tissues, including subcutaneous adipose tissue, tibial artery, transformed fibroblast, tibial nerve, and thyroid (Supplemental Table 2). Considering multi-tissue eQTL posterior probabilities for ADH1B available in the GTEx dataset, rs10516440 showed effects on multiple tissues, including subcutaneous adipose tissue, tibial artery, heart (left ventricle), lung, skeletal muscle, tibial nerve, sun-exposed skin (lower leg), thyroid, and whole blood (Figure 3). Rs10516440 is located in the upstream region of ADH gene cluster, and it affects gene expression of all ADH1 genes in multiple tissues (FDR < 5%; Supplemental Table 3). In our genome-wide association study (GWAS) of alcohol dependence (AD) [Gelernter and others 2014], the minor allele rs10516440*G, which correlated with reduced expression of ADH1 genes, was nominally associated with increased AD risk (N = 8,788, z = 3.952, p = 7.74*10−5) with nearly equal contribution from both African-Americans (N = 4,141, z = 2.63, p = 8.52*10−3) and European-Americans (N = 4647, z = 2.96, p=3.07*10−3), although, lacking the associated informatics, this was not stressed in that prior publication. This observation may serve to highlight the possible advantages of identifying functional SNPs based on various annotations.
