IBDkin: Fast estimation of kinship coefficients from identity by descent segments

Abstract

Motivation: Estimation of pairwise kinship coefficients in large datasets is computationally challenging because the number of related individuals increases quadratically with sample size. Results: We present IBDkin, a software package written in C for estimating kinship coefficients from identity by descent (IBD) segments. We use IBDkin to estimate kinship coefficients for 7.95 billion pairs of individuals in the UK Biobank who share at least one detected IBD segment with length ≥ 4 cM.