Decoding the regulatory genetic architecture of endometriosis using AlphaGenome

22

23 Endometriosis is a complex, estrogen-dependent disease with a strong genetic component. 24 Although genome-wide association studies (GWAS) have identified multiple susceptibility 25 loci, most associated variants reside in noncoding regions, limiting biological interpretation 26 and causal gene identification. Moreover, GWAS gene prioritization is limited by incomplete 27 tissue-specific annotation coverage (e.g., GTEx, ENCODE, fine-mapping, Mendelian 28 randomization, and network-based methods). We therefore applied the AlphaGenome 29 artificial intelligence framework to prioritize endometriosis-associated variants based on 30 predicted uterus-specific regulatory effects. 31

32 We analysed the top 10,000 endometriosis-associated single-nucleotide polymorphisms 33 (SNPs) identified by previously published GWAS by Rahmioglu et al , using AlphaGenome 34 across multiple genomic output types. Uterus-specific predictions with high-confidence 35 effects (|quantile score| ≥ 0.90) were grouped into major regulatory modalities. 36 AlphaGenome-prioritized SNPs within ±500 kb of known GWAS loci were classified into 37 tiers based on the number of supported regulatory modalities, with broader support indicating 38 stronger multilayer regulatory evidence. Effect allele frequency, linkage disequilibrium (LD), 39 and overlap with previously published endometriosis-associated variants were also assessed. 40

41 AlphaGenome generated uterus-specific, 147,033 high-confidence signals across 10,000 42 endometriosis-associated variants, spanning six regulatory modalities including gene 43 expression, promoter activity, chromatin accessibility, transcription factor binding, histone 44 modification, and RNA splicing. Within the 42 established endometriosis GWAS loci, 45 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 3 AlphaGenome identified 42 alternative sub-threshold SNPs with stronger predicted uterus-46 specific regulatory effects than the published GWAS lead variants. Nineteen AlphaGenome-47 prioritized SNPs were classified as tier 1, showing support across all six regulatory 48 modalities, compared with five GWAS lead SNPs. Linkage disequilibrium analysis identified 49 eight tier 1 SNPs with weak-to-low LD (r² < 0.5) relative to the corresponding GWAS lead 50 variants, regulating majority of genes involved in estrogen-driven proliferation and 51 inflammatory signalling, highlighting their potential relevance to endometriosis pathogenesis. 52 Additionally, we identified 167 genome-wide significant SNPs outside 42 published GWAS 53 lead SNP loci including six tier 1 SNPs (rs1482061, rs7772579, rs6557140, rs2982571, 54 rs12631337 and rs79626929), encompassing genes nearby ESR1/ 6q25.1, substantiating 55 biological relevance for endometriosis pathogenesis. 56

57 AlphaGenome-based regulatory prioritization refined endometriosis-associated genome-wide 58 association study loci by identifying variants with stronger predicted uterus-specific 59 functional relevance. These findings provide a regulatory framework for prioritizing 60 candidate variants and genes for downstream functional validation in endometriosis. 61

AlphaGenome; endometriosis; GWAS; regulatory variant prioritization; uterus-62 specific regulation; non-coding variants; multimodal genomics; RNA splicing. 63 64 65 66 67 68 69 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 4

70 Endometriosis is a chronic inflammatory disorder affecting approximately 5-10% of women 71 of reproductive age and is characterized by the presence of endometrial-like tissue outside the 72 uterus, most commonly on pelvic organs [1]. The disease is associated with debilitating 73 pelvic pain, infertility, and reduced quality of life, and diagnosis is often substantially delayed 74 because definitive confirmation typically requires surgical visualization of lesions [2,3]. 75 Current treatment options remain limited, relying mainly on hormonal suppression or 76 invasive approaches to surgically remove the lesions, both of which have limitations and 77 important impact on women health. Biologically and clinically, endometriosis is considered a 78 heterogeneous condition characterized by variability in lesion types, disease stage, infertility, 79 and pain manifestations, suggesting that multiple pathogenic mechanisms contribute to 80 disease susceptibility and presentation. 81 Genetic factors make a major contribution to endometriosis risk, with heritability estimated at 82 approximately 50%, with substantial proportion attributed to common genetic variation [4,5]. 83 Recent large-scale genome-wide association studies (GWAS) have expanded understanding 84 of the genetic architecture of endometriosis [4,6–13]. In particular, a meta-analysis including 85 60,674 cases and 701,926 controls identified 42 genome-wide significant endometriosis-86 related loci [3]. Fine-mapping of these loci further resolved six high-confidence candidate 87 causal variants, including variants at or near SYNE1, HOXA10, HOXC10, LINC00629, ESR1, 88 and LNC-LBCS genes, all located in non-coding regions [3]. More recently, a large multi-89 ancestry GWAS and integrated multi-omics analysis identified 80 genomic regions associated 90 with endometriosis risk, including 37 newly reported loci [14]. Multi-omics integrative 91 analyses in several tissues have further linked endometriosis genetic risk to pathways 92 involved in cell differentiation, immune and hormonal regulation, tissue remodeling, and 93 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 5 inflammation [3,14]. Consistent with these findings, many implicated loci map near genes 94 involved in hormone signaling, uterine development, immune regulation, adhesion, and 95 angiogenesis [1,7,9,10,12,13,15–25], which are vital biological processes related to the 96 pathogenesis of endometriosis. However, despite these advances, the regulatory 97 consequences of most endometriosis-associated variants remain incompletely understood, 98 particularly because the majority reside in non-coding regions, making gene prioritization and 99 mechanistic interpretation from GWAS signals alone inherently challenging. As a result, 100 current GWAS-based approaches largely resolve association signals at the locus level rather 101 than identifying causal genes or regulatory mechanisms, particularly in hormonally regulated 102 tissues such as the uterus, where long-range and context-dependent gene regulation is 103 prominent. To address this limitation, there is a need for approaches that can systematically 104 link non-coding genetic variation to downstream regulatory effects across disease-relevant 105 tissues and molecular layers. 106 In this context, recent advances in deep learning models such as AlphaGenome represent a 107 major step forward in sequence-to-function modeling by enabling simultaneous prediction of 108 thousands of functional genomic output tracks at base-pair resolution across multiple 109 regulatory modalities, including gene expression, chromatin accessibility, transcription factor 110 (TF) binding, histone modifications, promoter activity, three-dimensional chromatin 111 organization and RNA splicing [26]. Unlike earlier models like SpliceAI [27], BPNet [28], 112 ProCapNet [29], among others, which were designed for more specific regulatory tasks or 113 were constrained by trade-offs between input sequence length and output resolution, 114 AlphaGenome integrates long-range genomic context of up to 1 Mb with high-resolution 115 output, enabling the modelling of both local and distal regulatory effects within a unified 116 framework [26]. The model has demonstrated state-of-the-art or near state-of-the-art 117 performance across multiple benchmarking tasks and has shown strong ability to predict the 118 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 6 molecular consequences of non-coding variants across diverse regulatory layers [26]. 119 Notably, AlphaGenome has also been shown to recapitulate known disease-associated 120 regulatory mechanisms, including non-coding variant effects near the TAL1 oncogene [30], 121 where it successfully captured coordinated changes in TF binding, chromatin accessibility, 122 and gene expression associated with oncogenic activation. Such performance highlights its 123 value as a framework for linking non-coding genetic variation to downstream molecular 124 consequences across multiple regulatory modalities [26]. Given that the majority of disease-125 associated GWAS variants as in endometriosis, lie in non-coding regions [3,5], tools such as 126 AlphaGenome may provide an important opportunity to bridge the gap between genetic 127 association and biological mechanism. This is particularly important for loci where fine-128 mapping, expression and methylation quantitative trait locus, or sub-phenotype analyses 129 suggest functional relevance, yet the precise molecular direction and breadth of regulatory 130 perturbation remain unresolved. 131 In this study, we integrated endometriosis GWAS risk loci [3,14] with uterus-specific 132 AlphaGenome predictions to investigate the regulatory architecture of disease-associated 133 variants across multiple molecular layers including transcription, promoter, chromatin 134 accessibility, TF binding, histone modification, three-dimensional chromatin organization 135 and RNA splicing. The variants based on this multimodal approach were scored to prioritize 136 loci which could predict regulatory interpretation of endometriosis. By combining large-scale 137 genetic association data with deep learning-based regulatory prediction, this work seeks to 138 refine the functional interpretation of endometriosis risk loci and provide new insight into the 139 tissue-relevant regulatory mechanisms underlying disease susceptibility and symptom 140 heterogeneity. 141 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 7

142 Data source and variant processing 143 A total of 10,000 SNPs (p < 3.36 × 10/i2 5) were selected from the endometriosis GWAS meta-144 analysis reported in study by Rahmioglu et al , [3] (Additional file 2: Table S1). The initial 145 variant information extracted from the GWAS source included chromosome, genomic 146 position in hg19, and the corresponding effect and non-effect alleles. To ensure compatibility 147 with AlphaGenome, which requires variant coordinates in the GRCh38/hg38 reference 148 genome, all SNPs were converted from hg19 to hg38 using the Ensembl REST API. Variants 149 were mapped individually, and only successfully converted variants were retained for 150 downstream analysis. RsIDs were then annotated using Ensembl variation records based on 151 hg38 genomic coordinates (Additional file 2: Table S2). For AlphaGenome prediction, 152 variants were encoded such that the alternate allele corresponded to the GWAS effect allele, 153 and the reference allele corresponded to the GW AS non-effect allele, after checking 154 compatibility with the hg38 genomic position. The processed variants were formatted into a 155 tab-separated input file compatible with AlphaGenome, including variant ID, chromosome in 156 chr format, hg38 genomic position, reference/non-effect allele, and alternate/effect allele 157 (Fig. 1, Additional file 1 and Additional file 2: Table S2). Results were further cross-158 referenced with a recent multi-ancestry endometriosis GWAS dataset [14] to investigate the 159 regulatory modalities associated with these genetic signals. 160 161 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 8 162 Figure 1. AlphaGenome workflow for prioritizing endometriosis-associated SNPs . A 163 total of 10,000 endometriosis-associated GWAS SNPs ( p < 3.36×10 /i1/i1 ) were converted 164 from hg19 to hg38 and analyzed using AlphaGenome with a 1 Mb sequence context across 11 165 regulatory genomic output types. Predictions were restricted to uterus-specific biosamples, 166 and high-confidence signals (|quantile score| ≥ 0.90). Genomic output types were grouped 167 into six regulatory modalities like gene expression, promoter activity, chromatin accessibility, 168 transcription factor binding, histone modification, and RNA splicing. Each SNP was assigned 169 a tier based on regulatory modalities (Tier 1 = all six modalities; Tier 5 = ≤ 2 modalities). 170 Tiered SNPs were evaluated across three groups: published GWAS lead SNPs at 42 genome-171 wide significant loci, AlphaGenome-prioritized alternative SNPs within ±500 kb of these 172 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 9 loci, and outside-locus SNPs. Results were further cross-referenced with a recent multi-173 ancestry endometriosis GWAS [14] to validate novel regulatory signals. 174 175 AlphaGenome-based variant effect prediction 176 Variant effect prediction was performed using the AlphaGenome Python SDK version 0.6.1 177 within a Conda-based computational environment. The model was configured for Homo 178 sapiens using a 1 Mb sequence context window, allowing both proximal and distal regulatory 179 elements surrounding each variant to be incorporated into the prediction [26]. For each SNP, 180 a genomic interval centered on the variant position was extracted and used as input to the 181 model. Variant effects were computed by comparing predicted regulatory signals between the 182

and alternate alleles using the variant-scoring framework implemented in 183 AlphaGenome. The AlphaGenome generates predictions across 11 raw genomic output types 184 based on experimental techniques such as RNA sequencing (RNA-seq), Precision Run-On 185 coupled with cap analysis (PRO-cap), Cap Analysis of Gene Expression (CAGE), Assay for 186 Transposase-Accessible Chromatin sequencing (ATAC-seq), DNase sequencing (DNase-seq), 187 Transcription Factor Chromatin Immunoprecipitation Sequencing (TF ChIP-seq), Histone 188 Chromatin Immunoprecipitation sequencing (Histone ChIP-seq), splice sites, splice junctions, 189 splice-site usage, and contact maps. For each variant-track combination, AlphaGenome 190 produced annotations describing the regulatory context and predicted effect size. These 191 included variant-level information, such as variant_id and the scored genomic interval; gene-192 level annotations including gene_id, gene_name, gene_type, and gene_strand; assay-specific 193 attributes including output type, assay title, track name, and track strand; and biosample 194 metadata including biosample name, biosample type, life stage, and ontology terms. 195 Additional regulatory descriptors, such as TF identity, histone mark type, and tissue 196 annotations, including GTEx tissue labels, were retained where available. Quantitative 197 outputs included raw prediction scores and normalized quantile scores, representing the 198 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 10 magnitude and direction of variant-associated regulatory effects. All predictions were retained 199 for downstream filtering and analysis. 200 Biosample-level exploratory analysis 201 To characterize the distribution of AlphaGenome predictions across biological contexts, 202 prediction outputs were summarized at the biosample level using metadata from each 203 genomic output types (Additional file 2: Table S3). For each biosample, aggregate metrics 204 were calculated to describe prediction abundance, strength, and regulatory breadth, including 205 the total number of variant-track signals, the number of high-confidence signals defined by 206 absolute quantile score ≥ 0.90, positive and negative high-confidence signal counts, the 207 number of available genomic output types, distinct output types, represented regulatory 208 modalities, and quantile-score summary statistics such as maximum and mean absolute 209 quantile score (Additional file 2: Table S4). Because biosamples differed in the number of 210 available AlphaGenome tracks, signal counts were also normalized by track availability. 211 Normalized total and high-confidence signal burdens were calculated by dividing the 212 corresponding signal counts by the number of available genomic output types for each 213 biosample, providing track-adjusted estimates of prediction burden. Biosamples were first 214 summarized across all biosample types and then filtered to retain ‘tissue biosamples’ for the 215 main tissue-level exploratory analysis, reducing potential over-representation of cell lines or 216 isolated cell types (Additional file 2: Table S4). Tissues were ranked by total and normalized 217 signal burden, while regulatory breadth was assessed using the number of distinct output 218 types and regulatory modalities. These summaries guided tissue-context selection for 219 downstream uterus-specific analyses. 220 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 11 Uterus-specific filtering and high-confidence prediction selection 221 Given the uterine origin of endometriosis, AlphaGenome predictions were filtered using 222 biosample annotations. Predictions corresponding to uterus-related biosamples were retained 223 based on metadata fields such as biosample_name, tissue annotation, and associated ontology 224 information. AlphaGenome quantile scores were used to quantify the relative magnitude of 225 predicted variant effects across different assays and biosamples. The quantile score represents 226 the relative strength of a variant-induced predicted signal change compared with a 227

distribution of model predictions. Positive quantile scores indicate an increase in 228 predicted signal associated with the alternate allele, whereas negative quantile scores indicate 229 a decrease relative to the reference allele. Given the large scale and heterogeneity of the 230 generated prediction output, a stringent high-confidence threshold was applied. Only 231 predictions with an absolute quantile score ≥ 0.90 were retained for the main uterus-specific 232 regulatory analyses. This threshold was used to prioritize strong predicted regulatory effects 233 while reducing noise across the multi-modal AlphaGenome output (Fig. 1, Additional file 1 234 and Additional file 2: Table S5). 235 Annotation of uterus-specific AlphaGenome predictions 236 The uterus-specific prediction table was annotated by mapping each AlphaGenome variant_id 237 to its corresponding rsID using the hg38 SNP reference file generated during variant 238 preprocessing (Additional file 2: Table S2). Locus information was incorporated by matching 239 annotated rsIDs to the curated lead-SNP locus table derived from the 42 genome-wide 240 significant endometriosis loci reported in the published GWAS study [3] (Additional file 2: 241 Tables S6 and S7). To characterize the spatial relationship between predicted regulatory 242 variants and their associated genes, the distance between each SNP and the corresponding 243 gene body was calculated in base pairs using hg38 gene annotation coordinates. For each 244 variant-gene pair, the SNP position was compared with the annotated start and end 245 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 12 coordinates of the matched gene. Variants located within the gene body were assigned a 246 distance of 0 bp, whereas intergenic variants were assigned the shortest linear distance to the 247 nearest gene boundary. Additional gene-level features, including gene start, gene end, and 248 gene strand, were incorporated into the final annotated uterus-specific prediction table 249 (Additional file 2: Table S8). 250 Regulatory category grouping and tier classification 251 For SNP-level downstream interpretation, uterus-specific high-confidence AlphaGenome 252 predictions were collapsed into integrated regulatory modalities rather than analyzed as 253 individual raw genomic output types. Although AlphaGenome generated predictions across 254 11 raw genomic output types, only the genomic tracks represented after uterus-specific 255 biosample filtering and high-confidence filtering were used for tier-based classification. 256 Accordingly, six uterus-specific regulatory modalities were defined: gene expression, 257 promoter activity/transcription initiation, chromatin accessibility, TF binding, histone 258 modification, and RNA splicing. RNA-seq predictions were assigned to the gene expression 259 category, reflecting predicted transcript abundance. CAGE predictions were assigned to 260 promoter activity, reflecting transcription start site-associated regulatory activity. ATAC-seq 261 and DNase-seq predictions were grouped under chromatin accessibility, representing 262 predicted open chromatin regions. TF ChIP-seq predictions were assigned to TF binding, 263 whereas histone ChIP-seq predictions were assigned to histone modification, representing 264 chromatin-state-associated regulatory signals. Splicing-related predictions, such as splice 265 junctions and splice-site usage, were grouped under RNA splicing, capturing predicted effects 266 on splice junctions and splice-site usage (Fig. 1). 267 Other AlphaGenome predicted genomic output types, including PRO-cap, splice sites, and 268 contact maps, were retained in the global AlphaGenome output summary but were not present 269 among the uterus-specific high-confidence tracks used for downstream SNP interpretation. 270 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 13 Therefore, these outputs were not included in the uterus-specific tiering framework (Fig. 2). 271 Tier classification was based only on the six collapsed regulatory modalities described above. 272 Each SNP was summarized according to the number of distinct uterus-specific high-273 confidence regulatory modalities in which it showed predicted regulatory effects. The SNPs 274 were then classified into five tiers according to regulatory-modality breadth: Tier 1, support 275 across all six modalities; Tier 2, support across five modalities; Tier 3, support across four 276 modalities; Tier 4, support across three modalities; and Tier 5, support across two or fewer 277 modalities. This six-regulatory modality tiering framework was applied consistently to the 278 published GWAS lead SNPs, AlphaGenome-prioritized SNPs, and outside-GW AS-lead-SNP 279 loci (Fig. 1). 280 Locus-based grouping of 42 lead GWAS SNPs and selection of AlphaGenome-281 prioritized representative SNPs 282 To compare uterus-specific AlphaGenome predictions with the published endometriosis 283 GWAS architecture, the 42 lead SNPs were assigned to the 42 genome-wide significant loci 284 identified in the GWAS meta-analysis [3] (Additional file 2: Table S7). In this GWAS 285 framework, loci were defined around lead SNPs representing the most significantly 286 associated variant within a regional association signal, using a ±500 kb window ( p < 5 × 10 -287 8). Following the same locus-based approach, the 42 published GW AS lead SNPs were used 288 to construct ±500 kb locus windows in hg38 coordinates. All uterus-filtered SNPs with high-289 confidence AlphaGenome predictions were assigned to these loci based on genomic position. 290 Where a SNP fell into overlapping locus windows, it was assigned to the nearest locus based 291 on distance to the center of the lead-SNP region. For each locus, retained SNPs were 292 summarized according to their AlphaGenome-predicted regulatory modalities. This included 293 the number of unique genomic track-level predictions, the number of supported regulatory 294 modalities, and the maximum absolute quantile score. SNP-level GWAS association p values 295 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 14 were mapped back to retained SNPs using chromosome and hg19 genomic position from the 296 original GW AS summary dataset (Additional file 2: Table S1). Within each locus, SNPs were 297 ranked according to GWAS significance and AlphaGenome regulatory support (Additional 298 file 2: Table S9). 299 To identify AlphaGenome-prioritized representative SNPs within the same GWAS loci, one 300 alternative SNP was selected per locus using the following criteria: the SNP had to be 301 genome-wide significant at p < 5 × 10/i2/i2 , distinct from the SNP with the smallest p value in 302 that locus, and supported by high-confidence uterus-specific AlphaGenome predictions. 303 Prioritization was based first on the number of genomic track-level predictions, followed by 304 the number of supported regulatory modalities, maximum absolute quantile score, and GWAS 305 p value (Additional file 2: Table S10). These SNPs are further referred as “AlphaGenome-306 prioritized SNPs” (Fig. 1 and Additional file 1). The final locus-level summary table 307 included, for each locus, the published GWAS lead SNP and the AlphaGenome-prioritized 308 representative SNP, together with genomic position, effect and non-effect alleles, effect allele 309 frequency, and GWAS association p value (Additional file 2: Table S11). Separate prediction 310 summary tables were also generated containing the full uterus-specific AlphaGenome 311 regulatory profiles for the published GWAS lead SNPs and AlphaGenome-prioritized SNPs 312 representing the 42 loci (Additional file 2: Tables S12 and S13). 313 Identification and characterization of outside-locus SNPs from the 10,000 314 endometriosis-associated GWAS SNPs 315 SNPs from the 10,000 GWAS-ranked variant set that did not fall within any of the 42 ±500 316 kb published GWAS lead SNP locus windows (n = 3,454) were classified as outside-locus or 317 unassigned SNPs (Fig. 1, Additional file 1 and Additional file 2: Table S14). These SNPs 318 were retained for a separate exploratory regulatory-track analysis. Genome-wide significant 319 outside-locus SNPs were identified using the same threshold applied in the locus-based 320 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 15 analysis, p < 5 × 10 /i2/i2 . Outside-locus SNPs were then filtered for the presence of at least 321 one high-confidence uterus-specific AlphaGenome prediction (absolute quantile score ≥ 322 0.90). Each retained outside-locus SNP was summarized according to the number of 323 supported regulatory modalities, number of unique genomic track-level predictions, 324 maximum absolute quantile score, and GWAS p value. The same six-category framework and 325 tier-classification system described above were applied to the outside-locus SNPs. SNP-level 326 information, including rsID, genomic position, GW AS p value, regulatory categories, 327 maximum absolute quantile score, and tier assignment, was recorded in the outside-locus 328 summary (Additional file 2: Table S15). 329 Linkage disequilibrium analysis between published endometriosis-associated GWAS 330 lead SNPs and AlphaGenome-prioritized SNPs 331 To evaluate whether AlphaGenome-prioritized SNPs represented the same association signals 332 as the corresponding published GWAS lead SNPs, pairwise linkage disequilibrium (LD) was 333 calculated for each published GWAS lead SNP-AlphaGenome-prioritized SNP pair (Fig. 1 334 and Additional file 1). LD was estimated using the 1000 Genomes Phase 3 European 335

panel to remain consistent with the ancestry-specific LD framework used in the 336 original GWAS study. For each of the 42 loci, the corresponding published GWAS lead SNP 337 and AlphaGenome-prioritized SNP were extracted from the European reference panel and 338 analyzed as an SNP pair. Both r 2 and D′ were calculated for each pair. The r 2 value was used 339 as the primary classification metric because it reflects how well one SNP predicts another, 340 whereas D ′ was retained as an additional LD descriptor but was not used for category 341 assignment. SNP pairs were classified into four LD categories based on r²: strong LD, 342 defined as r² ≥ 0.8; moderate LD, defined as 0.5 ≤ r2 < 0.8; weak LD, defined as 0.2 ≤ r2 < 343 0.5; and low LD, defined as r 2 < 0.2. For summary-level interpretation, strong and moderate 344 LD pairs were considered genetically linked to the published GWAS association signal, 345 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 16 whereas weak and low LD pairs were considered potentially distinct or partially independent 346 regulatory candidates within the same GWAS-defined locus. The resulting LD table included 347 the published GWAS lead SNP, the AlphaGenome-prioritized SNP, genomic positions, r2, D′ , 348 and LD category for each locus (Additional file 2: Table S16). 349 Cross-reference with multi-ancestry endometriosis-associated GWAS SNPs 350 To determine whether the outside-locus SNP set contained variants subsequently implicated 351 in endometriosis, the outside-locus/unassigned SNP list was compared with novel lead SNPs 352 reported in a more recent multi-ancestry GWAS study [14] (Fig. 1 and Additional file 1). 353 Reported novel SNPs were matched against the original 10,000-SNP AlphaGenome input 354 dataset using rsID and, where needed, chromosome-position information. For each matched 355 SNP, locus assignment status was checked to determine whether it fell within one of the 356 original 42±500 kb published GW AS lead SNP windows or belonged to the outside-357 locus/unassigned SNP group. Matched SNPs were then annotated with their AlphaGenome 358 regulatory-category support, maximum absolute quantile score, number of unique track-level 359 predictions, GWAS p value from the 10,000-SNP dataset, nearest gene reported in the new 360 GWAS, and tier classification. The final comparison table included overlap status with the 361 original 10,000-SNP dataset, outside-locus status, regulatory categories, and predicted uterus-362 specific AlphaGenome effects (Additional file 2: Table S17). 363 364 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 17

365 Global AlphaGenome prediction landscape across 10,000 endometriosis-associated 366 GWAS SNPs 367 AlphaGenome analysis of the 10,000 endometriosis-associated GWAS SNPs generated a total 368 of 325,451,290 variant-track prediction signals across 11 raw regulatory output types, in all 369 available biosamples, without restricting the initial analysis to uterus, which were further 370 grouped into seven broader functional regulatory modalities for interpretation (Fig. 2 and 371 Additional file 2: Table S18), including gene expression, promoter activity/transcription 372 initiation, chromatin accessibility, TF binding, histone modification, RNA splicing, and three-373 dimensional chromatin organization. The landscape of global predicted genomic outputs was 374 strongly dominated by RNA-seq, which accounted for 224,292,024 signals, corresponding to 375 68.9% of the total AlphaGenome output (Fig. 2). The next most represented output types 376 were TF ChIP-seq (9.9%), splice junctions (7.2%), and histone ChIP-seq (6.9%). In contrast, 377 CAGE (3.4%), DNase-seq (1.9%), A TAC-seq (1.0%), splice-site usage (0.7%), contact maps 378 (0.1%), PRO-cap (0.1%), and splice sites contributed smaller proportions of the total 379 prediction space (<0.1%) (Fig. 2 and Additional file 2: Table S18). When raw output types 380 were collapsed into broader regulatory modalities, gene expression represented the largest 381 component of the prediction landscape (68.9%), followed by TF binding (9.9%), RNA 382 splicing (7.9%), histone modification (6.9%), promoter activity/transcription initiation 383 (3.4%), chromatin accessibility (2.9%), and three-dimensional chromatin organization (0.1%) 384 (Fig. 2). 385 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 18 386 Figure 2. Global AlphaGenome prediction landscape across the 10,000-SNP dataset. 387 Nested donut chart showing the total number of AlphaGenome variant-track prediction 388 signals across 11 raw genomic output types in outer circle collapsed into seven regulatory 389 modalities in the inner circle. 390 391 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 19 Uterus-specific AlphaGenome predictions and prioritization reveals regulatory 392 heterogeneity across endometriosis GWAS loci 393 Biosample-level summarization of 10,000 endometriosis-associated GWAS SNPs revealed 394 substantial variation in AlphaGenome prediction signal representation across several tissues 395 (Additional file 2: Tables S4). The highest total number of variant-track prediction signals 396 were observed in stomach, spleen, adrenal gland, ovary, testis, liver, sigmoid colon, kidney, 397 spinal cord, and urinary bladder, indicating that raw signal abundance differed markedly 398 across tissues. This variation likely reflected differences in the number and diversity of 399 available AlphaGenome output types for each tissue. After normalization by the number of 400 available output types, differences in signal burden across tissue remained evident, showing 401 that the observed variation was not explained solely by output type availability. High-402 confidence signals were also unevenly distributed across tissues, indicating biological 403 variability in predicted regulatory activity (Additional file 2: Table S4). 404 405 406 407 408 409 410 411 412 413 414 415 Figure 3. Biosample-level AlphaGenome signal burden and regulatory breadth. 416 Each point represents a biosample plotted by the total number of AlphaGenome variant-track 417 prediction signals for 10,000 endometriosis-associated GWAS SNPs and the number of 418 detected regulatory modalities. Reproductive tissues are highlighted and annotated. Point 419 color indicates regulatory breadth among the highlighted tissues, ranging from one to six 420 detected modalities. Dashed vertical and horizontal lines define the burden and breadth 421 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 20 thresholds used to separate biosamples into low burden/narrow breadth, low burden/broad 422 breadth, high burden/narrow breadth, and high burden/broad breadth quadrants. 423 Among reproductive tissues, ovary and uterus were both strongly represented (Fig. 3). Ovary 424 contained 2,675,024 total prediction signals and 259,856 high-confidence signals across 425 seven genomic output types, whereas uterus contained 1,432,412 total prediction signals and 426 147,033 high-confidence signals across eight genomic output types spanning six regulatory 427 modalities (Fig. 4) such as gene expression (RNA-seq), promoter activity/transcription 428 initiation associated activity (CAGE), chromatin accessibility (combined ATAC-seq and 429 DNase-seq), histone modification (histone ChIP-seq), TF binding (TF ChIP-seq), and RNA 430 splicing (combined splice junctions and splice-site usage) (Additional file 2: Table S19). 431 RNA-seq was the dominant uterus-specific output type, contributing 1,132,788 predictions, 432 followed by splice junctions, CAGE, TF ChIP-seq, histone ChIP-seq, ATAC-seq, DNase-seq, 433 and splice-site usage. 434 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 21 435 Figure 4. Uterus Specific AlphaGenome prediction landscape across the 10,000-SNP 436 dataset. Nested donut chart showing the total number of AlphaGenome variant-track 437 prediction signals across 8 raw genomic output types represented in outer circle collapsed 438 into six regulatory modalities in inner circle. 439 440 Although ovary showed higher total and high-confidence signal counts, the uterus provided 441 the most disease-relevant tissue context for endometriosis and captured a broader regulatory 442 prediction landscape, with eight genomic output types compared with seven in the ovary. 443 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 22 Therefore, the uterine context was selected for downstream locus-level analyses despite the 444 overall signal distribution remaining largely dominated by transcript abundance–related 445 predictions. 446 Moreover, following tissue-specific prioritization, uterus-filtered high-confidence SNPs were 447 mapped to the 42 genome-wide significant endometriosis loci previously identified by GWAS 448 meta-analysis [3]. Using regulatory prioritization strategy, AlphaGenome identified 42 449 alternative “AlphaGenome-prioritized SNPs” that differed from the published GWAS lead 450 variants and were used for below comparative locus-level analysis (Fig. 5a). 451 Comparison of uterus-specific regulatory signal retention between GWAS lead and 452 AlphaGenome-prioritized SNPs 453 Before high-confidence filtering (absolute quantile score ≥ 0.90), both the published GWAS 454 lead SNPs and the AlphaGenome-prioritized SNPs retained predictions across the major 455 uterus-relevant regulatory modalities with broadly comparable total signal distributions (Fig. 456 5b and Additional file 2: Table S19). Consistent with this pattern, paired locus-level statistical 457 testing showed no significant differences ( p > 0.05) between the two SNP sets across the 458 tested modalities after false-discovery-rate (FDR) correction (Additional file 2: Table S20). 459 This indicates that, at the level of raw uterus-filtered prediction coverage, the two matched 460 42-SNP sets showed similar regulatory representation. 461 After applying the high-confidence threshold (absolute quantile score ≥ 0.90), a clearer 462 difference emerged between the two SNP sets. The AlphaGenome-prioritized SNPs retained 463 more uterus-specific high-confidence regulatory signals than the published GW AS lead SNPs 464 across most major regulatory modalities (Fig. 5b and Additional file 2: Table S19). Paired 465 Wilcoxon signed-rank tests across the 42 matched loci showed significantly higher high-466 confidence signal burdens for AlphaGenome-prioritized SNPs in RNA-seq, CAGE, ATAC-467 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 23 seq, DNase-seq, TF ChIP-seq, and histone ChIP-seq tracks after FDR correction (p < 0.05) 468 (Additional file 2: Table S20). In contrast, splice junctions and splice-site usage showed 469 nominal differences but did not differ significantly between the two SNP sets. Although 470 RNA-seq remained the dominant output type in both SNP sets, AlphaGenome-prioritized 471 SNPs retained more high-confidence non-expression regulatory signals than GW AS lead 472 SNPs, particularly for CAGE, ATAC-seq, DNase-seq, TF ChIP-seq, and histone ChIP-seq 473 (Fig. 5b and Additional file 2: Table S19). This shift was also visible after collapsing tracks 474 into regulatory modalities: compared with GW AS lead SNPs, AlphaGenome-prioritized SNPs 475 showed lower relative gene expression contribution (78.6% to 64.9%) and higher 476 contributions from promoter activity (3.3% to 6.7%), chromatin accessibility (6.0% to 477 11.1%), TF binding (2.7% to 5.4%), and histone modification (4.9% to 6.8%) after high-478 confidence filtering (Fig. 5c). These findings suggest that AlphaGenome prioritization 479 enriched variants with stronger uterus-relevant high-confidence regulatory effects beyond 480 gene expression alone. 481 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 24 482 Figure 5. Uterus-specific AlphaGenome prioritization of endometriosis-associated SNPs. 483 (a) Manhattan plot showing GWAS association strength for SNPs across chromosomes. Grey 484 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 25 points represent other SNPs in the 10,000-SNP dataset, red points indicate the 42 published 485 GWAS lead SNPs, and blue points indicate the 42 AlphaGenome-prioritized subthreshold 486 SNPs selected within corresponding GWAS loci. Selected loci are labelled, and the dashed 487 horizontal line indicates the genome-wide significance threshold of p = 5 × 10 /i2/i2 . (b) 488 Absolute number of retained uterus-specific AlphaGenome variant–track prediction signals 489 across major genomic output types for the two matched 42-SNP sets. Bars show counts for 490 published GWAS lead SNPs after uterus tissue filtering, published GWAS lead SNPs after 491 high-confidence filtering, AlphaGenome-prioritized SNPs after uterus tissue filtering, and 492 AlphaGenome-prioritized SNPs after high-confidence filtering. High-confidence filtering was 493 defined as an absolute quantile score ≥ 0.90. Asterisks indicate output types in which 494 AlphaGenome-prioritized SNPs showed significantly higher high-confidence signal burden 495 than published GW AS lead SNPs after FDR correction ( p < 0.05; paired Wilcoxon signed-496 rank test). (c) Bubble plot showing the proportional distribution of retained uterus-specific 497 prediction signals across regulatory modalities for the same two SNP sets and filtering stages. 498 Genomic output types were collapsed into six regulatory modalities. Bubble size and 499 percentage labels indicate the contribution of each modality to the total retained prediction 500 signals within each dataset/filtering stage. 501 502 AlphaGenome-prioritized SNPs show broader uterus-specific multi-modal regulatory 503 support than published GWAS lead SNPs 504 Uterus-specific AlphaGenome predictions showed that all 42 published GWAS lead SNPs 505 and alterative 42 AlphaGenome-prioritized SNPs retained at least one high-confidence 506 regulatory signal in the uterine context (Fig. 6a-d and Additional file 2: Table S12, S13). Tier-507 based stratification demonstrated substantial heterogeneity among the 42 GWAS lead SNPs, 508 with RNA-seq representing the most consistently supported regulatory category in the uterine 509 context (Fig. 6a). Overall, only 5 of the 42 published GWAS lead SNPs belongs to tier 1 510 category, while 4 additional lead SNPs were represented in tier 2 category. Most published 511 lead SNPs were assigned to tier 4 or tier 5, indicating that many statistically defined GWAS 512 lead SNPs showed limited predicted regulatory breadth in the uterine context (Fig. 6b). On 513 contrary, alternative AlphaGenome-prioritized lead SNPs demonstrated substantially stronger 514 uterus-specific regulatory support (Fig. 6c). Notably, 35 out of 42 AlphaGenome-prioritized 515 lead SNPs were classified within tier 1-3 categories, reflecting strong regulatory support 516 across transcriptomic, epigenomic, and splicing-related datasets in the uterus (Fig. 6d and 517 Table 1). This indicates a broader and more integrated regulatory landscape for the 518 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 26 AlphaGenome-prioritized SNPs compared with the GW AS lead SNPs (Additional file 2: 519 Table S13). On the other hand, the GWAS lead SNPs exhibited limited regulatory overlap 520 with fewer regulatory modalities predictions. Moreover, allele frequency comparisons 521 showed no systematic difference between AlphaGenome-prioritized and published GWAS 522 lead SNPs: prioritized SNPs had higher effect allele frequencies at 22 loci (0.426 ± 0.253) 523 and lower at 20 loci (0.385 ± 0.223; Additional file 2: Table S11). This balance suggests that 524 the richer regulatory signal observed in AlphaGenome-prioritized SNPs was not simply an 525 artifact of allele frequency. Collectively, these findings suggest that AlphaGenome-based 526 prioritization can refine GWAS loci by identifying variants with stronger predicted functional 527 relevance in the uterine regulatory context. 528 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 27 529 Figure 6. Uterus-specific regulatory profiles of GWAS lead and prioritized SNPs. ( a) 530 Heatmap showing high-confidence AlphaGenome predictions, defined as absolute quantile 531 score ≥ 0.90, across six integrated regulatory modalities for the 42 published GWAS lead 532 SNPs and the (b) UpSet plot showing the overlap of six integrated uterus-specific 533 AlphaGenome regulatory modalities among the 42 published GWAS lead SNPs. (c) Heatmap 534 showing high-confidence AlphaGenome predictions, defined as absolute quantile score ≥ 535 0.90, across six integrated regulatory modalities for the 42 AlphaGenome-prioritized 536 representative SNPs. (d) UpSet plot showing the combinatorial distribution of regulatory 537 support among the 42 AlphaGenome-prioritized SNPs selected within ±500 kb of the original 538 GWAS lead loci. In the heatmaps (a, c), positive predicted effects of the alternative allele are 539 shown in green, negative predicted effects in purple, and the absence of retained high-540 confidence predictions in grey. Colour intensity corresponds to the magnitude of the quantile 541 score. In the UpSet plots (b, d), the six integrated regulatory categories are highlighted in red 542 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 28 text, and bar heights represent the number of SNPs supported by each individual or combined 543 regulatory modality. 544 545 Table 1. Uterus-specific regulatory tier classification of endometriosis-associated loci. 546 547 Tier Number of regulatory modalities AlphaGenome predictions for the 42 lead SNPs representing the GWAS loci Representative loci from AlphaGenome-selected SNPs Number of loci Loci Number of loci Loci Tier 1 6 5 GREB1/2p25.1, SYNE1/6q25.1*, ABO/9q34.2, MLLT10/10p12.31, HOXC10/12p13.13* 19 ABO/9q34.2, CD109/6q13, CEP112/17q24.1, DNM3/1q24.3, GDAP1/8q21.11, GREB1/2p25.1, HOXC10/12p13.13*, IGF1/12q23.2, KDR/4q12, MLLT10/10p12.31, PDLIM5/4q22.3, PTPRO/12p12.3, RIN3/14q32.12, RNLS/10q23.31, SKAP1/17q21.32, SYNE1/6q25.1*, VEZT/12q22, VPS13B/8q22.2, WNT4/1p36.12 Tier 2 5 4 7p15.2/7p15.2, HOXA10/7p15.2*, RIN3/14q32.12, VEZT/12q22 11 7p15.2/7p15.2, BSN/3p21.31, CDKN2B-AS1/9p21.3, FRMD7/Xq26.2, FSHB/11p14.1, HOXA10/7p15.2*, ID4/6p22.3, NGF/1p13.2, SLC19A2/1q24.2, SRP14-AS1/15q15.1, and WT1/11p14.1 Tier 3 4 4 PDLIM5/4q22.3, HEY2/6q22.31, CDKN2B-AS1/9p21.3, FSHB/11p14.1 5 7p12.3/7p12.3, BMPR2/2q33.1, FAM120B/6q27, HEY2/6q22.31, . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 29 TEX11/Xq13.1 Tier 4 3 9 EBF1/5q33.3, FAM120B/6q27, FRMD7/Xq26.2, GDAP1/8q21.11, KCTD9/8p21.2, SKAP1/17q21.32, SLC19A2/1q24.2, VPS13B/8q22.2, WNT4/1p36.12 4 DLEU1/13q14.2, EBF1/5q33.3, ETAA1/2p14, KCTD9/8p21.2 Tier 5 2 14 7p12.3/7p12.3, ACTL9/19p13.2, CD109/6q13, CEP112/17q24.1, DLEU1/13q14.2, DNM3/1q24.3, ETAA1/2p14, ID4/6p22.3, IGF1/12q23.2, KDR/4q12, LINC00629/Xq26.3*, NGF/1p13.2, PTPRO/12p12.3, SRP14-AS1/15q15.1 2 ACTL9/19p13.2, LINC00629/Xq26.3* 1 6 ASTN2/9q33.1, BMPR2/2q33.1, BSN/3p21.31, RNLS/10q23.31, TEX11/Xq13.1, WT1/11p14.1 1 ASTN2/9q33.1 548 Legends: Loci were stratified according to the number of independent regulatory modalities 549 supported by uterus-specific AlphaGenome predictions. Six regulatory modalities were 550 considered: gene expression (RNA-seq), promoter activity (CAGE), chromatin accessibility 551 (ATAC-seq and/or DNase-seq), histone modification (histone ChIP-seq), transcription factor 552 binding (TF ChIP-seq), and RNA splicing (splice junction and/or splice-site usage). For each 553 locus, the presence of at least one high-confidence signal, defined as absolute quantile score ≥ 554 0.90, within a regulatory modality was counted as one unit of evidence. Each regulatory 555 modality was counted once per locus, regardless of the number of individual tracks, genes, or 556 transcripts detected within that category. Loci were classified into five tiers based on total 557 regulatory support: Tier 1, six categories; Tier 2, five categories; Tier 3, four categories; Tier 558 4, three categories; and Tier 5, two or fewer categories. Higher-tier loci represent variants 559 with broader multi-layer regulatory support across transcriptional, chromatin-associated, and 560 post-transcriptional processes. Asterisks indicate loci that were previously resolved by GWAS 561 fine-mapping as high-confidence candidate causal loci and were also identified in the present 562 AlphaGenome analysis. 563 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 30 AlphaGenome-prioritized SNPs are mostly linked to GWAS lead SNPs, while 14 loci 564 show potential independent regulatory candidates for endometriosis 565 LD analysis identified 14 loci in which the AlphaGenome-prioritized SNP showed weak-to-566 low LD with the corresponding published GWAS lead SNP, defined as r² < 0.5 (Additional 567 file 2: Table S16). These included nine loci with moderate LD, defined as 0.2 ≤ r² < 0.5: 568 CD109/6q13, ETAA1/2p14, GDAP1/8q21.11, KDR/4q12, RIN3/14q32.12, SLC19A2/1q24.2, 569 SRP14-AS1/15q15.1, TEX11/Xq13.1, and VEZT/12q22. Five additional loci showed low LD, 570 defined as r² < 0.2: SYNE1/6q25.1, CDKN2B-AS1/9p21.3, LINC00629/Xq26.3, 571 GREB1/2p25.1, and WNT4/1p36.12. These 14 weak-to-low LD loci may represent distinct or 572 partially independent regulatory candidates within the same GWAS-defined regions. 573 Importantly, among these, eight AlphaGenome-prioritized SNPs were classified as tier 1, 574 three as tier 2, and one each as tier 3, tier 4, and tier 5, showing high-confidence support 575 across all six integrated uterus-specific regulatory modalities. 576 In contrast, the remaining 28 of 42 loci showed at least moderate LD, defined as r² ≥ 0.5, 577 between the published GWAS lead SNP and the AlphaGenome-prioritized SNP. Of these, 22 578 loci showed strong LD, defined as r² ≥ 0.8, whereas six loci showed moderate LD, defined as 579 0.5 ≤ r² < 0.8. Among these 28 strong- or moderate-LD loci, 11 AlphaGenome-prioritized 580 SNPs were classified as tier 1, eight as tier 2, four as tier 3, three as tier 4, and two as tier 5. 581 These results indicate that, in most loci, AlphaGenome prioritization identified SNPs that 582 remained genetically linked to the original GWAS association signal while providing 583 stronger predicted uterus-specific regulatory support. At the same time, several low-LD 584 prioritized SNPs also showed broad regulatory support, suggesting potential additional 585 regulatory candidates within the same GWAS-defined regions. The complete list of published 586 GWAS lead SNPs, AlphaGenome-prioritized SNPs, r² values, D ′ values, and LD 587 interpretation and tier-based categories is provided in Additional file 2: table S16. 588 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 31 Outside-locus analysis identifies 167 additional genome-wide significant SNPs with 589 uterus-specific regulatory support 590 Among SNPs located outside the 42 published GWAS lead SNP locus windows, 167 SNPs 591 reached genome-wide significance (p < 5 × 10 /i2/i2 ) and showed high-confidence 592 AlphaGenome predictions in at least one uterus-specific regulatory modality (Additional file 593 2: Table S14-S15). Of these outside-locus SNPs, six were classified as tier 1, including 594 rs1482061, rs7772579, rs6557140, rs2982571, rs79626929, and rs12631337. Five of these 595 SNPs mapped to chromosome 6 within the 6.q25.1 cluster (hg38) which spans a regulatory 596 block including genes such as SYNE1, ESR1, CCDC170, AKAP12, ZBTB2, ARMT1, and 597 RMND1 (Fig. 7a and Additional file 2: Table S21). Notably, these variants were not captured 598 within the predefined 42 GWAS lead SNP locus windows yet exhibited tier 1 regulatory 599 support in the uterine context, suggesting that the AlphaGenome-based prioritization 600 identifies additional high-confidence regulatory variants within known disease-relevant 601 genomic regions that are not recovered by conventional locus-based GWAS lead SNP 602 definitions. 603 The remaining tier 1 SNP out of six outside-locus SNPs (rs12631337) mapped to 604 chromosome 3 at chr3:50,161,104 in hg38 coordinates and was located within a gene-rich 605 region containing multiple protein-coding genes, including SEMA3F, GNAI2, HYAL2, RBM5, 606 RBM6, TUSC2, UBA7, TMEM115, IP6K1, APEH, SEMA3B, RASSF1, IFRD2, SLC38A3, 607 MAPKAPK3, and NPRL2 (Fig. 7b and Additional file 2: Table S21). Other genome-wide 608 significant outside-locus SNPs were distributed across lower regulatory tiers, with 9 SNPs 609 classified as tier 2, the 14 SNPs as tier 3, the 38 SNPs as tier 4, and 100 SNPs as tier 5. 610 Within tier 5, 51 SNPs showed support across two regulatory modalities, whereas 49 SNPs 611 retained only one high-confidence regulatory modality. Thus, although most genome-wide 612 significant outside-locus SNPs showed narrower regulatory support, a smaller subset 613 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 32 displayed broad multi-layer uterus-specific regulatory evidence. These findings indicate that, 614 beyond the predefined published GWAS loci, other genome-wide significant variants may 615 carry uterus-specific regulatory information relevant for downstream functional prioritization 616 (Additional file 2: Table S14-S21). 617 618 Figure 7. Genomic positions of endometriosis-associated SNPs on chromosomes 6 and 3. 619 Schematic representation of selected endometriosis-associated genomic loci showing GWAS 620 lead SNPs, AlphaGenome-prioritized SNPs, outside-locus SNPs, and nearby Ensembl-621 annotated AlphaGenome predicted protein coding genes using GRCh38 coordinates. (a) 622 Chromosome 6q25.1–6q25.2 locus containing the ESR1/SYNE1 region. (b) Chromosome 623 3p21.31–3p21.2 locus containing genes surrounding the selected chromosome 3 SNPs. 624 Outside-locus SNPs are shown in blue, the GWAS lead SNPs are shown in orange, and the 625 AlphaGenome-prioritized SNPs are shown in purple. Gene boxes indicate Ensembl-annotated 626 genes, and genomic positions are shown in megabases according to GRCh38. 627 628 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 33 Cross-reference with multi-ancestry GWAS SNPs reveals high-confidence uterus-629 specific regulatory support for newly reported endometriosis loci 630 To assess whether uterus-specific AlphaGenome predictions supported recently reported 631 endometriosis-associated variants, the AlphaGenome-ranked 10,000-SNP dataset was cross-632 referenced with 37 novel lead SNPs reported in a recent GW AS study [4,5]. Of these, 22 633 variants were present in the original 10,000-SNP dataset, whereas 15 were absent (Additional 634 file 2: Table S17). All 22 overlapping SNPs were located outside the original 42 published 635 endometriosis GWAS loci, based on the ±500 kb lead-SNP window definition, and none 636 corresponded to the original GWAS lead SNPs. The overlapping SNPs showed variable 637 degrees of uterus-specific regulatory support. Based on the number of high-confidence 638 AlphaGenome regulatory modalities detected, one SNP was classified as tier 1, three as tier 2, 639 four as tier 3, five as tier 4, and nine as tier 5 (Additional file 2: Table S17). Notably, 13 of 640 the 22 overlapping SNPs belonged to tiers 1-4, indicating support across three to six 641 regulatory modalities. The strongest signal was observed for rs17 99293, reported near 642 ARHGAP26, which showed high-confidence predictions across all six integrated categories, 643 tier 1 (Fig. 8). 644 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 34 645 Figure 8. Cross-reference of multi-ancestry endometriosis GW AS SNPs with uterus-specific 646 AlphaGenome-predicted regulated genes. Multi-ancestry endometriosis GW AS SNPs overlapping 647 the top 10,000 AlphaGenome-ranked variants are categorized into tier 1-4 according to the number of 648 supported high-confidence regulatory modalities. The left column shows the SNP identifier and 649 AlphaGenome tier. The middle column shows the nearest gene for the SNPs in the multi-ancestry 650 GW AS, and the right column shows genes or transcripts predicted by AlphaGenome to be regulated in 651 the uterus-specific dataset. Highlighted genes indicate common genes identified both as multi-652 ancestry GW AS-reported genes and as AlphaGenome-predicted regulated targets. 653 654 Importantly, several genes reported as nearest genes in the recent GWAS were also detected 655 as AlphaGenome-predicted regulated genes in the uterus-specific output, supporting 656 concordance between the external GW AS annotation and the present regulatory predictions 657 (Fig. 8). For example, rs2290358 was reported near CALD1, and AlphaGenome also 658 predicted regulation of CALD1, together with additional transcripts including AKR1B1, 659 TMEM140, BPGM, AKR1B15 , and nearby unannotated Ensembl transcripts. Similarly, 660 rs1799293 was reported near ARHGAP26 and showed AlphaGenome-predicted effects 661 involving ARHGAP26 as well as FGF1. The rs7974900 variant, reported near the SSPN 662 region, was also linked to predicted regulation of SSPN and additional transcripts including 663 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 35 BHLHE41, ITPR2, and RASSF8 . Other examples included rs223346 near UBE2D3, 664 rs62246055 near FOXP1, rs10824194 near ADK, and rs2834747 near RUNX1, for which the 665 GWAS-reported nearby gene was also represented among the AlphaGenome-predicted 666 regulated genes or transcripts (Fig. 8 and Additional file 2: Table S22). Together, this cross-667

indicates that a subset of recently reported multi-ancestry endometriosis GWAS 668 SNPs already present in the 10,000-SNP AlphaGenome dataset showed high-confidence 669 uterus-specific regulatory evidence. In several cases, AlphaGenome not only recovered the 670 GWAS-reported nearby gene but also identified additional potentially regulated transcripts 671 and multilayer regulatory modalities, providing functional support for these newly reported 672 endometriosis-associated loci. 673 674

675 Encyclopedia of DNA Elements (ENCODE) project data and subsequent Mendelian 676 randomization studies have consistently shown that ~90% of GWAS signals across complex 677 diseases map to non-coding functional elements [31,32]. Endometriosis is not an exception. 678 The majority of GWAS signals identified for endometriosis map to non-coding or intergenic 679 regions of the genome [33,34] suggesting that these variants confer disease risk through 680 regulatory rather than protein-altering mechanisms [35]. However, functional interpretation 681 of these loci remains challenging because the lead SNP at a GW AS locus selected based on 682 statistical association strength rather than functional evidence, may not be necessarily the 683 causal or most functionally relevant variant. In regions of LD, the lead associated SNP may 684 act as a proxy for nearby variants with stronger functional effects on gene regulation. 685 Moreover, assigning GWAS loci to causal genes remains challenging because the 686 functionally affected gene is not always the nearest gene to the associated variant. This is 687 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 36 particularly relevant for non-coding risk variants, whose functional impact and target genes 688 are often difficult to define [36], because such variants may influence distal regulatory 689 elements, chromatin interactions, or non-coding RNA-mediated regulatory networks. 690 Endometriosis, being a heterogenous disorder involving multiple molecular mechanisms, is 691 also characterized by dysregulated non-coding RNA networks, including miRNAs, lncRNAs, 692 and circRNAs in endometriotic lesions, eutopic endometrium, and peripheral immune 693 compartments [37–39]. These regulatory networks interact with epigenetic modifications, 694 chromatin remodeling, and TF activity to coordinate tissue-specific gene expression. 695 Importantly, because many non-coding RNA loci and regulatory elements reside within non-696 coding genomic regions, GWAS SNPs mapping to these regions may contribute to disease 697 susceptibility through disruption of regulatory and post-transcriptional mechanisms that are 698 not captured by protein-coding-centric analyses. Consequently, integrative computational 699 frameworks capable of modeling long-range regulatory effects are increasingly required for 700 functional interpretation of disease-associated non-coding variants. 701 Within this context, application of the AlphaGenome framework [26] enabled systematic 702 functional annotation of 10,000 endometriosis-associated GWAS SNPs in a uterus-specific 703 regulatory landscape. By leveraging up to 1 Mb of surrounding genomic sequence context for 704 each variant, AlphaGenome facilitated evaluation of variant effects across multiple regulatory 705 modalities, including gene expression, promoter activity, chromatin accessibility, TF binding, 706 histone modification, and RNA splicing. We used a tier-based classification framework for 707 integration of these predictions which further enabled refinement of uterus-specific regulatory 708 prioritization across GWAS loci. Collectively, these findings support the utility of long-range 709 sequence-based modeling approaches for improving identification and prioritization of 710 potentially functional non-coding variants underlying endometriosis susceptibility. 711 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 37 Previous GWAS follow-up studies have largely relied on experimentally derived functional 712 annotation resources, including expression quantitative trait locus datasets such as GTEx, 713 regulatory annotations from ENCODE and related epigenomic consortia, fine-mapping, 714 Mendelian randomization, and computational causal-gene prioritization methods [40]. More 715 recent network-based approaches, such as SigNet, integrate within-locus evidence with cross-716 locus gene and regulatory network information to prioritize likely causal genes at GWAS loci 717 [36]. Although these approaches are valuable for moving from association signals toward 718 candidate genes, they remain constrained by the availability, resolution, and tissue relevance 719 of the underlying annotation datasets, particularly for less studied diseases like endometriosis. 720 In particular, many resources have incomplete coverage of disease-relevant uterine context 721 and often prioritize genes rather than directly modelling variant-level regulatory effects 722 across multiple molecular layers. In contrast, our analysis enabled integrated evaluation of 723 endometriosis-associated variants across multiple predicted regulatory layers within a uterus-724 specific context. This provides a complementary framework for interpreting non-coding 725 GWAS signals beyond statistical association alone. 726 The comparison between published GWAS lead SNPs [3] and AlphaGenome-prioritized 727 SNPs showed that all published GWAS lead SNPs retained at least one predicted regulatory 728 signal in the uterus-specific analysis, indicating that the published GW AS loci remain 729 functionally relevant. This was consistent with partial concordance with independent fine-730 mapping evidence reported by Rahmioglu et al . [3], where four of six high-confidence non-731 coding variants also showed multi-layer regulatory support in the present analysis. However, 732 AlphaGenome-based prioritization identified alternative SNPs within the same GWAS-733 defined regions that exhibited broader multimodal regulatory support, suggesting improved 734 functional resolution beyond statistical association alone. Notably, only five published GWAS 735 lead SNPs were classified as tier 1, whereas 19 AlphaGenome-prioritized SNPs were 736 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 38 assigned to tier 1, reflecting a substantially higher proportion of variants with convergent 737 support across all six regulatory categories following re-prioritization. These results are 738 consistent with previous work showing that LD can make it difficult to distinguish lead 739 associated variants from nearby causal or functionally relevant variants within the same 740 GWAS locus [40]. 741 A subset of loci showed low LD between published GWAS lead SNPs (42 loci) and 742 AlphaGenome-prioritized SNPs, indicating that the most statistically significant GWAS SNPs 743 do not always correspond to the strongest predicted regulatory effect. Notably, among 14 loci 744 with weak-to-low LD (r² < 0.5), eight loci ( RIN3/14q32.12, GDAP1/8q21.11, KDR/4q12, 745 VEZT/12q22, CD109/6q13, WNT4/1p36.12, GREB1/2p25.1, SYNE1/6q25.1) were classified 746 as tier 1 with six regulatory modalities. Collectively, these findings highlight the potential of 747 regulatory annotation to refine causal inference and identify potential functional variation 748 beyond standard LD-based interpretation of GWAS signals. 749 Among these prioritized tier 1 loci, the WNT4/1p36.12,VEZT/12q22 and KDR/4q12 loci are 750 of particular relevance, as they have been consistently identified among the most significant 751 SNPs associated with endometriosis [41–47]. These loci also showed weak-to-low LD (r² < 752 0.5) between the AlphaGenome-prioritized SNPs and the corresponding GWAS lead variants, 753 suggesting that the strongest association signals do not necessarily reflect the most likely 754 causal regulatory variants. Together, these features provide informative examples of how 755 regulatory annotation can refine causal variant prioritization beyond published GWAS lead 756 SNPs. The WNT4/1p36.12 contains a high-affinity estrogen receptor alpha-binding site at the 757 locus, suggesting direct hormonal regulation of its expression [48]. WNT4 protein expression 758 is reduced in both eutopic and ectopic endometrium in endometriosis compared to normal 759 endometrium [49], it also regulates stromal cell proliferation, apoptosis, and decidualization, 760 processes essential for endometrial tissue remodelling. These alterations may disrupt 761 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 39 endometrial homeostasis and promote aberrant tissue remodelling, contributing to the 762 development and persistence of endometriotic lesions. Moreover, the expression of WNT4 in 763 normal peritoneum suggests that endometriosis arising via metaplasia cannot be completely 764 excluded [50]. Similarly, the VEZT/12q22 locus is characterised by increased VEZT 765 expression in ectopic compared with eutopic endometrium in women with endometriosis 766 [51], and its mRNA and protein expression shows a menstrual cycle stage-specific increase in 767 endometrial glands during the secretory phase [52]. The VEZT/12q22 locus contains an NF-768 κ B binding site, suggesting that risk variants may influence inflammatory signaling pathways 769 in endometriosis through modulation of NF-κ B–mediated transcriptional regulation [52]. The 770 KDR/4q12 locus encodes VEGFR2 gene , the principal mediator of VEGF-driven 771 angiogenesis and endothelial proliferation [47]. The risk variant (rs17773813[G]) at 772 KDR/4q12 locus is associated with disease severity, with larger effect sizes in stage III/IV 773 relative to stage I/II endometriosis [47] [3]. Collectively, these loci represent functionally 774 coherent examples spanning hormonal regulation ( WNT4/1p36.12), inflammatory signalling 775 (VEZT/12q22), and angiogenesis ( KDR/4q12) involved in endometriosis related pathways. 776 Thus, at well-established endometriosis susceptibility loci, AlphaGenome consistently 777 prioritized alternative SNPs, classifying them as tier 1 across all six regulatory modalities. 778 Notably, these SNPs were absent from the original GWAS reports and showed weak-to-low 779 LD (r² < 0.5) with the corresponding lead variants, underscoring the potential of regulatory 780 annotation to refine causal variant identification beyond conventional GW AS-based 781 approaches. 782 The alternate SNP was also suggested at the GREB1/2p25.1 locus, which was likewise among 783 the eight tier 1 loci showing weak-to-low LD with the published GWAS lead SNP , and 784 encodes an early estrogen-responsive gene and a key co-regulator of estrogen receptor 785 activity in hormone-driven tissues. Increased GREB1 expression has also been reported in 786 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 40 peritoneal endometriotic lesions compared with eutopic endometrium, suggesting a role in the 787 estrogen-driven proliferative environment of ectopic lesions [53]. Similarly, the 788 SYNE1/6q25.2 locus, another low-LD tier 1 locus identified by AlphaGenome, encompassing 789 ESR1, is one of the most architecturally complex susceptibility regions in endometriosis 790 genetics. This locus is regulated through long-range chromatin interactions involving ESR1, 791 CCDC170, ARMT1 , and SYNE1 genes, with distal regulatory elements and transcription 792 factors coordinating ESR1 expression [54]. This suggests that intronic variants in 793 SYNE1/6q25.2 locus may function as non-coding regulatory elements rather than protein-794 altering variants. Consistently, the multimodal regulatory activity identified at the 795 SYNE1/6q25.2 locus by AlphaGenome supports the functional relevance of this region in 796 estrogen-dependent endometriosis pathogenesis. 797 Finally, the CD109/6q13 locus, also belonging to the subset of eight tier 1 loci in which the 798 AlphaGenome-prioritized SNP exhibited weak-to-low LD with the published GWAS lead 799 variant, encodes a GPI-anchored glycoprotein that negatively regulates TGF- β signalling 800 [55,56]. Given the well-established involvement of TGF- β signalling in endometriosis 801 pathogenesis including inflammation, epithelial-mesenchymal transition, angiogenesis, and 802 fibrosis [57], CD109 represents a key modulator of these processes. Hence these 803 AlphaGenome-prioritized tier 1 loci, particularly those showing low LD with published 804 GWAS lead SNPs, represent examples of improved functional prioritization beyond statistical 805 association alone and further support the contribution of regulatory mechanisms to 806 endometriosis susceptibility. 807 Another important finding in our study was the outside-locus analysis, which extended the 808 AlphaGenome interpretation beyond the published 42 GWAS lead SNP windows. Although 809 the primary analysis was restricted to SNPs located within ±500 kb of each lead SNP , 3,454 810 SNPs lay outside these predefined regions, including 167 genome-wide significant SNPs with 811 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 41 high-confidence uterus-specific regulatory predictions. While most of these variants showed 812 limited regulatory support, six SNPs (rs1482061, rs7772579, rs6557140, rs2982571, 813 rs79626929, and rs12631337) were classified as tier 1 variants. Five of these SNPs 814 (rs1482061, rs7772579, rs6557140, rs2982571, and rs79626929) are located on chromosome 815 6 and are predicted to regulate genes within ESR1 /6q25.1 and SYNE1/6q25.2 loci including 816 ESR1, SYNE1, AKAP12, ZBTB2, ARMT1 and RMND1 (Fig. 7a). In our study, these variants 817 are designated as outside-locus SNPs because they fall outside the 500 kb window used for 818 locus definition, despite being in proximity to the published GWAS lead SNP rs71575922 819 and the AlphaGenome-prioritized SNP rs147631975 regulating SYNE1 gene. Notably, these 820 variants cluster near the outside-locus tier 1 SNPs, suggesting the presence of an extended 821 regulatory region spanning the ESR1/6q25.1-SYNE1/6q25.2 locus. 822 Among the outside-locus SNPs, tier 1 variant rs7772579 is in strong LD (r² = 0.98) with the 823 previously reported endometriosis-associated variant rs2206949 in ESR1 gene [46]. 824 Additionally, tier 2 SNP rs1971256, located in CCDC170 has also been reported as an 825 endometriosis risk variant [46]. The identification of multiple prioritized variants within 826 ESR1/6q25.1-SYNE1/6q25.2 region, further supports that this is a robust and consistently 827 replicated endometriosis susceptibility loci. ESR1 gene, encoding estrogen receptor α showed 828 robust genome-wide significant association across multiple meta-analyses, with several 829 independent signals at this locus further substantiating its central role in estrogen-driven 830 lesion growth and inflammation [8,46]. Nearby genes, including SYNE1, CCDC170, 831 AKAP12, ZBTB2, ARMT1, and RMND1 have also been implicated through GWAS and 832 transcriptome-wide analyses, showing expression correlations with ESR1 and cell-specific 833 expression in endometrial epithelial cells during the secretory phase, implicated in 834 implantation impairment, often associated with endometriosis [8,46,58]. AKAP12, while not 835 mapped to published GWAS lead SNPs, showed negative correlation with ESR1 and 836 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 42 functions in cytoskeletal and signalling regulation, supporting its potential modulatory role 837 within this network [59]. 838 Similarly, tier 1 outside-locus SNP rs12631337 predicted to regulate tumor suppressor gene 839 cluster on chromosome 3, including SEMA3F, GNAI2, HYAL2, RBM5, RBM6, TUSC2, 840 UBA7, TMEM115, IP6K1, APEH, SEMA3B, RASSF1, IFRD2, SLC38A3, MAPKAPK3, and 841 NPRL2 (Fig. 7b) . This SNP is in proximity to the BSN /3p21.31 locus together with the 842 GWAS lead SNP rs1352889. However, rs12631337 falls outside the predefined ±500 kb 843 locus boundary and was therefore classified as an outside-locus variant despite its likely 844 regulatory connection to this region. Among the genes predicted to be regulated by this SNP, 845 RASSF1 is particularly notable because its tumor-suppressor isoform, RASSF1A, has been 846 reported to show consistent epigenetic dysregulation in endometriosis, including promoter 847 hypermethylation in both ectopic and eutopic endometrium, reduced expression, and 848 association with disease severity [60]. 849 These findings indicate that additional genome-wide significant variants outside conventional 850 GWAS locus boundaries may also carry strong regulatory relevance in uterine tissue. Given 851 that GW AS locus definitions are constrained by lead SNP selection and fixed window sizes, 852 these findings suggest that outside-locus regulatory prioritization can complement standard 853 locus-based analyses by identifying potentially functional variants not captured within 854 established GWAS regions. The comparison of AlphaGenome-prioritized SNPs with multi-855 ancestry endometriosis GWAS SNPs [14] also supports the broader interpretation of 856 regulatory modalities. Of the 37 multi-ancestry lead SNPs, 22 were already present in the 857 10,000 SNP GWAS dataset [3] and all belonged to the outside-locus category rather than to 858 the 42 GWAS lead SNP windows. This suggests that SNPs located outside the original locus 859 framework may subsequently be identified as endometriosis-associated signals as larger and 860 more comprehensive GWAS datasets become available. Moreover, while in the multi-861 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 43 ancestry GWAS study, these SNPs were mapped to specific genes, AlphaGenome predicted 862 multiple regulatory target genes at the same loci, including the previously mapped genes, 863 thereby broadening the functional interpretation of these regions. Among these multi-ancestry 864 SNPs, rs1799293, previously mapped to ARHGAP26 gene was classified as tier 1 with 865 support across six regulatory modalities. AlphaGenome additionally identified FGF1 as 866 regulatory target gene for this SNP alongside the previously mapped gene ARHGAP26. 867 Promoter polymorphism -1385A/G (rs30411) at FGF1 has been associated with 868 endometriosis risk, with the A allele showing reduced frequency in endometriosis cases, 869 implicating altered angiogenic signaling [61]. In parallel, ARHGAP26 is downregulated in 870 ectopic and eutopic endometrium compared with normal endometrium and was negatively 871 associated with the severity of menorrhagia [62]. Together, the outside-locus and multi-872 ancestry GWAS comparison results suggest that functional interpretation may benefit from 873 extending beyond previously defined lead-SNP windows. While the 42 published loci remain 874 the central framework of this study, the outside-locus findings highlight additional genome-875 wide significant variants with predicted uterine regulatory effects. These variants may 876 provide useful candidates for future fine-mapping, replication, and functional validation as 877 larger GW AS datasets continue to refine the genetic architecture of endometriosis. 878

879 This study demonstrates that integrating GWAS signals with uterus-specific regulatory 880 predictions from AlphaGenome can refine the functional interpretation of endometriosis risk 881 loci. By extending analysis beyond published lead SNPs to include additional GWAS-ranked 882 and surrounding variants, we show that non-lead variants can exhibit stronger predicted 883 regulatory effects and multimodal support across loci. This re-prioritization framework, 884 supported by tissue-specific filtering and multi-layer regulatory annotation, improves the 885 identification of likely causal non-coding variants and provides a more nuanced view of the 886 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 44 genetic architecture underlying endometriosis. However, the findings are based on 887 computational predictions and require experimental validation. In addition, uterus-specific 888 filtering may not capture all disease-relevant cell types, and the strict prioritization threshold 889 may exclude variants with moderate functional effects. The fixed locus window and LD 890 assumptions may also miss long-range or ancestry-specific regulatory mechanisms. Overall, 891 this approach provides a refined and scalable framework that can be incorporated into 892 standard GWAS analysis pipelines to enhance variant prioritization and functional 893 interpretation, thereby supporting more systematic identification of candidate causal variants 894 in endometriosis and related complex traits. 895 Abbreviations 896 ATAC-seq: Assay for Transposase-Accessible Chromatin sequencing 897 CAGE: Cap Analysis of Gene Expression 898 DNase-seq: DNase sequencing 899 GTEx: Genotype-Tissue Expression 900 GWAS: Genome-wide association study 901 PRO-cap: Precision Run-On coupled with cap analysis 902 RNA-seq: RNA sequencing 903 SNP: Single-nucleotide polymorphism 904 TF ChIP-seq: Transcription factor chromatin immunoprecipitation sequencing 905 eQTL: Expression quantitative trait locus 906 LD: Linkage disequilibrium 907 3D: Three-dimensional 908 Declarations 909 Ethics approval and consent to participate 910 Not applicable. 911 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 45 Consent for publication 912 Not applicable. 913 Data Availability 914 All data analysed and presented in this study are included in this published article and its 915 additional files. Additional global AlphaGenome raw prediction outputs generated from the 916 analysis of the 10,000 SNPs and the analysis code are available from the corresponding 917 author upon reasonable request. 918 Competing interests 919 The authors declare that they have no competing interests. 920 Funding 921 The present study was funded by: Horizon Europe grant (NESTOR, grant no. 101120075) of 922 the European Commission; Novo Nordisk Foundation (grant no. NNF24OC0092384); 923 Estonian Research Council (grants nos. PSG1082, PRG1076); Swedish Research Council 924 (grant no. 2024-02530); Estonian Ministry of Education and Research Centres of Excellence 925 grant TK214 name of CoE and the Sigrid Jusélius Foundation. 926 Authors’ contributions 927 Sanu Bifal Maji conceived the study design, performed the AlphaGenome-based variant 928 annotation and prioritization analyses, conducted data processing, generated figures and 929 tables, interpreted the results, and drafted the manuscript. Alberto Sola-Leyva, Apostol 930 Apostolov, Lucia Blanco-Rodriguez contributed to data interpretation, methodological 931 refinement, and manuscript revision. Andres Salumets and Amruta D. S. Pathare supervised 932 the study, contributed to conceptualization and interpretation of the results, and critically 933 revised the manuscript. All authors read and approved the final manuscript. 934 935 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 46

936 1. Zondervan KT, Becker CM, Koga K, Missmer SA, Taylor RN, Viganò P. Endometriosis. 937 Nat Rev Dis Primers. 2018;4:9. https://doi.org/10.1038/s41572-018-0008-5 938 2. Nnoaham KE, Hummelshoj L, Webster P, d’Hooghe T, De Cicco Nardone F, De Cicco 939 Nardone C, et al. Impact of endometriosis on quality of life and work productivity: a 940 multicenter study across ten countries. Fertility and Sterility. 2011;96:366-373.e8. 941 https://doi.org/10.1016/j.fertnstert.2011.05.090 942 3. Rahmioglu N, Mortlock S, Ghiasi M, Møller PL, Stefansdottir L, Galarneau G, et al. The 943 genetic basis of endometriosis and comorbidity with other pain and inflammatory conditions. 944 Nat Genet. 2023;55:423–36. https://doi.org/10.1038/s41588-023-01323-z 945 4. Treloar SA, O’Connor DT, O’Connor VM, Martin NG. Genetic influences on 946 endometriosis in an Australian twin sample. Fertility and Sterility. 1999;71:701–10. 947 https://doi.org/10.1016/S0015-0282(98)00540-8 948 5. Saha R, Pettersson HJ, Svedberg P, Olovsson M, Bergqvist A, Marions L, et al. Heritability 949 of endometriosis. Fertility and Sterility. 2015;104:947–52. 950 https://doi.org/10.1016/j.fertnstert.2015.06.035 951 6. GTEx Consortium. Genetic effects on gene expression across human tissues. Nature. 952 2017;550:204–13. https://doi.org/10.1038/nature24277 953 7. Ruth KS, Beaumont RN, Tyrrell J, Jones SE, Tuke MA, Yaghootkar H, et al. Genetic 954 evidence that lower circulating FSH levels lengthen menstrual cycle, increase age at 955 menopause and impact female reproductive health. Hum Reprod. 2016;31:473–81. 956 https://doi.org/10.1093/humrep/dev318 957 8. Marla S, Mortlock S, Houshdaran S, Fung J, McKinnon B, Holdsworth-Carson SJ, et al. 958 Genetic risk factors for endometriosis near estrogen receptor 1 and coexpression of genes in 959 this region in endometrium. Molecular Human Reproduction. 2021;27:gaaa082. 960 https://doi.org/10.1093/molehr/gaaa082 961 9. Hautakangas H, Winsvold BS, Ruotsalainen SE, Bjornsdottir G, Harder A VE, Kogelman 962 LJA, et al. Genome-wide analysis of 102,084 migraine cases identifies 123 risk loci and 963 subtype-specific risk alleles. Nat Genet. 2022;54:152–60. https://doi.org/10.1038/s41588-964 021-00990-0 965 10. Jones A V , Hockley JRF, Hyde C, Gorman D, Sredic-Rhodes A, Bilsland J, et al. Genome-966 wide association analysis of pain severity in dysmenorrhea identifies association at 967 chromosome 1p13.2, near the nerve growth factor locus. Pain. 2016;157:2571–81. 968 https://doi.org/10.1097/j.pain.0000000000000678 969 11. Painter JN, Anderson CA, Nyholt DR, Macgregor S, Lin J, Lee SH, et al. Genome-wide 970 association study identifies a locus at 7p15.2 associated with endometriosis. Nat Genet. 971 2011;43:51–4. https://doi.org/10.1038/ng.731 972 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 47 12. Johnston KJA, Adams MJ, Nicholl BI, Ward J, Strawbridge RJ, Ferguson A, et al. 973 Genome-wide association study of multisite chronic pain in UK Biobank. Zhu X, editor. 974 PLoS Genet. 2019;15:e1008164. https://doi.org/10.1371/journal.pgen.1008164 975 13. Ruth KS, Campbell PJ, Chew S, Lim EM, Hadlow N, Stuckey BG, et al. Genome-wide 976 association study with 1000 genomes imputation identifies signals for nine sex hormone-977 related phenotypes. Eur J Hum Genet. 2016;24:284–90. 978 https://doi.org/10.1038/ejhg.2015.102 979 14. Koller D, He J, Løkhammer S, Aranda S, Qiu D, Davtian D, et al. Multi-ancestry 980 genome-wide association and integrated multi-omics analyses of endometriosis and its 981 clinical manifestations. Nat Genet [Internet]. 2026 [cited 2026 May 9]; 982 https://doi.org/10.1038/s41588-026-02582-2 983 15. Cheng J, Deng Y , Zhou J. Role of the Ubiquitin System in Chronic Pain. Front Mol 984 Neurosci. 2021;14:674914. https://doi.org/10.3389/fnmol.2021.674914 985 16. Ding Y-Q, Luo H, Qi J-G. MHCII-restricted T helper cells: an emerging trigger for 986 chronic tactile allodynia after nerve injuries. J Neuroinflammation. 2020;17:3. 987 https://doi.org/10.1186/s12974-019-1684-0 988 17. Ilad RS, Fleming SD, Bebington CR, Murphy CR. Ubiquitin is associated with the 989 survival of ectopic stromal cells in endometriosis. Reprod Biol Endocrinol. 2004;2:69. 990 https://doi.org/10.1186/1477-7827-2-69 991 18. García-Gómez E, Vázquez-Martínez ER, Reyes-Mayoral C, Cruz-Orozco OP, Camacho-992 Arroyo I, Cerbón M. Regulation of Inflammation Pathways and Inflammasome by Sex 993 Steroid Hormones in Endometriosis. Front Endocrinol. 2020;10:935. 994 https://doi.org/10.3389/fendo.2019.00935 995 19. Gregus AM, Levine IS, Eddinger KA, Yaksh TL, Buczynski MW. Sex differences in 996 neuroimmune and glial mechanisms of pain. Pain. 2021;162:2186–200. 997 https://doi.org/10.1097/j.pain.0000000000002215 998 20. Lenert ME, Avona A, Garner KM, Barron LR, Burton MD. Sensory Neurons, 999 Neuroimmunity, and Pain Modulation by Sex Hormones. Endocrinology. 2021;162:bqab109. 1000 https://doi.org/10.1210/endocr/bqab109 1001 21. Fung JN, Girling JE, Lukowski SW, Sapkota Y , Wallace L, Holdsworth-Carson SJ, et al. 1002 The genetic regulation of transcription in human endometrial tissue. Human Reproduction. 1003 2017;32:893–904. https://doi.org/10.1093/humrep/dex006 1004 22. O’Mara TA, Spurdle AB, Glubb DM, Endometrial Cancer Association Consortium. 1005 Analysis of Promoter-Associated Chromatin Interactions Reveals Biologically Relevant 1006 Candidate Target Genes at Endometrial Cancer Risk Loci. Cancers. 2019;11:1440. 1007 https://doi.org/10.3390/cancers11101440 1008 23. Zondervan KT, Cardon LR, Kennedy SH. The genetic basis of endometriosis: Current 1009 Opinion in Obstetrics and Gynecology. 2001;13:309–14. https://doi.org/10.1097/00001703-1010 200106000-00011 1011 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 48 24. Nathan A, Reinhardt P, Kruspe D, Jörß T, Groth M, Nolte H, et al. The Wilms tumor 1012 protein Wt1 contributes to female fertility by regulating oviductal proteostasis. Human 1013 Molecular Genetics. 2017;26:1694–705. https://doi.org/10.1093/hmg/ddx075 1014 25. Jones MR, Goodarzi MO. Genetic determinants of polycystic ovary syndrome: progress 1015 and future directions. Fertility and Sterility. 2016;106:25–32. 1016 https://doi.org/10.1016/j.fertnstert.2016.04.040 1017 26. Avsec Ž, Latysheva N, Cheng J, Novati G, Taylor KR, Ward T, et al. Advancing 1018 regulatory variant effect prediction with AlphaGenome. Nature. 2026;649:1206–18. 1019 https://doi.org/10.1038/s41586-025-10014-0 1020 27. Jaganathan K, Kyriazopoulou Panagiotopoulou S, McRae JF, Darbandi SF, Knowles D, 1021 Li YI, et al. Predicting Splicing from Primary Sequence with Deep Learning. Cell. 1022 2019;176:535-548.e24. https://doi.org/10.1016/j.cell.2018.12.015 1023 28. Avsec Ž, Weilert M, Shrikumar A, Krueger S, Alexandari A, Dalal K, et al. Base-1024 resolution models of transcription-factor binding reveal soft motif syntax. Nat Genet. 1025 2021;53:354–66. https://doi.org/10.1038/s41588-021-00782-6 1026 29. Cochran K, Yin M, Mantripragada A, Schreiber J, Marinov GK, Shah SR, et al. 1027 Dissecting the cis-regulatory syntax of transcription initiation with deep learning [Internet]. 1028 Genomics; 2024 [cited 2026 Apr 13]. https://doi.org/10.1101/2024.05.28.596138 1029 30. Mansour MR, Abraham BJ, Anders L, Berezovskaya A, Gutierrez A, Durbin AD, et al. An 1030 oncogenic super-enhancer formed through somatic mutation of a noncoding intergenic 1031 element. Science. 2014;346:1373–7. https://doi.org/10.1126/science.1259037 1032 31. Cano-Gamez E, Trynka G. From GWAS to Function: Using Functional Genomics to 1033 Identify the Mechanisms Underlying Complex Diseases. Front Genet. 2020;11:424. 1034 https://doi.org/10.3389/fgene.2020.00424 1035 32. Maurano MT, Humbert R, Rynes E, Thurman RE, Haugen E, Wang H, et al. Systematic 1036 localization of common disease-associated variation in regulatory DNA. Science. New York, 1037 N.Y .; 2012;337:1190–5. https://doi.org/10.1126/science.1222794 1038 33. Garibaldi-Ríos AF, Rodríguez-Gutiérrez PG, García-Díaz JM, Zúñiga-González GM, 1039 Figuera LE, Gómez-Meda BC, et al. Regulatory Effects of Endometriosis-Associated Genetic 1040 Variants: A Multi-Tissue eQTL Analysis. Diseases. 2025;13:248. 1041 https://doi.org/10.3390/diseases13080248 1042 34. Rahmioglu N, Nyholt DR, Morris AP, Missmer SA, Montgomery GW, Zondervan KT. 1043 Genetic variants underlying risk of endometriosis: insights from meta-analysis of eight 1044 genome-wide association and replication datasets. Human Reproduction Update. 1045 2014;20:702–16. https://doi.org/10.1093/humupd/dmu015 1046 35. Hindorff LA, Sethupathy P, Junkins HA, Ramos EM, Mehta JP, Collins FS, et al. 1047 Potential etiologic and functional implications of genome-wide association loci for human 1048 diseases and traits. Proc Natl Acad Sci USA. 2009;106:9362–7. 1049 https://doi.org/10.1073/pnas.0903103106 1050 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 49 36. Mousavi Z, Arvanitis M, Duong T, Brody JA, Battle A, Sotoodehnia N, et al. 1051 Prioritization of causal genes from genome-wide association studies by Bayesian data 1052 integration across loci. Rigoutsos I, editor. PLoS Comput Biol. 2025;21:e1012725. 1053 https://doi.org/10.1371/journal.pcbi.1012725 1054 37. Ghafouri-Fard S, Dashti S, Taheri M. PCAT1: An oncogenic lncRNA in diverse cancers 1055 and a putative therapeutic target. Experimental and Molecular Pathology. 2020;114:104429. 1056 https://doi.org/10.1016/j.yexmp.2020.104429 1057 38. Panir K, Schjenken JE, Robertson SA, Hull ML. Non-coding RNAs in endometriosis: a 1058 narrative review. Human Reproduction Update. 2018;24:497–515. 1059 https://doi.org/10.1093/humupd/dmy014 1060 39. Abbaszadeh M, Karimi M, Rajaei S. The landscape of non-coding RNAs in the 1061 immunopathogenesis of Endometriosis. Front Immunol. 2023;14:1223828. 1062 https://doi.org/10.3389/fimmu.2023.1223828 1063 40. Long E, Williams J, Zhang H, Choi J. An evolving understanding of multiple causal 1064 variants underlying genetic association signals. The American Journal of Human Genetics. 1065 2025;112:741–50. https://doi.org/10.1016/j.ajhg.2025.01.018 1066 41. Meidyana S, Isfandiary S, Primariawan RY . WNT4 (rs7521902 and rs16826658) 1067 polymorphism and its association with endometriosis – A systematic review and meta-1068 analysis. European Journal of Obstetrics & Gynecology and Reproductive Biology. 1069 2024;295:111–7. https://doi.org/10.1016/j.ejogrb.2024.01.038 1070 42. Albertsen HM, Chettier R, Farrington P , Ward K. Genome-Wide Association Study Link 1071 Novel Loci to Endometriosis. Huang L-M, editor. PLoS ONE. 2013;8:e58257. 1072 https://doi.org/10.1371/journal.pone.0058257 1073 43. Mafra F, Catto M, Bianco B, Barbosa CP, Christofolini D. Association of WNT4 1074 polymorphisms with endometriosis in infertile patients. J Assist Reprod Genet. 1075 2015;32:1359–64. https://doi.org/10.1007/s10815-015-0523-1 1076 44. Meidyana S, Isfandiary S, Primariawan RY . WNT4 (rs7521902 and rs16826658) 1077 polymorphism and its association with endometriosis – A systematic review and meta-1078 analysis. European Journal of Obstetrics & Gynecology and Reproductive Biology. 1079 2024;295:111–7. https://doi.org/10.1016/j.ejogrb.2024.01.038 1080 45. Painter JN, Anderson CA, Nyholt DR, Macgregor S, Lin J, Lee SH, et al. Genome-wide 1081 association study identifies a locus at 7p15.2 associated with endometriosis. Nat Genet. 1082 2011;43:51–4. https://doi.org/10.1038/ng.731 1083 46. Sapkota Y , Steinthorsdottir V , Morris AP , Fassbender A, Rahmioglu N, De Vivo I, et al. 1084 Meta-analysis identifies five novel loci associated with endometriosis highlighting key genes 1085 involved in hormone metabolism. Nat Commun. 2017;8:15539. 1086 https://doi.org/10.1038/ncomms15539 1087 47. Steinthorsdottir V , Thorleifsson G, Aradottir K, Feenstra B, Sigurdsson A, Stefansdottir L, 1088 et al. Common variants upstream of KDR encoding VEGFR2 and in TTC39B associate with 1089 endometriosis. Nat Commun. 2016;7:12350. https://doi.org/10.1038/ncomms12350 1090 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 50 48. Pavlič ev M, McDonough-Goldstein CE, Zupan AM, Muglia L, Hu Y-C, Kong F, et al. A 1091 common allele increases endometrial Wnt4 expression, with antagonistic implications for 1092 pregnancy, reproductive cancers, and endometriosis. Nat Commun. 2024;15:1152. 1093 https://doi.org/10.1038/s41467-024-45338-4 1094 49. Liang Y , Li Y , Liu K, Chen P, Wang D. Expression and Significance of WNT4 in Ectopic 1095 and Eutopic Endometrium of Human Endometriosis. Reprod Sci. 2016;23:379–85. 1096 https://doi.org/10.1177/1933719115602763 1097 50. Gaetje R, Holtrich U, Engels K, Kissler S, Rody A, Karn T, et al. Endometriosis may be 1098 generated by mimicking the ontogenetic development of the female genital tract. Fertility and 1099 Sterility. 2007;87:651–6. https://doi.org/10.1016/j.fertnstert.2006.07.1533 1100 51. Meola J, Rosa E Silva JC, Dentillo DB, Da Silva WA, Veiga-Castelli LC, De Souza 1101 Bernardes LA, et al. Differentially expressed genes in eutopic and ectopic endometrium of 1102 women with endometriosis. Fertility and Sterility. 2010;93:1750–73. 1103 https://doi.org/10.1016/j.fertnstert.2008.12.058 1104 52. Holdsworth-Carson SJ, Fung JN, Luong HTT, Sapkota Y , Bowdler LM, Wallace L, et al. 1105 Endometrial vezatin and its association with endometriosis risk. Hum Reprod. 2016;31:999–1106 1013. https://doi.org/10.1093/humrep/dew047 1107 53. Pellegrini C, Gori I, Achtari C, Hornung D, Chardonnens E, Wunder D, et al. The 1108 expression of estrogen receptors as well as GREB1, c-MYC, and cyclin D1, estrogen-1109 regulated genes implicated in proliferation, is increased in peritoneal endometriosis. Fertil 1110 Steril. 2012;98:1200–8. https://doi.org/10.1016/j.fertnstert.2012.06.056 1111 54. Lung DK, Reese RM, Alarid ET. Intrinsic and Extrinsic Factors Governing the 1112 Transcriptional Regulation of ESR1. Horm Cancer. 2020;11:129–47. 1113 https://doi.org/10.1007/s12672-020-00388-0 1114 55. Nilufer R, Karina B, Paraskevi C, Rebecca D, Genevieve G, Ayush G, et al. Large-scale 1115 genome-wide association meta-analysis of endometriosis reveals 13 novel loci and 1116 genetically-associated comorbidity with other pain conditions [Internet]. Genomics; 2018 1117 [cited 2026 May 19]. https://doi.org/10.1101/406967 1118 56. Galarneau G, Fontanillas P, the Celmatix Research Team, the 23andMe Research Team, 1119 Clementi C, Hu-Seliger T, et al. Genome-wide association studies on endometriosis and 1120 endometriosis-related infertility [Internet]. Genetics; 2018 [cited 2026 May 19]. 1121 https://doi.org/10.1101/401448 1122 57. Young VJ, Brown JK, Saunders PTK, Duncan WC, Horne AW. The Peritoneum Is Both a 1123 Source and Target of TGF- β in Women with Endometriosis. Dettman R, editor. PLoS ONE. 1124 2014;9:e106773. https://doi.org/10.1371/journal.pone.0106773 1125 58. Guare LA, Das J, Caruth L, Rajagopalan A, Akerele AT, Brumpton BM, et al. Expanding 1126 the genetic landscape of endometriosis: Integrative -omics analyses implicate key genes and 1127 pathways in a multi-ancestry study of over one million women [Internet]. Obstetrics and 1128 Gynecology; 2024 [cited 2026 May 19]. https://doi.org/10.1101/2024.11.26.24316723 1129 59. Li H. Physiologic and pathophysiologic roles of AKAP12. Sci Prog. 1130 2022;105:368504221109212. https://doi.org/10.1177/00368504221109212 1131 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint 51 60. Wu Y , Zhang M, Zhang X, Xu Z, Jin W. Methylation status and protein expression of 1132 RASSF1A in endometriosis. Oncology Letters. 2016;11:4107–12. 1133 https://doi.org/10.3892/ol.2016.4512 1134 61. Kang S, Li S-Z, Wang N, Zhou R-M, Wang T, Wang D-J, et al. Association between 1135 genetic polymorphisms in fibroblast growth factor (FGF)1 and FGF2 and risk of 1136 endometriosis and adenomyosis in Chinese women. Human Reproduction. 2010;25:1806–11. 1137 https://doi.org/10.1093/humrep/deq128 1138 62. Jiang C, Gong W, Chen R, Ke H, Qu X, Yang W, et al. RhoA/ROCK/ARHGAP26 1139 signaling in the eutopic and ectopic endometrium is involved in clinical characteristics of 1140 adenomyosis. J Int Med Res. 2018;46:5019–29. https://doi.org/10.1177/0300060518789038 1141 1142 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 30, 2026. ; https://doi.org/10.64898/2026.06.27.26356730doi: medRxiv preprint