Site-dependent transcriptomic signatures of endometriosis are conserved across hormonal states

9 Endometriosis, a chronic condition in which endometrial tissue grows at other sites in the body, 10 produces lesions whose gene expression profiles vary by anatomical location and menstrual 11 cycle stage. The extent to which these location-dependent transcriptional differences persist 12 across hormonal states has remained unknown. Here, we compile a comprehensive single-cell 13 RNA sequencing atlas of 672,051 cells across 112 donors from publicly available sources 14 spanning the eutopic endometrium, menstrual effluent, ovaries, peritoneum, and ectopic 15 endometrium from peritoneal and ovarian sites. By training machine learning classifiers on 16

tissue signatures, we identify a subset of cells within ectopic lesions that retain a 17 uterine transcriptomic identity, which we term “core” lesion cells, and which are distinguished 18 from surrounding host tissue populations. We show that transcriptomic differences between core 19 lesion cells at both ectopic sites are robust across all phases of the menstrual cycle and in patients 20 receiving exogenous hormonal therapy, and validate these findings in independent datasets. Cell 21 type- and tissue-specific gene signatures derived from these core populations are sufficient to 22 classify disease status and lesion subtype in independent bulk tissue data across 168 patients. 23 These findings establish that although the core lesion cells maintain elements of eutopic 24 endometrium identity, they also contain information about the anatomical site of the lesion. This 25 anatomic-specificity provides a potential framework for cycle-independent endometriosis 26 diagnosis and subtyping. 27 28 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 3 Main 29 Endometriosis, a chronic inflammatory disease affecting approximately 10% of reproductive-age 30 women, is characterized by the growth of endometrial tissue outside of the uterus 1,2. The clinical 31 presentation is heterogeneous, ranging from dysmenorrhea (painful periods) and chronic pelvic 32 pain to infertility, and diagnosis , which requires laparoscopic surgery with histological 33 confirmation, is delayed by an average of 6 -10 years3,4. Over this interval, disease can progress 34 from minimal to severe stages as classified by the revised American Society for Reproductive 35 Medicine (rASRM) system5. Beyond staging, endometriosis can be typed by lesion location, most 36 commonly as ovarian endometriomas or peritoneal lesions, which are thought to represent distinct 37 pathological entities with divergent structural architectures and potentially distinct pathogeneses6. 38 However, no adequate non-invasive tools currently exist to distinguish one subtype from the other. 39 The growth and maintenance of ectopic lesions is driven in large part by estrogen signaling and 40 disrupted progesterone responsiveness 7–9, with additional menstrual hormones implicated in 41 disease progression 10,11. These hormonal dependencies have motivated the use of ovulation -42 blocking therapies as the primary medical treatment for endometriosis and have shaped the 43 assumption that hormonal cycling is a major driver of transcriptomic heterogeneity within ectopic 44 tissue. Indeed, the eutopic endometrium undergoes dramatic transcriptomic remodeling across the 45 menstrual cycle12,13, and this responsiveness is at least partially preserved in ectopic lesions6,14. To 46 investigate the cellular and molecular landscape of endometriosis across these hormonal states, 47 single-cell RNA sequencing (scRNA -seq), a technique that measures gene activity in individual 48 cells, has been applied to profile eutopic and ectopic tissues 14–17. These studies have revealed 49 transcriptomic heterogeneity across tissue sites and cell types, yet hormonal variation across the 50 menstrual cycle and between treated and untreated patients remains deeply confounded with 51 ectopic site, leaving it unclear which variable principally shapes the molecular identity of lesion 52 subtypes. 53 Here, we integrate publicly available scRNA -seq datasets spanning the eutopic endometrium, 54 menstrual effluent, ovaries, peritoneum, and ectopic endometrium from both peritoneal and 55 ovarian sites to construct a comprehensive multi -tissue reference atlas. Using cell type -level 56 classifiers trained on reference tissue signatures, we identify the core endometrial-like populations 57 within each ectopic site and characterize how these populations differ from the surrounding host 58 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 4 tissue microenvironment. We demonstrate that ectopic site, rather than hormonal state, is the 59 primary axis of transcriptomic variation among core lesion mesenchymal lineage cells, indicating 60 the robustness of potential RNA biomarkers across the menstrual cycle. Finally, we leverage the 61 cell type- and tissue-specific signatures of core ectopic cells to stratify independent bulk tissue 62 samples by disease status and lesion location, establishing a molecular framework for non-invasive 63 endometriosis diagnosis that is independent of hormonal condition. 64

65 Construction of a comprehensive multi-tissue reference atlas for endometriosis 66 To define cell types across endometriosis lesions in the context of their potential lesion sites, we 67 assembled an integrated reference atlas comprising 672,051 cells from 112 donors spanning 68 eutopic endometrium, menstrual effluent, ovaries, peritoneal biopsies, and ectopic endometrium 69 from both peritoneal (EcP) and ovarian (EcO) sites (Figure 1a,b). The atlas was constructed by 70 harmonizing publicly available single -cell RNA sequencing (scRNA -seq) datasets across these 71 tissues. Specifically, to represent eutopic and menstruating endometrial tissue, we incorporated the 72 Human Endometrial Cell Atlas (HECA) 17, which provides harmonized metadata across publicly 73 available eutopic endometrium samples, together with menstrual effluent profiles from Shih et al.18 74 that include healthy controls, symptomatic individuals, and endometriosis-positive individuals. To 75 represent ovarian tissue, we integrated healthy ovarian datasets from Jones et al.19, Tabula Sapiens 76 2.020, and Fonseca et al 15. To include ectopic tissue and enable direct comparisons between both 77 peritoneal and ovarian ectopic sites without introducing sampling bias, we restricted inclusion to 78 studies that profiled both EcP and EcO lesions . This included Fonseca et al. and Tan et al .16. To 79 incorporate a peritoneal tissue reference, we also included the EcP-adjacent samples from Tan et 80 al., and the unaffected ovary samples from Fonseca et al. as an additional ovary reference . 81 Although these datasets originated from different laboratories and collection time points, they 82 exhibited comparable read depth and gene detection rates (Supplementary Figure 1a,b), supporting 83 their suitability for joint integration. 84 To establish unified cell identities across these diverse tissues, we performed large language model 85 (LLM)-based de novo cell type annotation21, generating broad cell class labels across all integrated 86 datasets (Figure 1c; Methods). The resulting LLM annotations mapped closely to the expert -87 curated annotations in the HECA ( Supplementary Figure 1c) and aligned well with canonical 88 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 5 marker gene expression for each cell type (Figure 1d), validating the unified labels. The most 89 notable discrepancies between the LLM -based and HECA annotations involved T cell 90 subpopulations and red blood cells; the LLM -guided approach classified a greater proportion of 91 cells as blood cells, consistent with a conservative interpretation of the strong hemoglobin signal 92 observed broadly across the atlas in previous studies (Figure 1d). Beyond re producing known 93 annotations, integration coupled with LLM-based labeling highlighted additional populations and 94 transitional states not emphasized in the source studies. These included myofibroblast-like cells in 95 the eutopic endometrium (Figure 1d, Supplementary Figure 1c) and transitional epithelial cell 96 states co -expressing canonical epithelial and stromal markers. Additionally, we identified a 97 mesothelial cell population, which was most abundant in the ectopic peritoneal tissue (Figure 1c, 98 Supplementary Figure 1d). 99 Taken together, atlas integration substantially reduced batch effects across tissues, datasets, and 100 disease states (Supplementary Figure 1d-f) while yielding unified cell type annotations spanning 101 all sampled compartments. This comprehensive multi-tissue reference atlas provides a foundation 102 for defining tissue-specific transcriptomic signatures and for deconvolving cell type-level signals 103 in endometriosis. 104 Somatic mutations link ectopic lesions to eutopic endometrium 105 The prevailing hypothesis for the pathogenesis of endometriosis is that retrograde menstruation 106 leads to endometrial tissue implant ing within the peritoneal space 22. If ectopic lesions originate 107 from displaced eutopic endometrium, then both tissues should share a common mutational history. 108 To assess this, we identified 1,064 high -confidence somatic variants across paired eutopic (300 109 variants) and ectopic (764 variants) endometrium samples in the Fonseca et al. dataset (Methods) 110 (Supplementary Figure 2a). We first compared the mutational processes operating in each tissue 111 by analyzing their mutation spectra. The 6-class mutation spectrum revealed similar proportional 112 contributions in both tissues ( Supplementary Figure 2a), and the 96 -class trinucleotide spectrum 113 confirmed broad similarity between the two sites (r = 0.87) ( Supplementary Figure 2 b). Both 114 spectra exhibited a prominent SBS5 signature, characterized by frequent C>T and T>C mutations, 115 consistent with prior observations across eutopic and ectopic endometrium23,24. 116 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 6 Having established that both tissues undergo similar mutagenic processes, we next asked whether 117 the identified somatic mutations were subject to selection. We computed the ratio of 118 nonsynonymous to synonymous mutations (dN/dS) and evaluated whether the observed ratio 119 exceeded random expectations by generating null distributions across 100,000 permutations. In 120 each permutation, an alternative base was randomly assigned , weighted by the observed 121 trinucleotide spectrum, at the sequence-specific context of each coding variant position (Methods). 122 The observed dN/dS was indistinguishable from the null distribution ( Supplementary Figure 2c), 123 indicating that somatic mutations accumulate without detectable genome -wide selection for or 124 against protein-altering changes. Together with the shared mutagenic signatures, this suggests that 125 both tissues harbor a set of passenger mutations amenable to lineage tracing. 126 We therefore asked whether any somatic variants were shared between paired eutopic and ectopic 127 samples, and identified sets of shared mutations in two patients ( Supplementary Figure 2d). To 128 assess whether these shared variants reflected clonal relationships, we analyzed the clonal cell 129 fraction, defined as the proportion of cells confidently carrying each mutation (Methods). In both 130 patients, the shared mutations were present in comparably small fractions of cells at both sites 131 (Supplementary Figure 2e), consistent with an oligoclonal structure in which the ectopic lesion 132 does not arise from a single massively expanded clone. To determine whether any clonal expansion 133 nonetheless occurred at ectopic sites, we compared variant allele frequencies (VAFs) of all called 134 mutations across paired eutopic and ectopic samples in all four patients ( Supplementary Figure 135 2f). Across four of five paired samples, ectopic sites exhibited modestly elevated VAFs, suggesting 136 a degree of clonal expansion following the initial seeding event. 137 Finally, to understand whether somatic mutations accumulate differentially across cell types, we 138 mapped called variants back to their cell type of origin (Supplementary Figure 2g). Consistent with 139 previous results25, the stromal and fibroblast fraction carry variant alleles at a relatively abundant 140 rate. This result is not consistent with prior bulk observations that epithelial cells harbor the most 141 mutations, likely owing to lower cell numbers that led to reduced sequencing depth in the bulk 142 analysis. Notably, the distribution of mutations between eutopic and ectopic endometrium varied 143 substantially across cell types, motivating a cell-type-resolved analysis of the ectopic lesion. 144 Cell type-level classifiers distinguish endometrial -like from host tissue cells within ectopic 145 lesions 146 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 7 During atlas integration, we observed that certain ectopic endometrium stromal cell populations 147 co-clustered with ovarian stromal cells (Supplementary Figure 1d), raising the question of whether 148 an underlying transcriptomic signature for each reference tissue could be leveraged to classify cell 149 types within the ectopic endometrium as endometrial derived or host -tissue derived with high 150 confidence (Figure 1a). To test this, we trained logistic regression classifiers on differentially 151 expressed, highly cell type-specific genes to distinguish endometrial (eutopic/menstrual effluent) 152 from non-endometrial reference (ovary/peritoneum) cell types (Methods). The classifiers achieved 153 strong performance across the most abundant cell types (Figure 2a) and accurately assigned the 154 majority of cells as endometrial or non -endometrial in origin ( Supplementary Figure 3a), with a 155 pooled false discovery rate of 1.5% across all cell types . Notably, the feature genes driving the 156 majority of our cell type-level classifiers were enriched for endometrium-specific genes as defined 157 by RNA expression in the Human Protein Atlas (HPA) 26 (Supplementary Figure 3b), further 158 validating the biological relevance of the discriminative signal. 159 Given that endometriosis is defined by the growth of endometrial-like tissue outside the uterus, we 160 next sought to determine whether cells within resected ectopic sites exhibited an endometrial 161 transcriptomic identity. Applying our classifiers across the entire dataset ( Supplementary Figure 162 3c), we found that the majority of cells at ectopic sites were transcriptomically more similar to 163 their host tissue of residence than to the endometrium. The proportion of lesion cells classified as 164 endometrial-like, hereafter referred to as "core" lesion cells , varied by lineage: endothelial cells 165 showed the lowest fraction (5%), followed by mesenchymal (8%), myeloid (11%), epithelial 166 (15%), and lymphoid (33%) lineage cells. These modest proportions are consistent with prior 167 observations that stromal cells of endometrial origin constitute a minority of endometrioma tissue14 168 and that immune populations within lesions represent an admixture of host - and endometrium-169 derived cells 27. Comparing across ectopic sites, we observed broadly comparable rates of 170 endometrial-like classification, with the most notable divergence occurring among immune 171 lineages, where EcO myeloid cells and EcP lymphoid cells were more frequently classified as 172 endometrial-like ( Supplementary Figure 3 d,e). The relative enrichment of endometrial -like 173 myeloid cells in EcO is consistent with previous observations that the predominant macrophage 174 population in endometriomas exhibits a tissue -resident phenotype that is also enriched in the 175 eutopic endometrium of patients with endometriosis16. 176 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 8 To test whether core populations faithfully recapitulate the eutopic endometrial microenvironment, 177 we performed differential cell -cell interaction analysis 28 comparing the eutopic endometrium 178 across all endometriosis patients against the core cells of both the EcO and the EcP sites (Methods). 179 In the EcO, core cells retained conserved interactions among mesenchymal populations relative to 180 non-core cells ( Supplementary Figure 4a), suggesting preservation of structural cell 181 communication networks. The EcP maintained this mesenchymal interaction, but it also displayed 182 a greater number of enriched core interactions involving epithelial populations, more closely 183 resembling the eutopic endometrium than the EcO ( Supplementary Figure 4b). This divergence 184 likely reflects the altered structural dynamics between EcO and EcP, as well as the reduced 185 prevalence of epithelial tissue within endometrioma samples 14–16. Consistent with prior reports , 186 core EcP cells also exhibited well-conserved epithelial-to-stromal crosstalk29. 187 To directly assess transcriptomic similarity to eutopic endometrium, we performed differential 188 expression analysis between diseased eutopic endometrium and both core and non -core 189 populations within each ectopic site separately (Supplementary Figure 5a,b; Methods). Across cell 190 types, core populations consistently exhibited fewer differentially expressed genes (DEGs) relative 191 to diseased eutopic endometrium than non-core populations at the same site (Figure 2c), suggesting 192 that core cells are transcriptomically closer to endometrial tissue. We also observed a higher DEG 193 burden per cell type in EcO than in EcP core cells, suggesting that endometrial-like populations in 194 endometrioma may be more divergent from diseased eutopic endometrium than those in peritoneal 195 lesions. 196 To interpret the biological basis of these differences, we partitioned DEGs for each cell type into 197 core-specific, non-core-specific, and shared gene programs between EcO and EcP. We focus on 198 macrophages, which exhibited amongst the largest DEG burdens at both sites (Supplementary 199 Figure 5a,b), to illustrate the functional differences between ectopic sites . In EcO, non -core-200 specific macrophage DEGs were enriched for cell migration pathways (Supplementary Figure 5c), 201 consistent with evidence that macrophages can traffic from host tissue into endometrial lesions 30. 202 The core-specific and shared programs in EcO macrophages highlighted cytokine response and 203 positive regulation of smooth muscle proliferation , with the core -specific program additionally 204 including activation pathways (Supplementary Figure 5d,e), aligning with the inflammatory milieu 205 of endometriosis and macrophage contributions to lesion vascularization 31,32. In EcP, non -core-206 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 9 specific genes reflected a macrophage expansion response similar to that in EcO, but with the 207 addition of local proliferation pathways rather than solely migration, and with a stronger 208 extracellular matrix organizing component, consistent with prior results33 (Supplementary Figure 209 5f). EcP shared and core -specific macrophage signatures showed enrichment for golgi lumen 210 acidification and positive regulation of cell death , respectively (Supplementary Figure 5g,h), the 211 latter raising the possibility of macrophage-associated dysfunction within the lesions. 212 Having established that core -specific DEGs capture distinct biology characterizing ectopic cells 213 relative to eutopic tissue, we next asked how conserved these transcriptomic programs were 214 between lesion sites. Strikingly, Jaccard indices between EcO and EcP core-specific gene sets were 215 relatively low across individual cell types (Figure 2 d), with the greatest site -conserved changes 216 occurring within immune populations. Grouping core-specific DEGs by immune lineage revealed 217 that the majority of enriched programs were site - and lineage -specific. Specifically, l ymphoid 218 DEG pathways reflected chemotaxis in EcP and response to cytokines in EcO, while myeloid 219 pathways reflected differentiation programs in EcP and catabolic processes in EcO (Figure 2f). 220 These site-dependent transcriptomic signatures suggest that the local tissue environment makes a 221 substantial contribution to the molecular identity of endometrial-like cells. 222 Menstrual cycle stage and exogenous hormones modulate ectopic endometrium in a cell type-223 specific manner 224 Fluctuating estrogen and progesterone levels across the menstrual cycle are widely thought to 225 modulate the function of the endometrial lesion-resident cell types, analogous to hormone -226 responsive programs in the eutopic endometrium12,17. We asked whether these changes within the 227 ectopic endometrium could explain the varying degrees of pain experienced throughout the 228 menstrual cycle by endometriosis patients . These patients are known to experience primary 229 dysmenorrhea, which is an acute symptom limited to the perimenstrual window and driven by 230 myometrial hypercontractility, vasoconstriction and ischemic pain 34. In addition to this primary 231 pain, endometriosis patients experience a secondary dysmenorrhea, which can begin in the 232 secretory phase of the menstrual cycle 35,36, indicative of a pain mechanism which is entirely 233 separate from the hypercontractility in primary dysmenorrhea37. We first quantified the number of 234 genes varying along the menstrual cycle for each cell type with respect to its compartment of origin 235 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 10 (Methods). This revealed substantial differences in hormone responsiveness between the eutopic 236 endometrium and the ectopic sites for both immune and stromal cell compartments 237 (Supplementary Figure 6a). 238 Given that the menstrual cycle comprises two distinct phases (proliferative and secretory) 239 associated with different levels of endometriosis-related pain, and that exogenous hormones 240 constitute an additional hormonal state, we next sought to identify phase -dependent expression 241 patterns within each cell type. We calculated mean expression profiles of gene –cell type–tissue 242 triplets across the three hormonal strata and clustered these profiles into six groups with similar 243 expression dynamics (Supplementary Figure 6b; Methods). To identify the underlying gene 244 programs, we annotated the clusters using gene ontology biological processes (Methods), which 245 revealed correspondence with established endometrial biology. Because endometriosis-associated 246 pain is initiated in the secretory phase35, we focused on the three clusters with expression peaks or 247 troughs in this interval. Cluster 3 , which peaks in the secretory phase, was characterized by the 248 response to zinc ion, an element known to fluctuate throughout the menstrual cycle and accumulate 249 during the secretory phase 38. Conversely, Cluster 1, which troughs in the secretory stage, was 250 enriched for extracellular matrix assembly, consistent with post -menstrual stromal regeneration 251 and matrix deposition39, while Cluster 6, which also reached its lowest expression in the secretory 252 phase, was defined by antigen presentation and alpha beta T cell activation with simultaneous 253 suppression of NK cell -mediated cytotoxicity, inconsistent with prior observations of immune 254 tolerance in the secretory state40,41. This failure to adopt the tolerogenic immune state that normally 255 characterizes the secretory phase 41 may reflect a blunted progesterone response, a hallmark of 256 endometriosis driven by loss of the stimulatory PR-B isoform42,43. 257 We partitioned the cycle-responsive genes in the peaking and troughing groups by their tissue and 258 cell type of origin (Supplementary Figure 6c , d). The greatest number of cycle -responsive genes 259 occurred in mesenchymal populations, including stromal cells, fibroblasts, and vascular smooth 260 muscle cells. Notably, across all cell types, these genes exhibited site -specific signatures, 261 indicating that eutopic and ectopic endometrium diverge in their hormonal response. 262 To characterize the functional implications of secretory -phase dynamics restricted to ectopic 263 tissue, we took the union of EcO and EcP genes across cell types. The secretory -peak programs 264 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 11 were most strongly enriched for macrophage immune response and fibroblast ECM development 265 (Supplementary Figure 6e), consistent with the heightened inflammatory and fibrotic states 266 reported in endometriotic lesions33,44. Programs troughing in the secretory state further reinforced 267 an inflammatory signature, with macrophages displaying elevated cytokine production 268 (Supplementary Figure 6f). Unexpectedly, NK cells exhibited increased cytotoxicity, diverging 269 from their canonical behavior during the secretory phase45 (Supplementary Figure 6f), reinforcing 270 the position that the ectopic immune compartment becomes inflammatory at the initiation of 271 secondary dysmenorrhea 37. In contrast, monocytes contributed an anti -inflammatory profile, 272 largely aligned with prior evidence that macrophage ontogeny and recruitment status give rise to 273 functionally divergent populations within endometriotic lesions 27,46. The convergence of these 274 inflammatory and cytotoxic profiles within fibrotic lesions provides a molecular framework for 275 the secondary dysmenorrhea of endometriosis that is mechanistically distinct from primary. In 276 place of an acute myometrial contraction, the lesion ’s immune cells sustain a hormonally driven 277 inflammatory state linking hormonal fluctuations to the modulation of pain across the menstrual 278 cycle in endometriosis patients. 279 Ectopic site modulates cell type -specific transcriptomic programs independently of 280 hormonal state 281 While our analysis shows that ectopic endometrial cells are hormone responsive, it also revealed 282 pronounced, site-specific transcriptomic divergences from the eutopic endometrium among core 283 ectopic cell types. We therefore asked which effect was larger: whether the transc riptomic 284 variation in the ectopic endometrium primarily reflects changes in hormonal state or whether the 285 ectopic site itself imposes an independent and dominant effect on cell state (Figure 3a). 286 To disentangle these two variables, we computed Spearman correlations between the mean batch-287 corrected principal component coordinates of each core cell type under two conditions: first, 288 between EcO and EcP cells in matched hormonal states, and second, between naturally cycling 289 cells and those from patients receiving exogenous hormonal treatment at matched sites. A greater 290 number of cell types exhibited higher correlation between different hormonal states than between 291 different sites (Figure 3b), indicating that ectopic site exerts a stronger influence on transcriptomic 292 identity than hormonal status. This effect was not uniform across compartments. The gap between 293 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 12 site and hormonal correlations was widest for mesenchymal populations, including stromal cells, 294 myofibroblasts, and vascular smooth muscle cells, whereas immune cell types were comparatively 295 stable across both perturbations, with slightly higher correlation between sites than hormonal states 296 (Figure 3b). This finding is notable given the longstanding emphasis on interpreting hormonal 297 cycling as a primary driver of transcriptomic variation in ectopic endometrial tissue . To 298 systematically characterize this site -dependent heterogeneity, we identified co -expressed gene 299 modules within the same cell types across all donors, ectopic sites, and hormonal states (Methods). 300 This analysis yielded thirty-three gene expression modules, twenty-six of which were significantly 301 enriched for at least one Gene Ontology Biological Process term and were retained for annotation 302 (Figure 3c; Supplementary Data Table 1; Methods). These modules were variably used across cell 303 types, with each module showing its highest mean usage in a single cell type (Figure 3c). 304 We next restricted analysis to modules whose usage was specific to a cell type , which yield ed 305 twenty-two modules, and t hree of the cell types had no s pecific module and were assigned the 306 most highly used annotated modu le, giving twenty-five cell type-module pairs spanning twenty-307 three unique modules (Supplementary Data Table 2; Methods) . We then tested whether module 308 usage differed between EcO and EcP independently of hormonal state (Methods). Eleven of the 309 twenty-five cell type-module pairs showed significant site -dependent deviation, and these were 310 predominantly associated with structural and endothelial populations, with only two immune-311 associated program reaching significance , suggesting that the structurally important cell types 312 within the lesions are what drive site dependence (Figure 3d; Supplementary Data Table 3). Gene 313 set enrichment analysis of these site-dependent modules converged on several recurring functional 314 axes. Endothelial cells preferentially engaged cell migration programs involving CDH5, CLDN5, 315 and ROBO4 (Supplementary Data Table 1) , consistent with established roles for endothelial 316 angiogenesis47 in endometriosis . Lymphatic endothelial cells showed increased enrichment of 317 blood vessel morphogenesis programs in EcP across both the Secretory and Exogenous Hormones 318 states, in line with reports of increased lymphatic vessel density in peritoneal endometriosis48. 319 Epithelial lineage cells diverged between ectopic sites through intermediate filament assembly 320 programs involving KRT19, KRT17, and KRT16 (Supplementary Data Table 1), genes implicated 321 in recognized hallmarks of endometriosis such as tissue remodeling and epithelial-mesenchymal 322 transition49. Among mesenchymal cell types, extracellular matrix (ECM) organization programs 323 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 13 included genes such as GAS6 (Supplementary Data Table 1), a known regulator of endometriosis-324 related fibrosis, ECM remodeling, and immunosuppression in other disease contexts 50,51. 325 Myofibroblasts showed deviations in pathways involving IGFBP1, IGFBP2, and IGFBP4 326 (Supplementary Data Table 1 ), which exhibit differential expression across lesion sites 52 and 327 regulate fibrosis53 and stromal cell proliferation54 in endometriosis. Stromal populations displayed 328 increased usage of programs related to cytoplasmic translation and ECM organization, reflecting 329 the extensive fibrosis and remodeling characteristic of different lesion sites6,14. Among the immune 330 cell programs with significant site dependency, CD8 T cells were related to the unfolded protein 331 response, encompassing heat shock proteins indicative of the oxidative stress environment in 332 endometriosis55. Collectively, these divergences reveal a previously unappreciated site 333 dependency of endometriosis -associated transcriptomic programs spanning structural, epithelial, 334 and immune cell compartments. 335 Although these programs demonstrated site -specific enrichment, their usage varied across 336 hormonal states, prompting us to ask whether the site -dependent differences persisted across 337 different hormonal milieus. Remarkably, the EcO-to-EcP program deviations across all cell types 338 were conserved regardless of whether individuals were receiving exogenous hormones or 339 menstrual cycle stage (Supplementary Figure 7a), indicating that the observed site -specific 340 differences were independent of hormonal state. To validate this finding, we performed differential 341 expression analysis between EcO and EcP mesenchymal lineage core lesion cells from donors 342 receiving exogenous hormones in the Tan et al. dataset and applied hierarchical clustering of the 343 top differentially expressed genes to the same cell types across a range of hormonal states in the 344 entirely held-out Fonseca et al. dataset. Strikingly, EcO and EcP cells clustered apart from one 345 another regardless of hormonal state or cell type ( Supplementary Figure 7b), demonstrating that 346 site-driven transcriptomic differences are robust across the full spectrum of hormonal conditions. 347 The genes driving this separation carry mechanistic relevance in endometriosis. TNXB, MME, and 348 RHOB56–58 have each been implicated in disease pathogenesis through ECM remodeling or 349 epithelial-mesenchymal transition, while TCF21 and IL32 are recognized regulators of the 350 inflammatory phenotype in endometriosis 59,60. ID4 has been identified through genome -wide 351 association studies (GWASs) as having an increased association with severe relative to mild 352 endometriosis61, raising the possibility that these site -dependent signatures may also capture 353 variation in disease severity independent of rASRM staging. 354 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 14 Cell type lineages exhibit transcriptomic responses varying with endometriosis severity 355 The identification of a severity-associated gene within our site-dependent signatures prompted us 356 to ask more broadly whether cell type lineages in ectopic lesions exhibit differential transcriptomic 357 responses to increasing disease severity, and whether the differentially expressed genes shared 358 across severity stages differ by lesion site. To address this, we performed differential expression 359 analysis comparing control healthy reference tissue against non -core lesion cells and eutopic 360 endometrium against core lesion cells for each lineage, then assessed whether the resulting effect 361 sizes scaled with disease severity (Methods). Comparing these effect sizes between mild and 362 severe endometriosis across lineages, we observed differential directionality: vascular and 363 epithelial cell populations displayed decreasing effect sizes with greater severity, whereas immune 364 populations showed increasing effect sizes ( Supplementary Figure 8a), suggesting lineage -365 dependent transcriptomic remodeling as disease progresses. 366 Given these divergent lineage responses, we next examined whether the intersections between 367 differentially expressed genes in mild and severe endometriosis at each lesion site were distinct. 368 The Ec P sites exhibited a more conserved pan severity signature than the Ec O sites 369 (Supplementary Figure 8b, c), indicating less overall transcriptomic difference associated with the 370 EcP. Among the shared differentially expressed genes, both sites displayed similar proportions of 371 cell type-specific and tissue-specific signatures, comprising 30 cell type-specific genes identified 372 as cell type -specific in Tabula Sapiens 62 alongside 14 genes with endometrium, ovary, or 373 peritoneum specificity as annotated by the Human Protein Atlas 63 (Supplementary Figure 8b,c; 374 Methods). This co-occurrence of both cell type - and tissue-specific genes within severity -shared 375 signatures raised the question of whether such signatures carry diagnostic utility beyond our single-376 cell framework. 377 Cell type- and tissue-specific gene signatures are sufficient to stratify endometriosis 378 subtypes 379 To assess the diagnostic potential of the cell type- and tissue-specific signatures identified through 380 the severity analysis, we turned to independent bulk tissue microarray data spanning eutopic 381 endometrium and peritoneum from both endometriosis patients and control individuals, as well as 382 bulk lesion samples from deep infiltrating lesions, superficial peritoneal lesions, and 383 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 15 endometriomas64 (Figure 4a; Supplementary Figure 8a,b). Projecting the bulk expression of both 384 the cell type - and tissue -gene sets into PCA space revealed separation among tissue types 385 (Supplementary Figure 9a,b), and we hypothesized that a combined signature of both cell type - 386 and tissue-specific genes would further improve stratification (Figure 4a). 387 We therefore tested whether these signatures could classify disease status and lesion origin across 388 three clinically relevant comparisons (Methods). We first created a Random Forest classifier which 389 distinguishes between healthy and diseased eutopic endometrium. For this classifier , the tissue -390 specific genes , which included very few endometrium -specific markers , yielded a statistically 391 insignificant AUC of 0. 592 [0.492-0.691], whereas the cell type -specific signature achieved an 392 AUC of 0.821 [0.745-0.883] (Figure 4 b), demonstrating that cell type -specific signatures are 393 sufficient to deconvolve endometriosis status in the eutopic endometrium. A similar hierarchy 394 emerged for a classifier built to identify control versus diseased peritoneal tissue, where the tissue-395 specific classifier achieved a modest AUC of 0. 696 [0.564-0.826] and the cell type -specific 396 classifier reached 0.829 [0.702-0.930] (Figure 4 c). T he improved tissue -specific performance 397 relative to the endometrium comparison suggests a stronger contribution of tissue-level signatures 398 in distinguishing diseased peritoneum from control peritoneum. We then created a third classifier 399 for lesion subtype classification , where we identified if tissue came from a deep infiltrating 400 peritoneal lesion, a superficial peritoneal lesion or an endometrioma , the cell type-specific (AUC 401 = 0.904 [0.855-0.944]) performed worse than the combined tissue and cell type signature (AUC = 402 0.936 [0.891-0.970]) classifier (Figure 4d), reflecting an increased contribution of tissue -specific 403 genes in this context. Consistent with this, the tissue -specific signature alone was most effective 404 in classifying lesion type of origin (AUC = 0. 908 [0.863 -0.944]) (Figure 4d), further reinforcing 405 the presence of an underlying tissue -of-residence signature distinguishing lesion subtypes. 406 Notably, combining cell type - and tissue -specific signatures into a unified gene set yielded 407 minimal gains in classification accuracy compared to just the cell -type specific classifier , 408 consistent with the possibility that cell type -specific genes already encode sufficient tissue -level 409 information to distinguish disease states and lesion subtypes. 410 Several genes driving these classifiers have established biological relevance in endometriosis. 411 Across both healthy compared to patient tissue- and cell-type-specific classifiers, the stromal cell 412 genes CLDN11, previously proposed as an endometriosis biomarker65, and SERPINE2, a stromal-413 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 16 specific gene, were among the highest -contributing features (Supplementary Figure 9a-c). Both 414 genes showed a stepwise increase in expression from control to patient endometrium and 415 peritoneum, and from these tissues to lesions (Figure 4 e), with both genes reaching their highest 416 levels in endometrioma, consistent with the dense stromal compartment characteristic of this lesion 417 subtype. In the eutopic endometrium classifier, the top -contributing genes included immune-418 specific CD2, GZMH, and IL7R (Supplementary Figure 10a), consistent with prior evidence that 419 an enhanced inflammatory state in the eutopic endometrium may serve as an indicator of 420 endometriosis66. The peritoneum classifier also showed high importance of immune cell-specific 421 genes including RUNX3 and CCL4L2 (Supplementary Figure 10b), reflecting the established 422 inflammatory environment in the peritoneum of endometriosis patients67,68. In the lesion classifier, 423 stromal cell-specific genes such as SERPINE2 and CDLN11 (Figure 4e; Supplementary Figure 424 10c) were among the strongest contributors, consistent with the site -dependent mesenchymal 425 signatures identified in our earlier analyses. Notably, the lesion classifier assigned high importance 426 to EMX2, an endometrium -specific gene, with less or negligible importance for some ovary-427 specific genes ( GREB1 and OGN respectively) and the peritoneum -specific gene CFD 428 (Supplementary Figure 10c), suggesting that endometrial identity is one of the primary 429 discriminative features for lesion classification. 430 Altogether, these results demonstrate that cell type - and tissue-specific transcriptomic signatures 431 derived from the integrated single-cell atlas are sufficient to classify diseased from control tissue 432 in both the endometrium and peritoneum, while also capturing the spatial signatures that 433 distinguish lesion subtypes , establishing a molecular framework with potential translational 434 applications for endometriosis diagnosis and subtyping. 435

436 Endometriosis affects roughly one in ten women of reproductive age worldwide, yet after nearly a 437 century of research there are still no adequate tools for non -invasive diagnosis 1,2. This delay is 438 driven in large part by the persistent failure to identify robust non-invasive biomarkers at the blood, 439 urinary, or menstrual effluent level69. A fundamental barrier to biomarker discovery is the cellular 440 heterogeneity of ectopic lesions, which comprise an admixture of endometrial -derived cells and 441 host tissue constituents whose signals confound bulk molecular analyses. To address this, we 442 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 17 compiled, across multiple data sources, a comprehensive multi-tissue single-cell atlas, across 112 443 donors, of 672,051 cells across the eutopic endometrium, menstrual effluent, ovaries, peritoneum, 444 and ectopic endometrium from both peritoneal and ovarian sites. This enabled, for the first time, a 445 systematic deconvolution of endometrial -like transcriptomic signatures from the surrounding 446 microenvironment across multiple tissues, hormonal states, and disease severities. 447 A central finding of this study is that the majority of cells within resected ectopic lesions are 448 transcriptomically more similar to their host tissue of residence than to the eutopic endometrium. 449 This observation carries important implications for how single -cell studies of endometriosis are 450 interpreted. Previous comparisons of transcriptomic profiles between ectopic sites, without first 451 deconvolving the fraction of cells most similar to host tissue, are, in effect, comparisons between 452 two different tissue microenvironments rather than between the most disease-relevant endometrial-453 like populations. By training cell type-level classifiers on reference tissue signatures and applying 454 them across the ectopic endometrium, we were able to identify the subset of “core” lesion cells 455 which most closely resemble eutopic endometrial cells and to characterize them separately from 456 the host tissue compartment. This analytical framework is conceptually analogous to methods 457 developed for resolving cell -type location in single -cell data from spatially dependent gene 458 expression70,71, but applied here entirely within a single -cell context to resolve tissue -of-origin 459 identity. 460 The identification of core endometrial -like cells at both ectopic sites is broadly consistent with 461 Sampson's retrograde menstruation hypothesis 22, which posits that shed endometrial tissue 462 implants at ectopic locations to cause endometriosis . Recent genomic evidence has further 463 supported this model through the identification of shared somatic mutations between eutopic 464 endometrium and matched ectopic lesions 72,73. Our own analysis of paired eutopic and ectopic 465 samples from the Fonseca et al. dataset similarly identified shared variants between matched 466 tissues, alongside highly concordant mutational spectra dominated by the SBS5 signature 467 indicative of common mutagenic processes. Together, these shared mutations and spectral 468 concordance provide genomic evidence that cells in ectopic lesions descend from uterine 469 precursors, and the observation that variant allele frequencies are elevated in ectopic tissue relative 470 to eutopic endometrium suggests a modest clonal expansion at ectopic sites. The small clonal cell 471 fractions of shared mutations additionally point to an oligoclonal lesion structure, likely reflecting 472 both intrinsic oligoclonality of the seeded population and substantial contributions from the 473 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 18 surrounding tissue microenvironment. Our transcriptomic data complement this genomic 474 evidence: the retention of a uterine gene expression identity in core lesion cells, despite their 475 residence in a foreign tissue microenvironment , is precisely what the retrograde menstruation 476 model predicts. At the same time, the fact that the majority of cells at ectopic sites resemble their 477 host tissue rather than the endometrium indicates that lesion establishment involves extensive 478 remodeling and recruitment of local cell populations, consistent with a model in which retrograde-479 seeded endometrial cells serve as a nidus that co-opts the surrounding tissue environment. 480 Our classifier -based approach provides a complementary transcriptomic perspective, 481 demonstrating that the endometrial-like fraction within resected tissue is modest, ranging from 5% 482 in endothelial cells to 33% in lymphoid cells, and that the non -core compartment carries distinct 483 transcriptomic programs. Notably, we observed differential immune cell distributions between the 484 core and non-core compartments across ectopic sites: non-core macrophages at ovarian sites were 485 enriched for programs related to cell migration, while those at peritoneal sites showed greater 486 enrichment for proliferative programs. The migration signature in ovarian non -core macrophages 487 is consistent with evidence that endometriosis triggers continuous monocyte recruitment to lesion 488 sites, with macrophages of distinct origins exerting opposing effects on lesion establishment74. We 489 also observed differences in cell -cell communication preservation between the eutopic 490 endometrium and the two ectopic sites, likely reflecting the differences in macro-level structural 491 architectures that characterize ovarian endometriomas, which form enclosed pseudocysts through 492 invagination of the ovarian cortex, versus those in peritoneal implants6,75. Taken together, these 493

demonstrate that core endometrial -like cells carry known transcriptomic signatures of 494 endometriosis and provide insight into how the two lesion microenvironments diverge through 495 disease progression. 496 The eutopic endometrium undergoes dramatic transcriptomic remodeling across the menstrual 497 cycle12,13,17, and ectopic lesions have been shown to retain some degree of hormonal 498 responsiveness6,8,14. Accordingly, hormonal variation has long been considered a primary driver 499 of transcriptomic heterogeneity in endometriotic tissue, and this assumption has influenced both 500 experimental design and therapeutic development. Our results refine this assumption , 501 demonstrating that the mesenchymal cell types are more transcriptomically similar across 502 hormonal states than across ectopic sites and that hierarchical clustering of differentially expressed 503 genes in an entirely held -out dataset separates EcO and EcP cells from one another regardless of 504 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 19 hormonal condition. Critically, however, this site dominance is itself cell -type-dependent and it 505 holds most strongly for the mesenchymal and structural populations, whereas immune cell types 506 retain greater hormonal responsiveness and modulate their functional state with cycle phase . The 507 stability of these site -dependent signatures echoes observations from other biological contexts 508 demonstrating that the local tissue microenvironment is the dominant determinant of cell state, as 509 has been established for specific tissue-resident macrophages76,77 and cross-tissue populations of 510 other cell types more broadly20. Our findings extend this principle to endometriosis, suggesting 511 that the ectopic tissue niche imprints a durable transcriptomic identity on endometrial -like 512 mesenchymal cells that supersedes hormonal cycling, even as the immune compartment continues 513 to track hormonal state. This compartment split reconciles the dominance of site in subtype 514 discrimination with the hormonally tuned immune programs relevant to cyclical symptom 515 variation. 516 This hormone-independent site specificity has direct implications for biomarker development. A 517 longstanding challenge in endometriosis diagnostics has been the confounding effect of menstrual 518 cycle phase on candidate biomarker performance, necessitating cycle-matched sampling that limits 519 clinical scalability. Our observation that mesenchymal gene signatures for discriminating between 520 subtypes of endometriosis are conserved across hormonal states suggests that robust biomarkers 521 can, in principle, be identified for endometriosis subtyping without requiring cycle -phase 522 stratification. We demonstrated this by showing that cell type-specific signatures derived from all 523 cell types in combined single-cell atlases are sufficient to classify disease status in independent 524 bulk tissue microarray data for both the eutopic endometrium and the peritoneum, while tissue -525 specific signatures can only effectively distinguish lesion subtypes. These complementary scales 526 of molecular identity may correspond to distinct diagnostic modalities. 527 The plasma cell -free transcriptome has been shown to represent a linear combination of tissue -528 specific RNA contributions 62,78, with recent work demonstrating that cell -type-of-origin can be 529 resolved from circulating cell -free RNA at single -cell resolution 62. High tissue -specific 530 classification accuracy for lesion subtype in our analyses suggests that peripheral blood cell -free 531 mRNA could, in principle, be leveraged to identify and classify endometriosis subtypes non -532 invasively, as has been demonstrated for cancer detection and subtyping 79. In parallel, 533 endometriosis, by virtue of its transcriptomic signatures in the eutopic endometrium , is uniquely 534 amenable to diagnostic approaches based on menstrual effluent, a non -invasive medium that has 535 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 20 shown promise for endometriosis biomarker discovery 18,80,81. Given that the cell type -specific 536 signatures are sufficient to distinguish diseased from healthy bulk eutopic endometrium, the cell 537 type-specific fraction of cell -free mRNA within menstrual effluent represents a potential avenue 538 for future diagnostic development. 539 In summary, this study provides a multi -tissue, multi -condition single -cell reference atlas for 540 endometriosis that enables deconvolution of endometrial -like cells from their host tissue 541 microenvironment. We demonstrate that ectopic site, rather than hormonal state, is the dominant 542 driver of transcriptomic heterogeneity among the mesenchymal core lesion cells. The cell type - 543 and tissue-specific signatures identified here are sufficient to stratify disease status and lesion site 544 in independent bulk data, establishing a molecular framework that can be extended to non-invasive 545 diagnostics through peripheral blood cell -free transcriptomics and menstrual effluent analysis. 546 More broadly, this re -analysis paradigm of deconvolving spatial and cell type -specific disease 547 signatures from heterogeneous tissue samples is generalizable to any disease with a resident tissue 548 component, and offers a scalable approach for robust classifier development across heterogeneous 549 patient populations. Given that site -dependent transcriptomic signatures persist regardless of 550 hormonal state, the retrograde menstruation model gains further support from this analysis: lesions 551 at different anatomical sites maintain distinct molecular identities that reflect a combination of 552 their originating tissue and their local tissue context. 553

554 Single cell transcriptomics data integration and cell typing 555 We collected processed scRNA -seq datasets from previously published studies on human 556 endometrial tissue, human ovary tissue, and human endometriosis lesions. These datasets included: 557 the Human Endometrial Cell Atlas (Mareckova et al)17, Tabula Sapiens 2.0 Ovary Data20, Jones et 558 al19, Fonseca et al endometriosis lesion cells and unaffected ovary samples 15, Tan et al 559 endometriosis lesion cells and peritoneal cells16, and menstrual effluent single cell data from Shih 560 et al 18. The Human Endometrial Cell Atlas also included endometrial biopsy data from the 561 following studies: Wang et al 12, Garcia-Alonso et al 82, Tan et al 16, Lai et al 83, Fonseca et al 15, 562 Huang et al84. For the endometriosis lesions datasets only studies which included both peritoneal 563 and ovarian endometriosis lesions were included. 564 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 21 We integrated the data from different sources using Harmony85 using both dataset and individual 565 as our batch correction variables. The resulting Harmony -corrected principal components were 566 utilized to estimate a neighbor graph and generate a uniform manifold approximation and 567 projection (UMAP) visualization. Following embedding of the samples on a UMAP, we followed 568 the AnnDictionary automated cell type labeling pipeline21 using Claude Sonnet 3.5. This gave us 569 preliminary broad cell type labels. These labels, if they contained cell type names which were too 570 broad, such as “T Cell” were further annotated down to CD4, CD8, and T reg identities using the 571 same marker genes used in the HECA atlas. Additional manual cell typing was applied on a 572 lineage-to-lineage basis, where each individual cell type lineage was re -clustered independently 573 and more fine-grained cell typing was performed on the AnnDictionary assigned cell types using 574 literature-based marker gene expression. 575 Somatic Mutation Calling and Analysis 576 Somatic mutations were called de novo from scRNA-seq data from the paired eutopic and ectopic 577 samples found in Fonseca et al 10X 3’ single cell RNA sequencing data using SComatic86. In brief, 578 the raw .fastq files, which were downloaded from GEO accession GSE213216, were processed in 579 Cell Ranger 10.0.0, using GRCh38, and produced aligned BAM files. These files were then split 580 according to their cell type annotation and SComatic was used to pileup bases across individual 581 cell types, compare variants to the reference genome, annotate the variants, and filter them. The 582 variant filtering process removed sites which were present in a panel of normals and also removed 583 known RNA editing sites as described previously86. 584 To assess selection pressure on coding mutations we computed the ratio of nonsynonymous to 585 synonymous mutations (dN/dS) using the amino acid consequences of the observed mutations, by 586 intersecting variant positions with GENCODE CDS annotations from GRCh38. To generate a null 587 expectation under neutrality, we performed a permutation test across 100,000 iterations. In each 588 permutation, at each of the 213 coding variant positions, a random alternative base was sample d 589 with probabilities weighted by the observed trinucleotide spectrum at the mutation’s context. This 590 sampling was performed across all mutation sites followed by the calculation of the dN/dS. 591 Significance was assessed by computing a z -score and a p -value from the standard normal 592 distribution. 593 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 22 To identify shared mutations between eutopic and ectopic samples we used the pooled variants at 594 each patient-tissue pair to identify the same somatic changes. To further identify the clonal cell 595 fraction, the percentage of cells carrying a given mutation, we ran the de novo somatic variant 596 calling on the cell type BAM files independently, without bulking across cells. This resulted in 597 high confidence mutation calling for a select group of cells, these high con fidence single cell 598 mutation calls, were then used to estimate the fractional contribution of the shared mutations to 599 the overall population of cells. The pooled variants across our samples were utilized to call both 600 the variant allele frequency and accumulation of mutations per cell type. 601 Single cell transcriptomics-based endometrium classifier 602 After integrating the full single cell object we trained cell type specific binary classifiers to identify 603 the most endometrium-like cells in the lesions. To train this model we utilized the reference data 604 from the HECA, menstrual effluent, ovarian datasets, and peritoneum data from the Tan et al study. 605 For each annotated cell type we separated our data into the cells from the eutopic endometrium, or 606 menstrual effluent, and those from another tissue source. Cell types were included only when both 607 classes had 200 or greater cells per class. 608 We created ElasticNet classifier models trained on log -normalized gene expression for each cell. 609 Features were selected in a fold -specific manner to avoid information leakage. Within each 610 training split, we performed differential gene expression analyses with a Mann-Whitney-Wilcoxon 611 test comparing the endometrium to all other tissues. We filtered the genes for the classifier by 612 Benjamini Hochberg adjusted p -value, minimum absolute log2 fold -change (1), and minimum 613 detection frequency (20%). To reduce the influence of broadly expressed or ambiguously 614 annotated features, we excluded mitochondrial and ribosomal genes and additional low-specificity 615 gene families and identifiers using a fixed pattern -based filter. We also included the opportunity 616 for a tissue-specificity filtering step for the differentially expressed genes, via a cell type level tau 617 calculation87 across the different tissues . This was computed from tissue -wise mean expression 618 within the training split. 619 Classifiers were ElasticNet models which tuned regularization strength and sparsity by grid search 620 over multiple values of the inverse regularization parameter and the ElasticNet L1 ratio. 621 Performance was evaluated using nested cross-validation with donor-disjoint outer folds to ensure 622 that training and testing sets contained non -overlapping donors. Hyperparameters were selected 623 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 23 within each outer training fold using an inner cross -validation loop, and the inner folds were also 624 donor-aware. To mitigate donor imbalance in some configurations, we applied donor -balanced 625 sample weighting within training folds, reweighting cells inversely proportional to the number of 626 cells contributed by each donor within each class. 627 To make the downstream calls equivalent to a probability, we calibrated predicted probabilities 628 using sigmoid calibration fit on the training data within each outer fold, and then generated 629 calibrated probabilities for the held -out validation split. For each outer validation fold, we 630 quantified discrimination using area under the precision –recall curve and area under the ROC 631 curve. We additionally reported balanced accuracy, F1 score, and confusion matrices using an 80% 632 probability cut off. For each configuration and cell type, we trained a final model on all available 633 cells using the same feature -construction strategy, tuned hyperparameters using donor -aware 634 cross-validation when possible, and serialized the fitted model together with the selected feature 635 list and cross-validation summary. 636 Selection of the final model was performed through a parameter sweep across feature strategies 637 (differential-expression genes with or without tau filtering), feature-set sizes, filtering thresholds, 638 and negative -class definitions. To select a single model per cell type, we aggregated cross -639 validation metrics across runs and ranked candidates. We prioritized configurations meeting a 640 minimum F1-performance threshold of 0.9 in the tau constrained models, using non-tau filtered 641 models as a fallback when the tau-based models did not meet the minimal F1 threshold. Selected 642 runs were consolidated into a final model set and summarized with diagnostic plots comparing 643 performance, calibration, and feature-set size across cell types. 644 Differential Gene Expression 645 Differential expression was performed using a two -part hurdle model (MAST) 88 to identify 646 transcriptional changes within each cell type. For each cell type lesion cells were stratified using 647 the classifier derived probability, lesions were divided into cells above or below a fixed probability 648 threshold of 80%. For comparisons involving eutopic endometrium and lesion cells, reference 649 endometrium was additionally partitioned to just the samples having endometriosis. For severity-650 stratified analyses, samples were categorized into mild or severe disease based on stage groupings, 651 and comparisons were performed within severity strata. 652 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 24 To minimize confounding by menstrual cycle state across donors , differential expression was 653 performed using menstrual cycle stage as a covariate. Cells were grouped into discrete cycle-state 654 bins, and only cycle states represented in both case and control groups were analyzed. Differential 655 expression was then run independently within each cycle-state subset. For each gene, we extracted 656 a single, combined significance value from the hurdle component of the model to test for 657 differential expression between case and control. Effect sizes were reported as log fold changes 658 from the combined model and converted to log2 scale. Multiple testing correction was performed 659 using the Benjamini -Hochberg procedure to control the false discovery rate. For each run, the 660 output table included per -gene log2 fold change with confidence intervals, test statistics and P 661 values, and FDR-adjusted P values. 662 Comparing core and non-core cells to the eutopic endometrium 663 To first compare the core and non -core cells of the lesion cells in different sites to the eutopic 664 endometrium, we performed a stratified analysis by performing differential communication and 665 differential expression analysis. The differential communication analysis was performed with 666 scDiffComm28, in which both the ectopic peritoneal and ovarian core and non -core cells were 667 compared to the eutopic endometrium for a total of 4 separate comparisons. Then using the output 668 tables from scDiffComm we isolated the “FLAT ” interactions, those which were not changing 669 from the eutopic endometrium, with an FDR adjusted p -value of below 0.05 and an odds ratio 670 (OR) of greater than 1. Then to compare the core versus non-core interaction changes within each 671 of our lesion types we subtracted the ORA Score, combined adjusted p-value and log2(OR), of the 672 non-core cells from the core cells to identify stronger FLAT interactions in the core cell or non -673 core cell populations. 674 Additionally, we compared the core and non -core cells to the same cell types in the eutopic 675 endometrium of individuals with endometriosis using the strategy outlined above. We were then 676 able to make cross core and non-core comparisons by looking for unique genes. We also took the 677 Jaccard index between these core -specific gene sets for every cell type across the two separa te 678 ectopic sites. Finally, utilizing gseapy, we searched for the biological process gene ontology terms 679 in the differentially expressed genes for specific lineages across the sites. This was performed, in 680 brief, by taking the union of the different lineage differentially expressed genes, running the gseapy 681 enrich function, and identifying gene programs in which 3 or more genes were present and 682 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 25 removing gene sets that did not contain at least a 10% overlap of gene terms, and removing gene 683 sets which were redundant with 50% or more of overlapping genes. 684 Identifying Hormone Sensitive Gene Modules 685 In order to identify gene modules in endometrial cells which are responsive to different stages of 686 the menstrual cycle and hormone therapy we divided our dataset into tissue-cell type pairs, where 687 tissue is either eutopic, the core endometrial peritoneal lesion cells or the core endometrial ovarian 688 lesion cells. We then tested for differentially expressed genes across hormonal states (Proliferative, 689 Secretory, or Exogenous Hormones) using MAST, with donor included as a fixed effect . Cycle 690 responsive genes were identified by thresholding the genes for a Benjamini-Hochberg adjusted p-691 value of less than 0.05 and a maximum log2 fold change of greater than or equal to one, where the 692 peak-versus-trough fold-change was computed as the difference between the highest- and lowest-693 expressing phases for each gene. 694 To identify recurrent trajectory programs across the hormonal status which span the eutopic and 695 ectopic tissue we pooled all gene -tissue-cell type triplets. The mean expression of the cy cle-696 responsive genes for these triplets was then z-scored and subject to k-means clustering. We swept 697 the number of clusters from 3 to 10, where k=6 yielded the highest silhouette score (0.540). To 698 annotate each cluster with its dominant biological program, we took the union of genes whose 699 triples were assigned to that cluster and ran gene set enrichment analysis for biological processes 700 using gseapy. The resulting term list was filtered to adjusted p -value < 0.05 and greater than or 701 equal to 2 overlapping genes, then a greedy redundancy filter was applied that iteratively selected 702 terms by adjusted p -value while excluding terms whose gene content overlapped previously -703 selected terms by more than 50%. The top three non -redundant terms per cluster were used as 704 cluster labels. 705 For each cell type, cluster-assigned genes were partitioned into mutually-exclusive site categories 706 based on which tissue contexts they were called cycle-responsive in: eutopic-only, EcO-only, EcP-707 only, or shared (assigned in two or more tissues). Per-celltype site-partition counts were visualized 708 as stacked bars. We then identified the clusters which reached a peak or had their minimum in the 709 Secretory phase of the menstrual cycle. For each cluster group, we identified genes whose triples 710 were assigned to that cluster in at least one ectopic tissue (EcO or EcP) but not in eutopic , the 711 ectopic-only gene set per cell type. Ectopic -only genes from EcO and EcP were pooled per cell 712 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 26 type, and gene set enrichment analysis was run separately for the secretory -peak (C3) and 713 secretory-trough (C1 or C6) pooled gene sets. Cell types with fewer than two genes in the ectopic-714 only set were excluded from enrichment. 715 Core Endometrial Cell Modules 716 To identify co -expressed gene modules across the core endometrial cells of both ectopic sites , 717 consensus non-negative matrix factorization (cNMF) was applied to the raw gene expression count 718 matrix across matrix ranks ranging from 10 to 55, with 1000 iterations per K. The optimal number 719 of factors, K = 33, was determined by evaluating the trade -off between reconstruction error and 720 stability across K values. All factorization outputs from the 10 00 iterations at K = 33 were 721 aggregated and subjected to density filtering, yielding 33 robust gene modules. 722 cNMF Analysis 723 To interpret cNMF programs, we performed pathway enrichment on the top program genes using 724 gseapy. For each cNMF factor, the top 100 genes with the highest weights were selected for gene 725 ontology (GO) analysis. For 100 genes in each cNMF factor, pathways with an adjusted p-value < 726 0.01 were retained. Among these, up to 10 pathways were selected per factor , a candidate term 727 was discarded if more than 25% of its overlapping genes were shared with any single higher -728 ranked term retained for the same factor. We further enforced global non -redundancy by 729 discarding terms whose overlapping gene sets shared more than 25% of their genes with a term 730 already retained for a different factor. Because these filters operate on leading-edge gene sets rather 731 than on term identity, the same GO term may be retained for more than one factor when the 732 underlying genes are distinct. This procedure returned 47 terms across 26 of the 33 factors, with 733 one to four terms per factor. Where a factor was annotated with multiple terms, the term with the 734 lowest adjusted P value was used as its representative label. The complete mapping is provided in 735 Supplementary Data Table 1. Annotated terms and associated gene sets were compiled into a 736 program-to-pathway mapping used for labeling downstream visualizations and summarization. 737 To summarize cell -type specificity, each cell's program usage vector was normalized to sum to 738 one across all 33 programs. We then computed mean normalized usage of each annotated program 739 within each of 19 cell types and z-scored these means across cell types within each program (Figure 740 3c). Each program was assigned to the cell type at which its z -scored usage was maximal. For 741 downstream analysis we retained only programs whose usage was specific to a single cell type, 742 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 27 defined as a margin of at least 0.5 standard deviations between the highest and second-highest cell-743 type z-score; 22 of the 26 annotated programs met this criterion. Each cell type was allocated a 744 number of programs equal to the number of specific programs assigned to it. Three cell types had 745 no specific program and were instead allocated the annotated program with the highest raw mean 746 usage within that cell type; for two of these, the resulting program was already assigned to another 747 cell type and is shown under both. This yielded 25 cell -type-program pairs spanning 23 unique 748 programs (Supplementary Data Table 2). Because specificity is defined by relative enrichment 749 across cell types whereas the fallback rule uses absolute usage within a cell type, one program is 750 placed with stromal cells in Figure 3c and with glandular epithelial cells in Figure 3d. 751 To compare program activity across biological contexts, we defined condition labels as the 752 combination of lesion site class (EcO or EcP) and cycle -state bin. Within each cell type and 753 condition, we computed mean normalized usage of each program across the cells of the 754 corresponding cell type within each of the six conditions and plotted as raw mean usage. We tested 755 for differences in program usage between lesion sites while adjusting for hormonal state and 756 accounting for within-donor correlation. For each cell-type- program pair, we fit an ordinary least 757 squares model predicting per-cell program usage from lesion site and hormonal state (usage ~ site 758 + state) on the cells of that cell type, using cluster -robust standard errors with donor as the 759 clustering variable. Because the model is additive, the site coefficient estimates a single state-760 adjusted site effect; the reported contrast is EcO minus EcP, with positive values indicating higher 761 usage in EcO . Note that clustering adjusts standard errors for within -donor correlation but does 762 not remove donor-level confounding of the point estimates, as donor is nested within lesion site . 763 Multiple-hypothesis correction was performed using the Benjamini-Hochberg procedure across all 764 real P-values and results are contained within Supplementary Data Table 3. 765 To quantify global transcriptional differences between conditions while accounting for cell -type 766 composition, we computed pairwise distances in program -usage space using a cell -type–aware 767 centroid approach. For each cell type and condition, we computed the centroid of selected program 768 usage vectors (mean across cells) and calculated pairwise condition distances within each cell type 769 using mean absolute difference across program dimensions. We then aggregated distances across 770 cell types using the mean values across the conditions. To summarize site differences within each 771 cycle state, we extracted the cell type specific distances between matched site pairs within each 772 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 28 state and compared distances across states using paired Wilcoxon signed -rank tests across cell 773 types, followed by Benjamini Hochberg correction. 774 Hierarchical clustering of hormonally controlled DEGs 775 To assess the ability of differentially expressed genes in one hormonal state to stratify lesion 776 subtype in other states we first removed the Fonseca et al dataset from the overall single cell object. 777 We then performed differential expr ession analysis on the core lesion cells in the peritoneal 778 endometriosis samples against the core lesion cells in the endometrioma samples in the Tan et al 779 hormonally controlled samples. We then restrict the differentially expressed genes using an 780 adjusted p-value threshold of 0.01 and a log2fc threshold of 0. 75. Amongst these genes for each 781 of our mesenchymal cell types, we isolated the top 15 positive and negative direction genes for 782 each of the 4 cell types, including Stromal Cells, Vascular Smooth Muscle Cells, Fibroblasts, and 783 Myofibroblast like cells. We then observed the expression of these genes on the scaled log 784 normalized expression data from the Fonseca et al core lesion cells across a multitude of menstrual 785 cycle stages. We then applied hierarchical clustering to the pseudobulk expression for donor -786 tissue-celltype groups and visualized the dendrogram and the heatmap together. 787 Creation of tissue and cell type specific gene sets 788 To compare differential expression patterns across both disease severity and tissue groupings we 789 calculated differential gene expression between reference ovarian tissue and our non -core 790 endometrioma cells, then between reference peritoneum tissue and our non-core peritoneal lesion 791 cells, and between the eutopic endometrium of patients with endometriosis and the core 792 endometrioma or peritoneal lesion cells. If a gene for a comparison had an adjusted P value of less 793 than 0.05 it was kept for effect size analysis. To identify endometrioma lineage based effect sizes, 794 we grouped the individual cell types in both endometrioma based comparisons by lineage and 795 calculated the 250 largest magnitude differential expression values. This process was repeated for 796 both severity classes and the same was repeated for the peritoneal based differential expression 797 values. The lineage level comparisons were performed using a Mann-Whitney-Wilcoxon test with 798 a BH correction of the P values. 799 To assess whether severity strata shared similar transcriptional responses, we filtered the 800 differentially expressed genes to only include those with a log2 fold change of magnitude 1 or 801 greater and computed the intersection between the mild and severe differential expression genes 802 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 29 for each cell type in each comparison in a direction aware manner. We then took the union of the 803 DEG sets across all cell types within each severity grouping and plotted one venn diagram for each 804 tissue comparison to show a broad mild to severe gene overlap. 805 To interpret the ability of the intersecting genes to be deconvolved in a non -invasive manner we 806 quantified the number of tissue enriched/enhanced genes (HPA)26 which also had a Gini coefficient 807 of >0.6, and cell type specific gene sets as defined by the Gini coefficient of >0.8 for differentially 808 expressed genes between cell types in Tabula Sapiens v.2.0 as described previously62. For tissue 809 specificity, we only included genes which were specific for the endometrium or ovaries, and the 810 peritoneum specific set was approximated as being tissue specific genes for tissues in the 811 peritoneal cavity, such as the bladder, fallopian tubes, adipose tissue, and intestines . For each 812 intersection, we recorded whether the gene was cell type specific for any of its mapped cell types. 813 To directly compare tissue and cell type specificity within the shared mild to severe genes we 814 classified each gene as tissue or cell type specific. For each comparison, we visualized category 815 proportions using pie charts. We then saved the union of the cell type specific gene across all of 816 our comparisons, we also saved the union of tissue specific genes across all comparisons for 817 downstream analysis. 818 Machine Learning Classification of Endometriosis Lesions 819 Micro-array data was obtained from the Gene Expression Omnibus (GEO accession: 820 GSE141549)59, comprising batch-corrected, normalized microarray expression profiles from 205 821 samples across seven tissue categories: control endometrium, patient endometrium, control 822 peritoneum, patient peritoneum, deep infiltrating endometriosis, peritoneal lesions, and ovarian 823 endometrioma. We employed a patient-aware data partitioning approach during model evaluation, 824 making sure no individual donor was included in both the training and the validation samples. 825 We created three separate models for every comparison, a tissue -enriched gene set model 826 (described above), a cell type specific gene set model (described above), and a combined gene set 827 model. Genes not present in the expression dataset were excluded, and duplicate gene identifiers 828 were resolved by retaining the first occurrence. We first calculated the first and second princip al 829 components of the micro -array data for the three separate gene sets and observed visualization 830 across the different tissues present. 831 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 30 We then created machine learning models for each of our gene sets, using a Random Forest 832 classifier, to distinguish between (1) control and patient endometrium, (2) control and patient 833 peritoneum, and (3) between deep infiltrating endometriosis, peritoneal lesions, and ovarian 834 endometrioma. To train the model features were standardized to zero mean and unit variance prior 835 to model fitting. Hyperparameters were tuned via grid search and the class weights were also 836 balanced to account for unequal class frequencies. Nested cross validation was performed using 837 stratified group k -fold cross -validation (k=5) on both the inner and out er folds, ensuring all 838 samples from a given patient remained within the same fold. Hyperparameter selection was 839 performed via nested cross -validation where the inner loop was used for identifying optimal 840 parameters and the outer loop evaluation providing performance estimates. 841 Model performance was assessed using the area under the receiver operating characteristic curve 842 (ROC-AUC) for the out -of-fold prediction generated from the nested cross -validation. For our 843 multi-class classifier a macro-averaged one-versus-rest ROC curve was computed for every gene 844 set. Confidence intervals for ROC curves and AUC values were estimated via block bootstrapping 845 with the patient as the resampling unit. In short, for each of the 1,000 bootstrap iterations, N 846 patients were drawn with replacement from the N unique patients in the analysis, and all out -of-847 fold predictions belonging to the sampled patients were aggregated to form a bootstrap sample (so 848 a patient drawn twice contributed all of their samples twice). The 2.5th and 97.5th percentiles of 849 the bootstrap distribution defined the confidence bounds . The coefficients of a Random Forest 850 model trained across the entire dataset were then used to rank feature importance. 851 Data Availability 852 All data utilized in this experiment was from publicly available repositories which did not require 853 approval. The HECA was downloaded from a web portal which originated from its parent 854 publication: https://www.reproductivecellatlas.org/endometrium_reference.html. Shih et al 855 menstrual effluent data was available through National Center for Biotechnology 856 Information/Gene Expression Omnibus (GEO) accession number GSE203191. Tan et al lesion 857 and peritoneal data were available through a web portal: 858 https://singlecell.jax.org/datasets/endometriosis-2022. Fonseca et al lesion and ovary data w ere 859 available through GEO accession number GSE213216. Jones et al single cell ovary data was 860 available through GEO accession number GSE260685. Tabula Sapiens 2.0 could be accessed 861 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 31 through the CellxGene Web portal: https://cellxgene.cziscience.com/collections/e5f58829-1a66-862 40b5-a624-9046778e74f5. 863 Code Availability 864 All code for the work and the manuscript are available on GitHub at: 865 https://github.com/doughenze/endometriosis_sites. 866 Acknowledgments 867 The authors thank all members of the Quake Lab for valuable feedback and discussions throughout 868 the study. 869 Author Information 870 Department of Bioengineering, Stanford University, Stanford, CA, USA 871 Douglas E. Henze, George Crowley, and Stephen R. Quake 872 Department of Applied Physics, Stanford University, Stanford, CA, USA 873 Stephen R. Quake 874 Author’s Contribution 875 D.E.H, G.C., and S.R.Q conceptualized and designed experiments. D.E.H collected data. D.E.H 876 and G.C analyzed the data. D.E.H and S.R.Q discussed and interpreted the data. D.E.H designed 877 the figures. D.E.H and S.R.Q wrote the manuscript. D.E.H and S.R.Q supervised, directed, and 878 managed the study. All authors discussed the results and commented on the manuscript. 879 Competing Interest 880 The authors declare no competing interests. 881 Correspondence 882 Correspondence to Stephen R. Quake 883 884 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 32

885 1. Zondervan, K. T., Becker, C. M. & Missmer, S. A. Endometriosis. N. Engl. J. Med. 382, 886 1244–1256 (2020). 887 2. Giudice, L. C. Endometriosis. N. Engl. J. Med. 362, 2389–2398 (2010). 888 3. Taylor, H. S. Reimagining Endometriosis. Med 2, 481–485 (2021). 889 4. Zondervan, K. T. et al. Endometriosis. Nat. Rev. Dis. Primer 4, 9 (2018). 890 5. American Society For Reproductive Medicine. Revised American Society for 891 Reproductive Medicine classification of endometriosis: 1996. Fertil. Steril. 67, 817–821 892 (1997). 893 6. Nisolle, M. & Donnez, J. Peritoneal endometriosis, ovarian endometriosis, and 894 adenomyotic nodules of the rectovaginal septum are three different entities. Fertil. 895 Steril. 68, 585–596 (1997). 896 7. Bulun, S. E. et al. Endometriosis. Endocr. Rev. 40, 1048–1079 (2019). 897 8. Zeitoun, K. M. & Bulun, S. E. Aromatase: a key molecule in the pathophysiology of 898 endometriosis and a therapeutic target. Fertil. Steril. 72, 961–969 (1999). 899 9. Vercellini, P., Viganò, P., Somigliana, E. & Fedele, L. Endometriosis: pathogenesis and 900 treatment. Nat. Rev. Endocrinol. 10, 261–275 (2014). 901 10. La Marca, A., Carducci Artenisio, A., Stabile, G., Rivasi, F. & Volpe, A. Evidence for 902 cycle-dependent expression of follicle-stimulating hormone receptor in human 903 endometrium. Gynecol. Endocrinol. Off. J. Int. Soc. Gynecol. Endocrinol. 21, 303–306 904 (2005). 905 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 33 11. Sapkota, Y. et al. Meta-analysis identifies five novel loci associated with endometriosis 906 highlighting key genes involved in hormone metabolism. Nat. Commun. 8, 15539 907 (2017). 908 12. Wang, W. et al. Single-cell transcriptomic atlas of the human endometrium during the 909 menstrual cycle. Nat. Med. 26, 1644–1653 (2020). 910 13. Talbi, S. et al. Molecular phenotyping of human endometrium distinguishes menstrual 911 cycle phases and underlying biological processes in normo-ovulatory women. 912 Endocrinology 147, 1097–1121 (2006). 913 14. Liu, S. et al. Single-cell and spatial transcriptomic profiling revealed niche interactions 914 sustaining growth of endometriotic lesions. Cell Genomics 5, 100737 (2025). 915 15. Fonseca, M. A. S. et al. Single-cell transcriptomic analysis of endometriosis. Nat. 916 Genet. 55, 255–267 (2023). 917 16. Tan, Y. et al. Single-cell analysis of endometriosis reveals a coordinated transcriptional 918 programme driving immunotolerance and angiogenesis across eutopic and ectopic 919 tissues. Nat. Cell Biol. 24, 1306–1318 (2022). 920 17. Marečková, M. et al. An integrated single-cell reference atlas of the human 921 endometrium. Nat. Genet. 56, 1925–1937 (2024). 922 18. Shih, A. J. et al. Single-cell analysis of menstrual endometrial tissues defines 923 phenotypes associated with endometriosis. BMC Med. 20, 315 (2022). 924 19. Jones, A. S. K. et al. Cellular atlas of the human ovary using morphologically guided 925 spatial transcriptomics and single-cell sequencing. Sci. Adv. 10, eadm7506 (2024). 926 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 34 20. Quake, S. R. & The Tabula Sapiens Consortium. Tabula Sapiens reveals transcription 927 factor expression, senescence effects, and sex-specific features in cell types from 28 928 human organs and tissues. Preprint at https://doi.org/10.1101/2024.12.03.626516 929 (2024). 930 21. Crowley, G. et al. Benchmarking cell type and gene set annotation by large language 931 models with AnnDictionary. Nat. Commun. 16, 9511 (2025). 932 22. Sampson, J. A. Metastatic or Embolic Endometriosis, due to the Menstrual 933 Dissemination of Endometrial Tissue into the Venous Circulation. Am. J. Pathol. 3, 93-934 110.43 (1927). 935 23. Moore, L. et al. The mutational landscape of normal human endometrial epithelium. 936 Nature 580, 640–646 (2020). 937 24. Ólafsson, S. et al. Endometriosis lesions are oligoclonal structures derived from the 938 normal endometrium. Preprint at https://doi.org/10.64898/2026.02.25.708037 (2026). 939 25. Suda, K. et al. Different mutation profiles between epithelium and stroma in 940 endometriosis and normal endometrium. Hum. Reprod. 34, 1899–1905 (2019). 941 26. Uhlén, M. et al. Proteomics. Tissue-based map of the human proteome. Science 347, 942 1260419 (2015). 943 27. Henlon, Y. et al. Single-cell analysis identifies distinct macrophage phenotypes 944 associated with prodisease and proresolving functions in the endometriotic niche. 945 Proc. Natl. Acad. Sci. U. S. A. 121, e2405474121 (2024). 946 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 35 28. Lagger, C. et al. scDiffCom: a tool for differential analysis of cell–cell interactions 947 provides a mouse atlas of aging changes in intercellular communication. Nat. Aging 3, 948 1446–1461 (2023). 949 29. Burns, G. W. et al. Spatial transcriptomic analysis identifies epithelium-macrophage 950 crosstalk in endometriotic lesions. iScience 28, 111790 (2025). 951 30. Greaves, E. et al. Estradiol is a critical mediator of macrophage-nerve cross talk in 952 peritoneal endometriosis. Am. J. Pathol. 185, 2286–2297 (2015). 953 31. Ono, Y. et al. CD206+ macrophage is an accelerator of endometriotic-like lesion via 954 promoting angiogenesis in the endometriosis mouse model. Sci. Rep. 11, 853 (2021). 955 32. Bacci, M. et al. Macrophages are alternatively activated in patients with endometriosis 956 and required for growth and vascularization of lesions in a mouse model of disease. 957 Am. J. Pathol. 175, 547–556 (2009). 958 33. Wu, M.-H., Hsiao, K.-Y. & Tsai, S.-J. Endometriosis and possible inflammation markers. 959 Gynecol. Minim. Invasive Ther. 4, 61–67 (2015). 960 34. Dawood, M. Y. Primary dysmenorrhea: advances in pathogenesis and management. 961 Obstet. Gynecol. 108, 428–441 (2006). 962 35. Stratton, P. & Berkley, K. J. Chronic pelvic pain and endometriosis: translational 963 evidence of the relationship and implications. Hum. Reprod. Update 17, 327–346 964 (2011). 965 36. ACOG Committee Opinion No. 760: Dysmenorrhea and Endometriosis in the 966 Adolescent. Obstet. Gynecol. 132, e249–e258 (2018). 967 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 36 37. Begum, I. A. The connection between endometriosis and secondary dysmenorrhea. J. 968 Reprod. Immunol. 168, 104425 (2025). 969 38. Deuster, P. A., Dolev, E., Bernier, L. L. & Trostmann, U. H. Magnesium and zinc status 970 during the menstrual cycle. Am. J. Obstet. Gynecol. 157, 964–968 (1987). 971 39. Bruner, K. L., Eisenberg, E., Gorstein, F. & Osteen, K. G. Progesterone and transforming 972 growth factor-β coordinately regulate suppression of endometrial matrix 973 metalloproteinases in a model of experimental endometriosis. Steroids 64, 648–653 974 (1999). 975 40. Obeagu, E. I. & Obeagu, G. U. From menses to ovulation: the immunomodulatory role 976 of monocyte cytokines in menstrual cycle physiology and reproductive health - a 977 narrative review. Ann. Med. Surg. 87, 8453–8459 (2025). 978 41. Wira, C. R., Rodriguez-Garcia, M. & Patel, M. V. The role of sex hormones in immune 979 protection of the female reproductive tract. Nat. Rev. Immunol. 15, 217–230 (2015). 980 42. Attia, G. R. et al. Progesterone receptor isoform A but not B is expressed in 981 endometriosis. J. Clin. Endocrinol. Metab. 85, 2897–2902 (2000). 982 43. Burney, R. O. et al. Gene expression analysis of endometrium reveals progesterone 983 resistance and candidate susceptibility genes in women with endometriosis. 984 Endocrinology 148, 3814–3826 (2007). 985 44. Vissers, G., Giacomozzi, M., Verdurmen, W., Peek, R. & Nap, A. The role of fibrosis in 986 endometriosis: a systematic review. Hum. Reprod. Update 30, 706–750 (2024). 987 45. Gong, H. et al. The regulation of ovary and conceptus on the uterine natural killer cells 988 during early pregnancy. Reprod. Biol. Endocrinol. 15, 73 (2017). 989 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 37 46. Hogg, C. et al. Macrophages inhibit and enhance endometriosis depending on their 990 origin. Proc. Natl. Acad. Sci. 118, e2013776118 (2021). 991 47. Rocha, A. L. L., Reis, F. M. & Taylor, R. N. Angiogenesis and endometriosis. Obstet. 992 Gynecol. Int. 2013, 859619 (2013). 993 48. Reichelt, U. et al. High lymph vessel density and expression of lymphatic growth 994 factors in peritoneal endometriosis. Reprod. Sci. 19, 876–882 (2012). 995 49. Zhang, Q., Duan, J., Olson, M., Fazleabas, A. & Guo, S.-W. Cellular Changes Consistent 996 With Epithelial-Mesenchymal Transition and Fibroblast-to-Myofibroblast 997 Transdifferentiation in the Progression of Experimental Endometriosis in Baboons. 998 Reprod. Sci. 23, 1409–1421 (2016). 999 50. Pušić, M. N. et al. Growth Arrest-Specific Protein 6 Is Elevated in Endometriosis but 1000 Shows Poor Diagnostic Performance. Int. J. Mol. Sci. 26, 8348 (2025). 1001 51. Zhai, X. et al. Gas6/AXL pathway: immunological landscape and therapeutic potential. 1002 Front. Oncol. 13, 1121130 (2023). 1003 52. Klemmt, P. A. B., Carver, J. G., Kennedy, S. H., Koninckx, P. R. & Mardon, H. J. Stromal 1004 cells from endometriotic lesions and endometrium from women with endometriosis 1005 have reduced decidualization capacity. Fertil. Steril. 85, 564–572 (2006). 1006 53. Wang, X. et al. The role of insulin-like growth factor binding proteins in TGF-β1-induced 1007 fibroblast-myofibroblast transition during endometriosis fibrosis. Cell. Signal. 140, 1008 112362 (2026). 1009 54. Giudice, L. C., Dsupin, B. A., Gargosky, S. E., Rosenfeld, R. G. & Irwin, J. C. The insulin-1010 like growth factor system in human peritoneal fluid: its effects on endometrial stromal 1011 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 38 cells and its potential relevance to endometriosis. J. Clin. Endocrinol. Metab. 79, 1284–1012 1293 (1994). 1013 55. Clower, L., Fleshman, T., Geldenhuys, W. J. & Santanam, N. Targeting Oxidative Stress 1014 Involved in Endometriosis and Its Pain. Biomolecules 12, 1055 (2022). 1015 56. Harrington, D. J. et al. Tenascin is differentially expressed in endometrium and 1016 endometriosis. J. Pathol. 187, 242–248 (1999). 1017 57. Ma, R. et al. Identification of key genes associated with endometriosis and endometrial 1018 cancer by bioinformatics analysis. Front. Oncol. 14, 1387860 (2024). 1019 58. Yotova, I. et al. LINC01638 promotes epithelial-to-mesenchymal transition in 1020 endometriosis epithelial cells by up-regulating RHOB via HDAC1 suppression. Reprod. 1021 Biomed. Online 51, 104942 (2025). 1022 59. Ganieva, U. et al. Involvement of Transcription Factor 21 in the Pathogenesis of Fibrosis 1023 in Endometriosis. Am. J. Pathol. 190, 145–157 (2020). 1024 60. Lee, M.-Y. et al. Role of interleukin-32 in the pathogenesis of endometriosis: in vitro, 1025 human and transgenic mouse data. Hum. Reprod. 33, 807–816 (2018). 1026 61. Rahmioglu, N. et al. Genetic variants underlying risk of endometriosis: insights from 1027 meta-analysis of eight genome-wide association and replication datasets. Hum. 1028 Reprod. Update 20, 702–716 (2014). 1029 62. Vorperian, S. K. et al. Cell types of origin of the cell-free transcriptome. Nat. 1030 Biotechnol. 40, 855–861 (2022). 1031 63. Moufarrej, M. N. et al. Early prediction of preeclampsia in pregnancy with cell-free RNA. 1032 Nature 602, 689–694 (2022). 1033 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 39 64. Gabriel, M. et al. A relational database to identify differentially expressed genes in the 1034 endometrium and endometriosis lesions. Sci. Data 7, 284 (2020). 1035 65. Li, Q. et al. Identification of Candidate Gene Signatures and Regulatory Networks in 1036 Endometriosis and its Related Infertility by Integrated Analysis. Reprod. Sci. 29, 411–1037 426 (2022). 1038 66. Donnez, J. & Cacciottola, L. Endometriosis: An Inflammatory Disease That Requires 1039 New Therapeutic Options. Int. J. Mol. Sci. 23, 1518 (2022). 1040 67. Lousse, J.-C. et al. Peritoneal endometriosis is an inflammatory disease. Front. Biosci. 1041 4, 23–40 (2012). 1042 68. Ferrero, S., Vellone, V. G. & Barra, F. Pathophysiology of pain in patients with peritoneal 1043 endometriosis. Ann. Transl. Med. 7, S8 (2019). 1044 69. Nisenblat, V. et al. Blood biomarkers for the non-invasive diagnosis of endometriosis. 1045 Cochrane Database Syst. Rev. 2016, CD012179 (2016). 1046 70. Halpern, K. B. et al. Single-cell spatial reconstruction reveals global division of labour 1047 in the mammalian liver. Nature 542, 352–356 (2017). 1048 71. Moor, A. E. et al. Spatial Reconstruction of Single Enterocytes Uncovers Broad Zonation 1049 along the Intestinal Villus Axis. Cell 175, 1156-1167.e15 (2018). 1050 72. Suda, K. et al. Clonal Expansion and Diversification of Cancer-Associated Mutations in 1051 Endometriosis and Normal Endometrium. Cell Rep. 24, 1777–1789 (2018). 1052 73. Anglesio, M. S. et al. Cancer-Associated Mutations in Endometriosis without Cancer. 1053 N. Engl. J. Med. 376, 1835–1848 (2017). 1054 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 40 74. Murray, P. J. & Wynn, T. A. Protective and pathogenic functions of macrophage subsets. 1055 Nat. Rev. Immunol. 11, 723–737 (2011). 1056 75. Donnez, J. et al. Peritoneal endometriosis and “endometriotic” nodules of the 1057 rectovaginal septum are two different entities. Fertil. Steril. 66, 362–368 (1996). 1058 76. Lavin, Y. et al. Tissue-resident macrophage enhancer landscapes are shaped by the 1059 local microenvironment. Cell 159, 1312–1326 (2014). 1060 77. Gautier, E. L. et al. Gene-expression profiles and transcriptional regulatory pathways 1061 that underlie the identity and diversity of mouse tissue macrophages. Nat. Immunol. 1062 13, 1118–1128 (2012). 1063 78. Koh, W. et al. Noninvasive in vivo monitoring of tissue-specific global gene expression 1064 in humans. Proc. Natl. Acad. Sci. 111, 7361–7366 (2014). 1065 79. Larson, M. H. et al. A comprehensive characterization of the cell-free transcriptome 1066 reveals tissue- and subtype-specific biomarkers for cancer detection. Nat. Commun. 1067 12, 2357 (2021). 1068 80. Warren, L. A. et al. Analysis of menstrual effluent: diagnostic potential for 1069 endometriosis. Mol. Med. 24, 1 (2018). 1070 81. Wilson, T. R. et al. Analysis of menstrual effluent uncovers endometriosis-specific cell 1071 populations and impaired cellular pathway processes. Preprint at 1072 https://doi.org/10.1101/2025.08.21.671582 (2025). 1073 82. Garcia-Alonso, L. et al. Mapping the temporal and spatial dynamics of the human 1074 endometrium in vivo and in vitro. Nat. Genet. 53, 1698–1711 (2021). 1075 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 41 83. Lai, Z.-Z. et al. Single-cell transcriptome profiling of the human endometrium of 1076 patients with recurrent implantation failure. Theranostics 12, 6527–6547 (2022). 1077 84. Huang, X. et al. Single-cell transcriptome analysis reveals endometrial immune 1078 microenvironment in minimal/mild endometriosis. Clin. Exp. Immunol. 212, 285–295 1079 (2023). 1080 85. Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with 1081 Harmony. Nat. Methods 16, 1289–1296 (2019). 1082 86. Muyas, F. et al. De novo detection of somatic mutations in high-throughput single-cell 1083 profiling data sets. Nat. Biotechnol. 42, 758–767 (2024). 1084 87. Yanai, I. et al. Genome-wide midrange transcription profiles reveal expression level 1085 relationships in human tissue specification. Bioinformatics 21, 650–659 (2005). 1086 88. Finak, G. et al. MAST: a flexible statistical framework for assessing transcriptional 1087 changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome 1088 Biol. 16, 278 (2015). 1089 1090 1091 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 42 Figures 1092 1093 1094 1095 Figure 1: Integrated cell atlas across endometriosis related tissues 1096 A, Schematic overview of the data analysis pipeline, building references and training classifiers 1097 on reference tissue and applying them to lesions. B, Donor distribution, number of unique donors 1098 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 43 for every tissue-dataset pair. C, UMAP plot across the different tissues colored by broad cell 1099 type. D, Marker gene expression across the broad cell types. 1100 1101 1102 1103 1104 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 44 Figure 2: Tissue-of-origin classifiers reveal heterogeneous endometrial-like and host tissue 1105 populations within ectopic lesions. 1106 A, Grouped bar plot showing the relative performance amongst each of the cell type-specific 1107 classifiers. Sorted by macro F1. B, Percentage of lesion cells for each lineage which mapped to 1108 being endometrium-like (green) or not (gray). C, Empirical cumulative distribution function to 1109 show the relative percent of differentially expressed genes from the eutopic endometrium; lines 1110 are colored by site and core or non-core category. D, Ranked Jaccard Indices between the core-1111 specific differentially expressed genes for each of the cell types across the endometrioma and 1112 peritoneal endometriosis samples E, Gene ontology biological processes terms for the 1113 differentially expressed genes from the eutopic endometrium for myeloid and lymphoid cells 1114 within each of the sites. 1115 1116 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 45 1117 1118 Figure 3: Transcriptomic programs are more associated with site than hormonal status 1119 A, Schematic image of the primary axes of lesion cell variation being hormonal or spatial. B, 1120 Dumbbell plot showing the spearman correlation across sites in matched hormone cases or 1121 across hormones in matched site cases, with line meaning perfect agreement. C, Diagonalized 1122 heatmap showing the z-scored usage of different NMF modules across cell types. D, Bar plot 1123 showing the mean usage for each site-hormone pair across all cell types. * represents a 1124 statistically significant deviation between sites via a linear mixed model. 1125 1126 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 46 1127 Figure 4: Severity-associated cell type- and tissue-specific signatures classify disease status 1128 and lesion origin. 1129 A, PCA embedding of combined cell type- and tissue-specific gene signature in the bulk tissue 1130 micro array data. B, Performance of each of the different gene sets to diagnose bulk eutopic 1131 endometrium samples as having endometriosis or not as shown via a receiver operator 1132 characteristic (ROC) curve. C, Performance of each of the different gene sets to diagnose bulk 1133 peritoneum samples as having endometriosis or not as shown via a ROC curve. D, Performance 1134 of each of the different gene sets to stratify lesions by tissue of origin as shown via a ROC curve. 1135 E, Log normalized expression of PDLIM3 and SERPINE2 across all groups. Statistics shown are 1136 from a Wilcoxon rank sum test with a Benjamini-Hochberg correction across all possible 1137 comparisons for a given gene. * is for a corrected p-value below 0.05, ** is for a corrected p-1138 value below 0.01, and *** is for a corrected p-value below 0.001. All AUC curves show the 1139 calculated curve and a 95% confidence interval. 1140 1141 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 47 Supplementary Figures 1142 1143 1144 1145 Supplementary Figure 1: Semi-automated cell typing recovers previously identified cell 1146 populations 1147 A, Number of transcripts per cell, line on the violin plot indicates the median value. B, Number 1148 of genes per cell, line on the violin plot indicates the median value. C, Cluster map showing the 1149 relative percentage of HECA labels for each of our semi-automated pipeline created labels. D, 1150 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 48 UMAP representation of all cells colored by tissue. E, UMAP representation of all cells colored 1151 by Dataset. F, UMAP representation of all cells colored by endometriosis stage. 1152 1153 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 49 1154 Supplementary Figure 2: Somatic mutation accumulation in eutopic and ectopic 1155 endometrium 1156 A-B, 6-class (A) and 96 class (B) mutation spectra between the eutopic endometrium and the 1157 ectopic endometrium. C, Permutation test results showing expected against null dN/dS for the 1158 observed mutation spectra and mutation sites. D, Number of mutations shared between the 1159 eutopic and ectopic endometrium in a given patient. E, Percentage of cells, bulked across the two 1160 patients with shared mutations, that display the shared mutation in the eutopic (left) and ectopic 1161 (right) endometrium. F, cumulative density functions of the variant allele frequencies across the 1162 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 50 different patients eutopic and ectopic samples. G, Total number of mutations attributable to each 1163 cell type, colored by sample location. 1164 1165 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 51 1166 Supplementary Figure 3: Validation of cell type-level endometrial origin classifiers 1167 A, Confusion matrices of classifier predictions across populous cell types B, Bar plot showing 1168 the number of endometrium tissue-enhanced or tissue enriched HPA genes which were used in 1169 the classifier for each cell type. C, UMAP showing value of applied classifier across the whole 1170 integrated dataset. D-E, Stacked bar plots showing classifier values across different cell lineages 1171 in the (D) endometrioma and the (E) peritoneal endometriosis samples. 1172 1173 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 52 1174 Supplementary Figure 4: Differential cell-cell interaction between the eutopic 1175 endometrium and core or non-core cell populations 1176 A, Heatmap showing the difference in odds ratio scores of preserved cell-cell interactions from 1177 the eutopic endometrium to core and non-core cells in endometrioma. B, Heatmap showing the 1178 difference in odds ratio scores of preserved cell-cell interactions from the eutopic endometrium 1179 to core and non-core cells in peritoneal endometriosis. 1180 1181 1182 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 53 1183 1184 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 54 Supplementary Figure 5: Core- and non-core-specific differential gene expression between 1185 the eutopic endometrium and different lesion sites. 1186 A, Summary of differentially expressed genes as core-specific, non-core-specific, or shared for 1187 endometrioma samples across all cell types. B, Summary of differentially expressed genes as 1188 core-specific, non-core-specific, or shared across peritoneal endometriosis samples across all cell 1189 types. C-E, Gene Ontology Biological Processes for endometrioma macrophage (C) non-core 1190 specific, (D) shared, and (E) core-specific differentially expressed genes. F-H Gene Ontology 1191 Biological Processes for peritoneal endometriosis macrophage (F) non-core specific, (G) shared, 1192 and (H) core-specific differentially expressed genes. 1193 1194 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 55 1195 Supplementary Figure 6: Hormone-responsive gene modules explain endometriosis pain. 1196 A, Barplot showing the number of cycle-dependent differentially expressed genes for each cell 1197 type across the ectopic ovary (EcO), ectopic peritoneum (EcP), and eutopic endometrium 1198 (Eutopic) tissues. B, Hormone responsive cluster profiles identified through k-means clustering 1199 of the mean expression for a tissue, cell type, and gene triplet at each hormonal state. Clusters are 1200 labeled with the top 3 gene ontology terms for the union of all genes present in the cluster. C, 1201 Stacked bar plot representing the number of cycle-responsive genes in Cluster 3 (C3) for EcO, 1202 EcP, Eutopic, and shared between two or more. D, Stacked bar plot representing the number of 1203 cycle-responsive genes in Cluster 1 or Cluster 6 (C1 | C6) for EcO, EcP, Eutopic, and shared 1204 between two or more. E, Bubble plot showing gene ontology biological processes enriched in the 1205 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 56 ectopic only (EcO or EcP) genes for Cluster 3 across each cell type. F, Bubble plot showing gene 1206 ontology biological processes enriched in the ectopic only (EcO or EcP) genes for Cluster 1 or 1207 Cluster 6 across each cell type. 1208 1209 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 57 1210 Supplementary Figure 7: Site-dependent cell type signatures are conserved across the full 1211 spectrum of hormonal conditions. 1212 A, Boxplot showing the euclidean distance between the usage space for each individual cell type 1213 across sites for each hormone state. n.s represents no statistically significant deviation. B, 1214 Clustered heatmap showing z-scored expression of top 15 differentially expressed genes across 1215 mesenchymal cell types in a hormonally controlled dataset as applied to a non-hormonally 1216 controlled dataset. 1217 1218 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 58 1219 1220 Supplementary Figure 8: Severity associated differentially expressed genes contain cell 1221 type- and tissue-specific genes. 1222 A, Boxplot representation of the top 250 genes associated with each lineage across different 1223 severity classifications of endometriosis. B, Venn diagram representation of the differentially 1224 expressed genes split by severity across endometrioma (left) and percentage of severity shared 1225 genes that are non-specific, cell type-specific, tissue specific, or both (right). C, Venn diagram 1226 representation of the differentially expressed genes split by severity across peritoneal 1227 endometriosis (left) and percentage of severity shared genes that are non-specific, cell type-1228 specific, tissue specific, or both (right). 1229 1230 1231 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 59 1232 1233 Supplementary Figure 9: Bulk tissue microarray data can be separated by cell type- and 1234 tissue-specific gene signatures 1235 A-B, PCA projections of bulk tissue microarray data using cell type-specific (A) genes and 1236 tissue-specific (B) genes. 1237 1238 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 60 1239 1240 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint 61 Supplementary Figure 10: Combined cell type- and tissue-specific gene signature 1241 importances across diverse classifiers 1242 A, Combined cell type- and tissue-specific gene signature classifier importances for predicting 1243 control against diseased eutopic endometrium. B, Combined cell type- and tissue-specific gene 1244 signature classifier importances for predicting control against diseased peritoneum. C, Combined 1245 cell type- and tissue-specific gene signature classifier importances for multi-class prediction 1246 across three separate lesion types (deep infiltrating endometriosis, peritoneal endometriosis, and 1247 endometrioma). 1248 1249 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted July 16, 2026. ; https://doi.org/10.64898/2026.07.10.737856doi: bioRxiv preprint