Various SNPs were implicated as salient features in predicting the vulnerability to develop AUD. Interestingly, they varied by gender and ancestry. Moreover, females and males did not share any implicated SNPs, which may shed light on previous discrepancies observed in unstratified studies. One variant on chromosome 16 (rs4780836), previously associated in a large GWAS with alcohol consumption12, was found both in AA and EA females’ models suggesting gender-specificity of this susceptibility marker. While this study focused on individual SNPs from previous GWAS, future studies should aggregate information from a large number of potentially causal SNPs, such as Polygenic Risk Score (PRS), to increase features matching12, 54
