Two large GWAS of AUD-related phenotypes: AUD determined using ICD codes from the Million Veteran Program (MVP-AUD) (Kranzler et al., 2019) and scores derived from the problem subscale (questions 4–10) of the Alcohol Use Disorder Identification Test (AUDIT) from the UK Biobank (UKBB-AUDIT-P) (Sanchez-Roige et al., 2019) were used as the discovery datasets. MVP-AUD (N=202,004; 5,933,416 variants) (Kranzler et al., 2019) were obtained from the database of Genotypes and Phenotypes (dbGaP, phs001672). UKBB-AUDIT-P (N=121,604; 15,312,259 variants) were provided by authors of the original publication (Sanchez-Roige et al., 2019). Only European ancestry samples in both datasets were used due to the limited non-European ancestry samples available (with resultant insufficient statistical power) and complicated linkage disequilibrium structures in admixed populations (e.g., African American, Latinx). A/T or C/G variants were excluded to avoid strand ambiguity. The two GWAS used different phenotypes – one clinically ascribed in healthcare settings (i.e., ICD codes for AUD, requiring one inpatient or two outpatient ICD9/10 codes (Kranzler et al., 2019)) and another via self-report on a questionnaire (i.e., AUDIT) (Sanchez-Roige et al., 2019). Furthermore, the study cohorts differed
