Validating Genomewide Predictions of Genetic Variance in a Contemporary Breeding Program

Abstract

Predicting the genetic variance among progeny from a cross—prior to making said cross—would be a valuable metric for plant breeders to discriminate among possible parent combinations. The use of genomewide markers and simulated populations is one proposed method for making such predictions. Our objective was to assess the predictive ability of this method for three relevant quantitative traits within a breeding program regularly using genomewide selection. Using a training population of two‐row barley (Hordeum vulgare L.) lines, we predicted the mean (μ), genetic variance (VG), and superior progeny mean (μSP, mean of the best 10% of lines) of 330,078 possible parent combinations for Fusarium head blight (FHB) severity, heading date, and plant height. Twenty‐seven of these combinations were chosen to develop biparental populations, which were subsequently phenotyped for the same traits. We found that the predictive abilities (rMP) for μ and μSP were moderate to high (rMP = 0.46–0.69), whereas those for VG were lower (rMP = 0.01–0.48). Unsurprisingly, predictive ability was likely a function of trait heritability, as rMP estimates for heading date (the most heritable trait) were highest, and rMP estimates for FHB severity (the least heritable trait) were lowest. We observed strong negative bias when predicting VG (on average −83 to −96%), but the relative consistency of this bias across validation families indicates that it may have little impact when selecting crosses. We concluded that accurate predictions of VG and μSP are feasible, but as with any implementation of genomewide selection, reliable phenotypic data are critical.

A connected half‐sib family training population for genomic prediction in barley

“…A better option for increasing PA with CV A or CV W would be selection of parents that maximize within‐family genetic variance for all traits under selection. Predicting genetic variance of a cross has historically been a difficult prediction problem, but recent studies have described several methods to improve prediction of progeny variance based on genome‐wide molecular markers (Lehermeier, Teyssèdre, & Schön, 2017; Mohammadi, Tiede, & Smith, 2015; Osthushenrich et al., 2018) although variable results have been reported (Adeyemo & Bernardo, 2019; Lado et al., 2017; Neyhart & Smith, 2019). These methods require estimation of marker effects to predict progeny variance, meaning potential parents must be genotyped and phenotyped in a sufficiently large field experiment to accurately estimate marker effects.…”

“…(2019) may provide datasets that enable accurate parental selection for new breeding programs based on expected progeny genetic variance. Initial empirical results from Neyhart and Smith (2019) suggest this approach may work for high heritability traits, but not for those of low heritability.…”

Exploring the Realm of Possibilities: Trying to Predict Promising Crosses and Successful Offspring through Genomic Mating in Barley

“…The identification of the most promising crosses therefore rests on adequately predicting the mean and variance for high priority traits within large sets of potential crosses [2].…”

“…Thus, for each cross, the mean can be calculated as the mean of the parents' GEBVs, based on the estimation of single-nucleotide coancestry/distance across the entire genome, including a majority of mostly neutral loci, and do not provide predictions for each trait [2].…”

Multi-Trait Improvement by Predicting Genetic Correlations in Breeding Crosses

“…The predictive abilities for remaining trait 381 combinations were not significantly different from zero (P > 0.05; bootstrapping). The ability to 382 predict the genetic correlation appeared to coincide with the heritability of both traits; the entry-383 mean heritability in the TP (and in the VF) was 0.45 (0.11) for FHB severity, 0.96 (0.78) for 384 heading date, and 0.52 (0.74) for plant height (Neyhart and Smith 2019). 385…”