To identify DEGs potentially driven by environmental factors, we used DNA methylation as an environmental proxy in BAs. We first identified the top 1% of variable CpGs probably driven by unknown environmental factors. We identified these CpGs by removing variation attributable to technical and biological factors captured by the top five DNA methylation principal components, while preserving variation due to global ancestry. We then grouped these top variable CpGs into variable methylated regions (VMRs) for the caudate nucleus (89 samples, 12,051 VMRs), DLPFC (69 samples, 9,701 VMRs) and hippocampus (69 samples, 9,924 VMRs). In contrast to our differential expression analysis, we found few global ancestry differentially methylated regions (DMRs) (FDR < 0.05; n = 3, 1 and 8 for the caudate nucleus, DLPFC and hippocampus, respectively). However, we identified a larger number of local ancestry-associated DMRs (FDR < 0.05; n = 494, 260 and 265 for the caudate nucleus, DLPFC and hippocampus, respectively; Fig. 5a).
