The population-based UK Biobank cohort 51 consists of 501,726 individuals with genome-wide data for 93,095,623 autosomal genetic variants imputed using the HRC and UK10K reference panels 52. We used the variants from the HRC reference panel to conduct a 4-means clustering approach and used the first two principal components to identify a genetically homogenous subgroup of 462,065 individuals. We next removed 131,790 individuals that had a shared relatedness up to the third degree that were identified by UK Biobank based on kinship coefficients (> 0.044) calculated using the KING toolset 53. For these 131,790 removed individuals we then calculated a genomic relationship matrix and identified one individual to be reinstated from within each related group that had a genetic relatedness less than 0.025 with all other participants which allowed us to add an additional 55,745 individuals back into our sample. We then used a checksum based approach (https://personal.broadinstitute.org/sripke/share_links/checksums_download/) to identify and exclude 954 individuals from within the UK Biobank cohort that overlapped with the Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium (PGC) cohorts analysed by Wray, et
