it relies on a lower symptom burden (e.g., ≥2 of 11 vs ≥3 of 7 criteria for DSM-IV dependence), which may dilute identification of genetic effects on more severe forms of the disorder. However, when we examined the association between GWS variants and a count of DSM-5 criteria across alcohol and illicit drugs, these SNPs were associated with that count but not at GWS levels (p>5E-5). Finally, 39% of the controls met criteria for substance abuse – we elected to include these individuals to maintain the sample size. Consistent with this, results for chromosome 3 (AA p=2E-8) remained GWS while those for chromosomes 1 (EA p=3E-4; AA p=6E-5), 5 (AA p=7E-8) and 13 (AA p=4E-5) were attenuated in statistical significance but not in magnitude upon exclusion of individuals with abuse from the controls. Finally, the lack of replication was discouraging, although it is noteworthy that sample sizes for the replication cohorts were modest, and they may not have had sufficient power to replicate findings with these effect sizes. Nonetheless, upon meta-analysis, these variants did not show consistent genome-wide support indicating considerable heterogeneity across-samples, low power, or raising the possibility that the current findings are false positives.
