used estimates from the GWAS of alcohol consumption (AUDIT-C) in the Million Veterans Project (MVP) sample (Kranzler et al., 2019), the largest published GWAS of alcohol phenotypes with a multiancestry sample including 209,020 EAs and 57,340 AAs. We calculated alc-PRS for our AA subsample using PRS-CSx, an extension of the PRS-CS method that integrates GWAS summary statistics from multiple populations and leveraging linkage disequilibrium (LD) diversity across discovery samples to improve polygenic prediction in non-EA samples (Ruan et al., 2022). Given GWAS summary statistics and ancestry-matched LD reference panels, PRS-CSx calculates one polygenic score for each discovery sample, and integrates them by learning an optimal linear combination to produce the final PRS. For the AA sample in the present study, Alc-PRS was calculated using PRS-CSx with GWAS summary statistics from the MVP EA and AA samples and 1000 Genomes ancestry-matched reference panels. An average of 865,352 SNPs was included in the construction of alc-PRS for AAs.
