Linkage disequilibrium of admixed cohorts from ancestral reference panels and GWAS summary statistics

Reference-panel LD misrepresents admixed cohorts and biases summary-statistics heritability upward. This work composes the LD of an admixed cohort from ancestral reference panels and quantities available in GWAS summary statistics, so that heritability and related estimates for admixed cohorts, such as African-ancestry participants of All of Us, do not require access to in-sample genotypes.