As noted earlier, the risk of AUD development was approximately half (ORs of 0.53 and 0.57, respectively) for the bottom 5% of samples with the lowest PGS in the AOU and IB cohorts. These ORs were similar to the OR for rs1229984 estimated from the MVP dataset36 (OR, 0.57 [95% CI, 0.54-0.61]) and close to the OR estimated from the FinnGen dataset38 (OR, 0.41 [95% CI, 0.33-0.51]). The rs1229984 T allele frequencies were 4.2% in the AOU dataset, 2.9% in the IB dataset, 3.0% in the MVP dataset, and 0.5% in the FinnGen dataset. The low allele frequency may be the reason for the smaller OR in the FinnGen dataset due to the smaller number of T allele carriers. We did not include rs1229984 in our calculations of PGS because it did not pass the Hardy-Weinberg equilibrium test in the UKBB dataset; therefore, PGS and rs1229984 can be used together to identify those with a low risk of AUD. It is noteworthy that the results of the bottom 10% and 5% were similar; therefore, these findings suggest that we can
