our family sample based on a weighted and unweighted GRM, we obtained median estimates of SNP heritability of .29 for P3 amplitude and .27 for genetic factor scores. This represents approximately 40% to 50% of the heritable variance in each trait. These estimates are imprecise; 95% confidence intervals around them are necessarily large when derived from genetically unrelated individuals. Nevertheless, results of GCTA analyses indicate that much of the additive genetic influence in both endophenotypes is due to common genetic variants. GCTA that accounted for shared environmental influences within families and GCTA with the full sample (without a threshold of genetic relatedness) produced nearly identical estimates (.571 and .570, respectively). These numbers cannot be considered SNP heritability estimates, because they are driven by all factors that cause highly related individuals to have similar values of P3 amplitude, such as nonadditive genetic influences and rare variants that are not tagged by the SNPs on the genotyping array. However, the fact that they were virtually identical indicates that shared environmental influences were minimal. This is consistent with the fact that the 95% confidence interval around the estimate of C in biometric models of both twin data and family data included 0.
