Another limitation is that although we demonstrated these associations across different substances of abuse, these genetic variants may relate to different aspects of substance use including diagnosis, craving, age of onset, and sensitivity to substances among others. Our study also cannot be generalized to samples with different demographic characteristics. This was a high-risk sample comprised of individuals with alcoholic fathers, limiting generalizability to population representative samples. Future studies should explore these mediated relationships in more diverse samples. Lastly, these genetic associations were not replicated using a second independent sample. Given the importance of replicating genetic effects, significant caution is warranted when drawing inferences until findings are demonstrated using another sample. Replication is an important goal as new, cross-level, longitudinal studies become available.
