We manually annotated the traits studied in each GWAS as related to blood (176 SNPs), brain (61 SNPs) or other (546 SNPs) phenotypes (Supplementary file 1). For example, we annotated traits associated with neurological or psychiatric conditions as “brain” and markers of subtypes of blood cell markers as “blood”. We then used this list of SNPs to perform eQTL analysis. We first performed the eQTL analysis in a uniform way by only considering the subset of probes and SNPs detected in all tissue types, or 2929 SNP:probe pairs. This analysis identified eQTLs that were highly significant in all three tissues and additional eQTLs distinctly significant in either blood or brain tissues (Fig. 3). Of the shared eQTLs, three stood out as highly significant in all three tissues for three SNPs including a single mRNA probe, ILMN_1695585 that maps to the RPS26 gene on chromosome 12q13.2, within 500 KB of three GWAS SNPs associated with Type 1 diabetes (False discovery rate (FDR) corrected P < 1.45 × 10− 38 for association with rs11171739 in the frontal cortex, P < 6.72 ×
