No SNP was significantly associated with BD in the CADD GWAS. This is not surprising, given the relatively small sample. GWAS of psychiatric and behavioral phenotypes that have successfully identified and replicated individual effects of common SNPs have relied on very large samples (Rietveld et al. 2013; Ripke et al. 2013). Despite the lack of significance of any individual SNP, GCTA REML analysis estimated that 49.3% (SE=0.31, p=0.06) of the Caucasian ancestry sub-sample variation in BD could be accounted for by all of the genotyped SNPs. Conversely, a similar study found no evidence of variance in early adolescent (12-year-old) non-substance behavioral problems being attributable to common variants (Trzaskowski et al. 2013). This may suggest qualitative differences between genetic effects on BD at different ages, an effect that has been reported from twin models of comorbidity between dependence on different substances (Vrieze et al. 2012), which is a marker of BD.
