We assessed the global landscape of expression using principal components analysis (PCA) (Figure S1). Unlike SNP-based PCA plots for the same populations, all populations in the expression-based PCA plot do not separate distinctly by their continental ancestry. We assessed correlation of principal component (PC) 1 from the SNP-based PCA which separates African/non-African populations against all principal components from the expression PCA and found decay of correlation from the first 50 principal components maximized at PC3 and PC7 highlighting that gene expression differences, while not distinguishable by heredity alone, are partly shaped by it (Figure S2).
