The genetic regulation of the human brain is not completely understood. GWAS studies have shown that a majority of trait-associated variants reside in the non-coding genome. To study the effect of these variants on the brain, we assembled a multi-ancestry multi-tissue resource of paired RNA-DNA data from 4,656 individuals across 10,725 samples from publicly available cohorts. Using a mixed model that accounts for sample overlaps, diverse ancestry, and effect size heterogeneity across tissues, we mapped cis-eQTLs. We found that the meta-analysis significantly boosted discovery compared to individual cohorts helping us find 22,663 eGenes.