The core PGC MDD49 cohort set builds on the MDD29 sample from Wray et al.7 These cohorts provided individual phenotype and genotype data for quality control, imputation, and analysis. The 49 cohort names, sample sizes, and inclusion/exclusion criteria are summarised in Table S1 and more details can be found in Methods S1. Most cohorts provided both cases and controls. For case-only cohorts, cohorts were either merged or matching controls were obtained from other PGC groups. Cases met international consensus criteria for a lifetime diagnosis of major depressive disorder (ICD9, ICD10, DSM-IV, or DSM-5). Cases were classified into clinical studies (where diagnoses were established using structured diagnostic instruments by clinicians, trained interviewers, or medical record review) and community studies (where structured diagnostic instruments were used for diagnosis of life-time major depression, or self-report of an MDD diagnosis). Controls were either screened for the absence of MDD and other mood disorders or selected randomly from the population.
