In our analysis of the smaller African ancestry subsample, we found no significant association between externalizing polygenic scores and adolescent externalizing behavior, nor associations between externalizing polygenic score and T1 parental knowledge and T1 peer substance use. The lack of predictive polygenic association is likely due in part to the smaller sample size and thereby reduced power, and in part to the fact that the discovery sample of our externalizing polygenic score included only individuals of European ancestry (Karlsson Linnér et al. in press). Polygenic scores do not translate well across ancestries (Martin et al. 2017), and populations of non-European ancestry, particularly those of African ancestry, have been historically underrepresented in gene identification efforts (Popejoy and Fullerton 2016). This is a limitation of the field as a whole, and represents a serious issue for studies with diverse populations. Although we attempted to address this issue by also implementing a multi-ethnic polygenic scoring approach (Márquez-Luna et al. 2017), the predictive power of the polygenic score constructed using this approach in the African ancestry subsample was still severely attenuated when compared to
