We used ancestry-specific discovery GWAS results to calculate alcohol problems polygenic scores in the COGA EA and AA samples. For the EA sample, the alcohol problems PRS was derived using meta-analyzed GWAS weights (detailed in Barr et al., 2020) from the Psychiatric Genomic Consortium’s GWAS of alcohol dependence analyses (COGA removed; Walters et al., 2018) and UK Biobank’s problem subscale from the Alcohol Use Disorders Identification Test (AUDIT-P) analyses (Sanchez-Roige et al., 2018) for individuals of European ancestry. For the AA sample, the alcohol problems PRS was derived using GWAS weights from the Million Veteran Program GWAS of alcohol use disorder analysis for individuals of African ancestry (Kranzler et al., 2019). Higher polygenic scores indicated higher polygenic loading for alcohol problems.
