Zhou et al., 2020). For participants of African ancestry, we used GWAS summary statistics from a meta-analysis of the European ancestry in tandem with GWAS summary statistics from the PGC DSM-IV alcohol dependence analyses (COGA sample removed; Walters et al., 2018) and MVP alcohol use disorder analysis (Kranzler et al., 2019) in individuals of African ancestry. Because PRS-CSx improves predictive power for non-European ancestry samples with smaller GWAS (Ruan et al., 2022), we included both the European and African ancestry-derived polygenic scores in the AA sample in COGA, but just the European ancestry-derived polygenic scores in the EA sample in COGA.
