individual also has a positive family history of AUD, then they may be at elevated genetic risk for AUD and could potentially be used as proxy cases for AUD in GWAS studies. On the other hand, many family history positive individuals do not develop AUD but may be also at elevated genetic risk for AUD. Including these individuals as controls could reduce the statistical power when sample size is mall. Note in this study, we used the first-degree relatives to define the family history because many datasets may not have detailed family history information from distant relatives (Scheuner et al., 2006), and the data for first degree relatives are likely to be more accurate than those of distant relatives. We performed a sensitivity analysis using all relatives to define the family history, which increased the sample size and heterogeneity of FH+ and decreased the sample sizes and heterogeneity of FH? and FH−. All results were similar except that the difference of mean PRS between AUD cases and controls was not significant in FH? subsample, partially due to the dramatically reduced sample size (199 AUD cases and 135 controls).
