The ability to predict vulnerability and identify related predisposition biomarkers holds enormous possibilities including preventions tactics, treatments or simply avoidance. Equally important is the ability to identify resilience factors, those biomarkers or psychosocial “protective” characteristics, that can thwart or prevent the progress of alcohol dependence. Overall, our findings demonstrate the importance of embedded ancestry, gender and age in the calculation of model prediction of the development of AUD. This approach we argue, should be expanded to any diagnosis or prediction of treatment response. We further show that the model based on various features from different realms (genetics, electrophysiology and FH) outperform prediction models based on singular-realm features. Wider selection of features with a narrower approach when choosing the sample will generate better prediction scores, enabling accurate anticipation of the development of an undesirable disorder. We also identified specific robust features of EEG and SNP measurement for each gender/ancestry group, further deepening our understanding of the predisposition of brain mechanisms underlying the future development of AUD. Future studies are required to further validate these results with larger cohorts, sampling uniformity and wider selection of features.
