A novel linkage-disequilibrium corrected genomic relationship matrix for SNP-heritability estimation and genomic prediction

Abstract

Single nucleotide polymorphism (SNP)-heritability estimation is an important topic in several research fields, including animal, plant and human genetics, as well as in ecology. Linear mixed model estimation of SNP-heritability uses the structures of genomic relationships between individuals, which is constructed from genome-wide sets of SNP-markers that are generally weighted equally in their contributions. Proposed methods to handle dependence between SNPs include, “thinning” the marker set by linkage disequilibrium (LD)-pruning, the use of haplotype-tagging of SNPs, and LD-weighting of the SNP-contributions. For improved estimation, we propose a new conceptual framework for genomic relationship matrix, in which Mahalanobis distance-based LD-correction is used in a linear mixed model estimation of SNP-heritability. The superiority of the presented method is illustrated and compared to mixed-model analyses using a VanRaden genomic relationship matrix, a matrix used by GCTA and a matrix employing LD-weighting (as implemented in the LDAK software) in simulated (using real human, rice and cattle genotypes) and real (maize, rice and mice) datasets. Despite of the computational difficulties, our results suggest that by using the proposed method one can improve the accuracy of SNP-heritability estimates in datasets with high LD.

Accuracy of genomic selection and long‐term genetic gain for resistance to Verticillium wilt in strawberry

“…Cross‐validations were performed within and between years in all combinations, as outlined above. Narrow‐sense genomic heritability was estimated from the mean of 10,000 MCCV‐generated REML estimates of the variance components from G‐BLUP (Endelman, 2011; Mathew, Léon, & Sillanpää, 2018).…”

Genomic, Transcriptomic and Epigenomic Tools to Study the Domestication of Plants and Animals: A Field Guide for Beginners

“…Parametric methods, also known as model-based methods, assign individuals into a predefined number of K populations based on their genotypes and the allele frequency of each locus (Pritchard et al, 2000). Several parametric methods have been described that successfully analyze genomic datasets to infer population structure (e.g., Tang et al, 2006;Alexander et al, 2009;Raj et al, 2014), but one has to be careful when using them, as they assume linkage equilibrium and Hardy-Weinberg equilibrium in the dataset (Linck and Battey, 2019), so SNPs should be filtered accordingly before these methods can be confidently used (Wigginton et al, 2005;Mathew et al, 2018). Furthermore, parametric methods have been found to be susceptible to changes in the SFS generated by minor allele frequency thresholds that are commonly used to filter population genomics data because low-frequency polymorphisms are expected to contain information about recent events, which adds uncertainty to the assignation of individuals in populations that reflect ancient demographic events (Linck and Battey, 2019).…”

SNP heritability: What are we estimating?

“…1c). This corresponds to the LD corrected GRM, recently independently proposed as an alternative to the standard GRM based on empirical observations 13 .…”

“…Variance components, based on genomic relationship matrices, fitted using restricted maximum likelihood estimation, the so called G-REML method 3 , have been proposed, implemented in various tools [7][8][9][10] , and widely adopted as an approach to estimate Based on empirical observations, it has been suggested that G-REML estimates are biased under departure of the population from the model 4,8,11,12 . This has led to a number of variations on the G-REML approach based on various assumptions about the genetic architecture of the phenotype 8,[11][12][13] . These different models yield different estimates 14,15 .…”