The strongest association (inflation corrected 9.4×10−7≤p≤ 9.8×10−6) was detected with a group of 18 highly correlated variants (r2≥=0.95) within and flanking C15orf53. Interestingly, recent GWAS data has reported consistent evidence showing that variation in a region close to C15orf53 influences susceptibility for bipolar disorder 33–36. SNPs that influenced bipolar disorder susceptibility (rs12912251, rs2172835, and rs12899449) also show strong association with SC (inflation corrected p=9.4×10−7, p=3.1×10−6, and p=1.26×10−5, respectively) in COGA families severely affected by AD. The alleles that are associated with reduced risk for BD are also associated with lower dependence symptom counts. We did not detect association between symptom counts and other variants reported by the Psychiatric GWAS Consortium Bipolar Disorder Working Group (2011) to be strongly associated with bipolar disorder. This suggests a specific effect of this gene on risk for bipolar disorder and alcohol dependence rather than a more general shared underlying liability to both disorders. Studies of psychiatric disorders have shown that bipolar disorder and alcoholism commonly co-occur37–39, and that individuals with bipolar disorder have a greater likelihood of AUDs than the general population40. Approximately 46%
