Realized genome sharing in random effects models for quantitative genetic traits

Abstract

DNA copies inherited from the same ancestral copy by related individuals are said to be identical by descent (IBD). IBD gives rise to genetic similarities between related individuals. In quantitative genetics, two fundamental problems are heritability estimation and gene mapping for genetic traits. IBD plays a critical role in the study of both problems. When working with population-based samples where pedigree information is unavailable, it is essential to estimate IBD accurately from genetic marker data using pedigree-free methods. The estimated IBD can then be used in heritability estimation and gene mapping using random effects models. For pedigree-free IBD estimation, we showed that it is important to use the fact that DNA is inherited in segments as opposed to independent loci. As the single nucleotide polymorphism (SNP) marker panels become increasingly dense, the impact of allelic association (or linkage disequilibrium, LD) on accuracy of IBD estimation also grows. Through simulation studies, we demonstrated that adjusting for LD in the marker panel can lead to improved IBD estimation accuracy. For heritability estimation and gene mapping using random effects models, a difficult task is to specify the correlation structures of the random genetic effects, which are typically functions of IBD sharing over the putative causal genomic region. We provided formulas for the asymptotic bias and sampling error of heritability estimates, when the genetic correlation structures are potentially mis-specified. Mis-specification of the genetic correlation structures can occur due to inaccurate IBD estimation or mis-identification of the causal genome. We showed that such mis-specification can lead to substantial downward bias in heritability estimation, or loss of power in gene mapping.