Using effect sizes generated via a prior GWAS of AO-AD in the COGA family sample,15 we created GRS at varying P-value thresholds and examined their association with AO-AD, AD, AO-I, as well as liability to problematic drinking (AD-SX and Maxdrinks) in two independent and differently structured and ascertained datasets, SAGE and OZ-ALC. GRS, especially when including SNPs associated with AO-AD at more liberal P-value thresholds, were significantly associated with a range of these alcohol-related measures in those two independent and distinctly ascertained samples. In contrast, there was no evidence for replication of the top 10 most significantly associated variants from COGA in either SAGE (6.6 × 10−1–8.1 × 10−1) or OZ-ALC (8.7 × 10−1–8.9 × 10−1). This strongly underscores the idea that multiple common genetic effects contribute to the etiology of complex disorders like addictions.
