We selectively examined the admixed BA population (151 unique individuals; Table 1) to (1) characterize transcriptional changes associated with African or European genetic ancestry in neurotypical adults (age > 17) and (2) limit the potential confounding effects of systematic environmental factors that may differ between BAs and non-Hispanic WAs. We used RNA-seq data from the caudate nucleus (n = 122), dentate gyrus (n = 47), DLPFC (n = 123) and hippocampus (n = 133). The admixed BA donors showed a varied proportion of EA (STRUCTURE17; EA mean = 0.21, range = 0–0.62; Supplementary Fig. 1) consistent with previous reports and the history of the slave trade18,19. We used these continuous genetic ancestry estimates to identify differentially expressed features (genes, transcripts, exons and junctions) linearly correlated with ancestry proportion and adjusted for sex, age and RNA quality. This RNA quality adjustment included experiment-based RNA degradation metrics that account for batch effect and cell composition12,20. To increase detection power and improve effect size estimation, we applied the multivariate adaptive shrinkage (‘mash’21) method, which leverages the correlation structure of genetic ancestry effects across
