, D. Fibroblast growth factor receptor 2 promotes the proliferation, migration, and invasion of ectopic stromal cells via activation of extracellular-signal-regulated kinase signaling pathway in endometriosis. Bioengineered 13, 8360–8371. 28. Zhang, M. et al. Research advances in endometriosis-related signaling pathways: A review. Biomed. Pharmacother. 164, 114909 (2023). 29. Mortlock, S. et al. Tissue specific regulation of transcription in endometrium and association with disease. Hum. Reprod. 35, 377–393 (2020). Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 30. Kirkegaard, S., Uldall Torp, N. M., Andersen, S. & Andersen, S. L. Endometriosis, polycystic ovary syndrome, and the thyroid: a review. Endocr. Connect. 13, e230431 (2024). 31. Shih, Y .-H., Lu, C.-H. & Lung, C.-C. Risk of thyroid disorders after diagnosis of endometriosis: A 20-year retrospective cohort study. Maturitas 206, 108824 (2026). 32. Peyneau, M. et al. Role of thyroid dysimmunity and thyroid hormones in endometriosis. Proc. Natl. Acad. Sci. U. S. A. 116, 11894–11899 (2019). 33. Zondervan, K. T., Becker, C. M. & Missmer, S. A. Endometriosis. N. Engl. J. Med. 382, 1244–1256 (2020). 34. Bhattacharjee, E. et al. Genetic variants associated with spontaneous preterm birth in women from India: A prospective cohort study. Lancet Reg. Health Southeast Asia 14, 100190 (2023). 35. Yang, M.-Y . et al. SEAD reference panel with 22,134 haplotypes boosts rare variant imputation and genome-wide association analysis in Asian populations. Nat. Commun. 15, 10839 (2024). 36. Das, S. et al. Next-generation genotype imputation service and methods. Nat. Genet. 48, 1284–1287 (2016). 37. Fuchsberger, C., Abecasis, G. R. & Hinds, D. A. minimac2: faster genotype imputation. Bioinformatics 31, 782–784 (2015). 38. Taliun, D. et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed program. Nature 590, 290–299 (2021). 39. Yang, J., Lee, S. H., Goddard, M. E. & Visscher, P. M. GCTA: A tool for genome-wide complex trait analysis. Am. J. Hum. Genet. 88, 76–82 (2011). 40. Watanabe, K., Taskesen, E., van Bochoven, A. & Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. Nat. Commun. 8, 1826 (2017). 41. Wang, G., Sarkar, A., Carbonetto, P. & Stephens, M. A simple new approach to variable selection in regression, with application to genetic fine mapping. J. R. Stat. Soc. B 82, 1273–1300 (2020). 42. Ward, L. D. & Kellis, M. HaploReg: a resource for exploring chromatin states, conservation, and regulatory motif alterations within sets of genetically linked variants. Nucleic Acids Res. 40, D930-934 (2012). 43. Kagda, M. S. et al. Data navigation on the ENCODE portal. Nat. Commun. 16, 9592 (2025). 44. Võsa, U. et al. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat. Genet. 53, 1300–1310 (2021). 45. GTEx Consortium. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). 46. Hatton, A. A. et al. Genetic control of DNA methylation is largely shared across European and East Asian populations. Nat. Commun. 15, 2713 (2024). 47. Mortlock, S. et al. Global endometrial DNA methylation analysis reveals insights into mQTL regulation and associated endometriosis disease risk and endometrial function. Commun. Biol. 6, 780 (2023). 48. Alexander, T. A. & Machiela, M. J. LDpop: an interactive online tool to calculate and visualize geographic LD patterns. BMC Bioinformatics 21, 14 (2020). 49. Revised American Society for Reproductive Medicine classification of endometriosis: 1996. Fertil. Steril. 67, 817–821 (1997). 50. Willer, C. J., Li, Y . & Abecasis, G. R. METAL: Fast and efficient meta-analysis of genomewide association scans. Bioinformatics 26, 2190–2191 (2010). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Authors and Affiliations Sandhya Anand1,2 · Hrishikesh Munshi1 · Shashank Shekhar3 · Chaitanya Shembekar4 · Pushp Lata Sankhwar5 · Nilajkumar Bagde6 · Pramathes DasMahapatra7 · Pranay Phukan8 · Bimal John9 · Milind Shah10 · Deepali Kale11 · Ketki Kulkarni11 · Swati Mahobia12 · Niraj Mahajan13 · Anil Humane14 · Nagendra Sardeshpande15 · Reji Mohan16 · Muralidhar V. Pai17 · Kedar Padte18 · Atul Ganatra19 · Sonal Kumta20 · Sakshi Gangurde1 · Nayna Barada1 · Tabassum Khan1 · Shagufta Khan21 · Geetanjali Sachdeva22 · Gita D. Mishra23 · Smita D. Mahale24 · Grant W. Montgomery2,25 · Article in Press Scientific Reports https://doi.org/10.1038/s41598-026-69642-9 Sally Mortlock23  · Rahul K. Gajbhiye1 on behalf of on behalf of the ECGRI Consortium Sally Mortlock [email protected] Rahul K. Gajbhiye [email protected] 1 Clinical Research Laboratory, Indian Council of Medical Research National Institute for Research On Women’s Health (ICMR-NIRWoH, Formerly ICMR-NIRRCH), J. M. Street, Parel, Mumbai 400012, India 2 Institute for Molecular Bioscience, The University of Queensland, St Lucia, QLD 4067, Australia 3 Department of Obstetrics and Gynaecology, All India Institute of Medical Sciences, Jodhpur 342005, India 4 Omega Hospital, Nagpur 440015, India 5 Department of Obstetrics and Gynaecology, King George’s Medical University, Lucknow, Uttar Pradesh 226003, India 6 Department of Obstetrics and Gynaecology, All India Institute of Medical Sciences, Raipur 492099, India 7 Spectrum Clinic and Endoscopy Research Institute, Kolkata 700017, India 8 Department of Obstetrics and Gynaecology, Assam Medical College, Dibrugarh 786002, India 9 Minimally Invasive Surgery Unit, Credence Hospital, Thiruvananthapuram 695011, India 10 Naval Maternity Endoscopy & Infertility Center, Solapur 413001, India 11 Nowrosjee Wadia Maternity Hospital, Mumbai 400012, India 12 Sai Baba Nursing Home, Raipur 492001, India 13 Department of Obstetrics and Gynaecology, Topiwala National Medical College and BYL Nair Charitable Hospital, Mumbai 400008, India 14 Department of Obstetrics and Gynaecology, Government Medical College, Nagpur 440003, India 15 Worli Hospital for Women, Mumbai 400030, India 16 Department of Obstetrics and Gynaecology, Sri Avittam Thirunal Hospital (SAT), Thiruvananthapuram 695011, India 17 Sikkim Manipal Institute of Medical Sciences, Gangtok 737102, India 18 Dr Kedar’s Maternity & Infertility Hospital, Panaji 403001, India 19 Progeny IVF Centre & Clinic for Women, Mumbai 400080, India 20 Fortis Hospital, Mulund, Mumbai 400078, India 21 Gamete Immunobiology, Indian Council of Medical Research National Institute for Research On Women’s Health (ICMR-NIRWoH, Formerly ICMR-NIRRCH), Mumbai 400012, India 22 Cell Physiology and Pathology, Indian Council of Medical Research National Institute for Research On Women’s Health (ICMR-NIRWoH, Formerly ICMR-NIRRCH), Mumbai 400012, India 23 Australian Women and Girls’ Health Research (AWaGHR) Centre, School of Public Health, The University of Queensland, Herston, Qld 4006, Australia 24 Indian Council of Medical Research National Institute for Research On Women’s Health (ICMR-NIRWoH, Formerly ICMR-NIRRCH), Mumbai 400012, India 25 Department of Health Science and Technology, Aalborg University, Aalborg, Denmark