the SNPs, because the relationships between genetically unrelated subjects are not influenced by shared environment or non-SNP genetic variance. In a sample of related subjects, such as the families used in the present study, phenotypic relatedness and genetic relatedness are confounded in important ways that must be addressed. If estimates are based on everyone, then SNP-based genetic relatedness and phenotypic relatedness are confounded with rare and non-additive genetic relationships (e.g., MZ twins share almost all variants, including rare and common SNPs) and shared environmental effects due to shared family experiences.
