The MVP cohort has been previously described.38–40 GWAS was conducted in each of two tranches of data separately by ancestry, depending upon when the data became available. Ancestry was assigned using 10 principal components (PCs) and the 1000 genomes project phase 3 EUR and AFR reference within each tranche of data. For the analysis of the quantitative phenotype we also performed a GWAS in the UK Biobank sample. Finally, we conducted GWAS meta-analyses of traits related to depression using data from 4 large cohorts (Table 1, Figure 1 Upper Left): the Million Veteran Program (MVP), 34, 41 the PGC/UK Biobank,10 FinnGen, and 23andMe.9 For the ICD definition of depression, the phenotype with the most available data for the MVP cohort, there were 1,154,267 total subjects for primary meta-analysis. For the secondary case control meta-analysis, we performed a similar analysis except we replaced the MDD diagnosis from MVP with the SR-Depression GWAS for a total of 1,114,383 participants. For the secondary analysis of depressive symptoms by PHQ, we included 286,821 total participants from UKB and MVP. We also performed a GWAS
