LD patterns across the genome enable the calculation of genetic correlations between traits. This is because the observed association for a SNP is a product of both its own contribution toward a phenotype and the association of the SNPs that are in LD with it (Yang et al., 2011). Because SNPs in regions of high LD tag a greater proportion of the genome than SNPs in regions of low LD, SNPs in regions of high LD will have stronger associations than SNPs found in regions of low LD. Thus, by using the known LD structure of a reference SNP panel, the SNP heritability of a single phenotype or the genetic correlation of two phenotypes can be computed using LD score regression (Bulik-Sullivan et al., 2015a; Lee et al., 2016).
