Our estimate of 45% is still less than the 80% of phenotypic variance due to additive genetic effects (that is, the estimated heritability). One reason why the SNPs do not explain the full estimated heritability is that the SNPs on the arrays are not in complete LD with causal variants. The ability of the SNPs to explain the phenotypic variation caused by causal variants depends on the LD between all the causal variants and all the SNPs. Lack of complete LD is manifested as a difference between the genomic relationship between each pair of subjects j and k at the causal variants (Gjk) and the relationship between the same individuals calculated from the SNPs (Ajk). As causal variants are unknown, we cannot estimate their LD with observed SNPs directly. However, we can mimic it by considering the LD of the genotyped SNPs with one another. It is likely that the causal variants and the SNPs have different properties, so LD among SNPs is only a guide to LD between causal variants and SNPs. One way in which the causal variants
