Significant ML SVM models were calculated predicting remission from AUD for individuals who were previously diagnosed as AUD DSM-5. Sex and ancestry stratified analysis created an individualized model for each of the groups: EA males, EA females, AA males, and AA females (full details of the number of participants and matching age for each of the models in Supplementary Tables S1–S4). Table 1 summarizes the results of the significant predictive model scores across ancestry and sex (see Tables S7–S10 for full results), confirming the previous finding that the combined feature model (e.g., AA males and females models with EEG, PRS, medication, and demographic features) was more accurate than models based on single domain (Fig. 1). We found higher model accuracy when group’s ancestry was defined by genetics than by self-report, in EA males (p < 0.001) and AA males (p < 0.001) in models with only PRS as features (Supplementary Fig. S1). No difference was found in the females’ groups between the two types of ancestry definition. The AA male group combined feature model of PRS, EEG-FC, marital status, and
