nominally significantP < 0.05
To account for unmeasured confounders, within each population, we empirically determined the number of gene expression principal components (PCs) to adjust for in order to maximize the number of nominally significant eQTL associations identified ( P < 0.05).