The discrepancy between the high heritability estimates in twin and adoption studies on the one hand, and the inability to identify genes involved in these behaviors on the other hand, has been often referred to as the problem of the ‘missing’ heritability [46]. While some genome-wide association studies have been successful in identifying common SNPs, the majority of genetic variants that contribute to disease susceptibility remain undiscovered [29]. Moreover, these associated genes typically explain only a small proportion (<1%) of the genetic variance underlying the trait. The power calculation shows that our sample is unable to detect common genetic variants of small effect sizes that contribute to the variance in antisocial behavior. Yang et al. (2010) showed that it is likely that the heritability is not ‘missing’, at least in part, but that the SNPs that tag certain genes have a very small effect individually and might therefore not be detected in the analyses [30], [47]. We estimated that the total proportion of phenotypic variance explained by genome-wide SNPs when considered together is 0.55, with a standard error of 0.41.
