Quality control procedures were performed with PLINK43 and EIGENSOFT v4.2.44, 45 using array-based genotype calls. EA and AA subjects were analyzed separately. One subject was retained from each group of related subjects (identity-by-descent > 0.25). Subjects were excluded from the present study if they were ever more than six standard deviations from the mean of principal component (PC) 1, 2 or 3 during five iterations of PC analysis (PCA) in EIGENSOFT v4.2.44, 45 Association testing was performed with PLINK.43 The SNP was coded as the allele dosage, and age, sex and the first 10 PCs (to control for population stratification) were included as covariates. We observed acceptable genomic inflation factor (λ) values for our analyses that included 10 PCs and no additional steps were taken to include or exclude covariates from the model. The results from each array were combined by meta-analysis. SNPs were excluded from the meta-analyses if they were not present (either directly genotyped or well-imputed) in all studies. There were 7,445,053 SNPs for the AA sample, 5,675,219 SNPs for the EA sample, and 4,795,232 SNPs for the
