Third, our study is the first (to our knowledge) to include families of African ancestry in analyses of the environmental mechanisms that transmit genetic risk for alcohol problems across generations. We found a pattern of effects consistent with environmental inheritance whereby maternal AUD clinical criterion counts mediated the effects of nontransmitted maternal alleles on AUD clinical criterion counts and maximum drinks in the child generation. However, we did not find support for the hypothesis that parental relationship discord nor divorce mediated genetic influences on offspring alcohol outcomes in our subsample of African ancestry families. These null effects may be attributable to reduced statistical power due the smaller number of families in this subsample, and/or the limited predictive power of polygenic risk scores in African ancestry populations due to historic underrepresentation in genome-wide association studies. Consistent with this possibility, zero-order correlations indicated very few statistically significant associations between parental alleles and offspring alcohol outcomes, and no statistically significant associations between parental alleles and relationship discord and divorce. In view of these weak polygenic associations, and as with any nonsignificant result, we
