The gene-based PRSgene were calculated using variants located in AUD genes implicated in both AA and EA (i.e., P-values < 0.05 in both populations), consequently, they should be applicable to both populations. To test this proposition, an EA target dataset was also tested. As some EA samples of COGA and SAGE data were part of EA-PAU, they were not included as the target EA datasets, instead, EA individuals were drawn from the Indiana Biobank (https://indianabiobank.org/). The Indiana Biobank is a state-wide collaboration that provides centralized processing and storage of specimens that are linked to participants’ electronic medical information via Regenstrief Institute at Indiana University. All Indiana Biobank individuals included in this study were unrelated. AUD in Indiana Biobank was diagnosed based on ICD9 (303 and 305.0) and ICD10 (F10) codes. Individuals not diagnosed as AUD and without AUD-associated conditions such as alcohol-associated pancreatitis were defined as controls.
