Alcohol use disorder (AUD) is one of the most common public health problems [1] and both genetic and environmental factors contribute to risk. Estimates of the heritability of AUD range from 40% to 60% [2–4]. Recently, several large-scale genome-wide association studies (GWAS) of AUD-related phenotypes have reported many genetic variants associated with AUD [5–7]. These GWAS reiterated the highly polygenic underpinnings of AUD and related phenotypes where many variants contribute small effects on AUD. Consequently, polygenic risk scores (PRS) have proven to be a strong approach for assessing AUD genetic liability beyond the genome-wide significant variants [5, 7]. For instance, in our previous study of an European ancestry cohort [8], individuals comprising the top PRS decile were almost twice as likely to meet the criteria for AUD compared to all others, an estimate comparable to those published for the first-degree family history of AUD in national surveys [9, 10]. However, PRS analysis of AUD in admixed populations, such as African Americans (AA), suffer from poor performance due to the much smaller sample sizes of the discovery GWAS [5].
