GCTA results are given in Figure 2 and differ depending on the sample and the GRM cutoff. For comparison, the twin-estimated heritabilities are provided (gray), as well as the sum of the heritabilities and shared environmental estimates (black). First, the best estimate of the aggregate effects of common SNPs was produced from the largest sample of genetically unrelated individuals (the parents), as determined by using a cutoff of < .025 on the genetic relatedness matrix. The parents share environments but not genes, and resulting GCTA estimates will not be confounded with non-common-SNP genetic effects and/or strong shared environmental effects. As can be seen in Figure 2, the estimates of phenotypic variance accounted for by the aggregated SNPs range from .16–.22 in the sample of unrelated parents (displayed in red). All estimates were statistically significant at one-tailed p<.05 except DEP (Alcohol Dependence). The full sample estimates (no GRM cutoff; displayed in yellow) yielded much higher estimates, consistent with the notion that rare-SNP, non-additive, non-SNP, and/or shared environmental effects are contributing to phenotypic similarity, sometimes substantially. It appears, however, that the GCTA
