W66. TRANS-ANCESTRY AND SEX-STRATIFIED GENOME-WIDE ASSOCIATION META-ANALYSIS OF MAJOR DEPRESSIVE DISORDER DIAGNOSIS IN 500,000 CASES AND 2 MILLION CONTROLS

Background Large-scale genome-wide association studies (GWASs) have accelerated genomic discovery for major depressive disorder (MDD), but important limitations persist. To date, more than 77% of GWAS participants are of European ancestry, limiting transferability to non-European populations. Recent large-scale efforts have increasingly drawn on electronic health records (EHRs), yet inconsistent diagnostic coding across cohorts increases phenotypic heterogeneity and reduces statistical power. Most GWAS meta-analyses have also combined males and females in a single analysis despite converging evidence that MDD has a partially sex-differentiated genetic basis. A recent European-ancestry sex-stratified GWAS meta-analysis of MDD (∼195k cases) reported greater polygenicity in females and female-specific causal variants, but lacked the representation of non-European populations. Thus, the full extent of sex-specific genetic architecture and its downstream biological implications across diverse populations remains largely uncharacterized. Methods Here, the PsycheMERGE Network Diversity Initiative reports a sex-stratified, trans-ancestry GWAS meta-analysis of MDD diagnosis in over 500,000 cases and 2 million controls from 13 EHR-linked biobanks across the US, UK, and Denmark, with 22% of participants genetically similar to non-European reference populations. This work incorporates four key innovations: (1) a uniform EHR-based MDD phenotype applied across cohorts to reduce phenotypic heterogeneity; (2) a trans-ancestry analytic framework that pairs ancestry-stratified meta-analysis with joint trans-ancestry models, controlling for population structure and relatedness while maximizing participant inclusion; (3) sex-stratified GWAS and genome-wide genotype-by-sex interaction analyses, comparing the genetic architecture of MDD between males and females; and (4) integration of ancestry- and sex-stratified cell-type-specific single-nucleus transcriptome-wide association studies, performed using transcriptomic prediction models trained in a large human brain single-cell atlas. Results The genetic architecture of MDD across ancestry groups and between sexes is characterized at the genome-wide and locus levels, alongside ancestry- and sex-specific gene and cell-type associations. The primary phenotype is benchmarked against alternative EHR definitions and prior depression GWAS to assess concordance and specificity. The alternative diverse-ancestry GWAS approaches are also compared for locus discovery and statistical fine-mapping resolution. To our knowledge, these analyses represent the largest sex-stratified trans-ancestry GWAS of MDD, as well as the first ancestry- and sex-stratified, cell-type-specific transcriptomic analysis of MDD GWAS. Discussion By jointly leveraging ancestral diversity and sex stratification, this study disentangles shared, population-specific, and sex-specific genetic signals for MDD, improves fine-mapping through cross-ancestry linkage disequilibrium information, and connects risk loci to cell-type-level biological mechanisms that may contribute to established sex differences in MDD prevalence, symptom profiles, and comorbidity patterns. The analytic framework presented here is designed to be scalable and should serve as a foundation for sex-informed, ancestrally inclusive genomic research in psychiatry and other complex traits.