Furthermore, we performed a Genome-wide Complex Trait Analysis (GCTA) to estimate the proportion of the heritability of liability to adult antisocial behavior that can be explained by testing the SNPs on the GWAS chips simultaneously [29], [30]. One individual per family was selected for the analysis. We used only genotyped SNPs. To reduce the potential for bias, SNPs that had a Hardy-Weinberg p-value<10−3, had >5% missingness in all samples, or showed evidence of differential missingness between cases and controls (p<0.01), were removed. In this way only good quality SNPs genotyped across all genotyping platforms were retained. A total of 278.570 SNPs remained after quality control. A stringent cut-off of 0.025 was used to remove pairs of individuals that show evidence of cryptic relatedness. The final sample comprised 160 cases and 2012 controls. Analysis was performed using the GCTA software and all 22 autosomes were fitted in the model simultaneously. The prevalence estimate was 0.035% as estimated in the phenotypic sample.
