Academic journal article Genetics

Improving Response in Genomic Selection with a Population-Based Selection Strategy: Optimal Population Value Selection

Academic journal article Genetics

Improving Response in Genomic Selection with a Population-Based Selection Strategy: Optimal Population Value Selection

Article excerpt

(ProQuest: ... denotes formulae omitted.)

IN genomic selection (GS), genome-wide genetic markers and phenotypic observations are used to estimate marker effects that can subsequently be used to accurately predict breeding values of individuals that have only been genotyped (Meuwissen et al. 2001). GS improves upon marker-assisted selection by more effectively capturing the effects of all quantitative trait loci (QTL). Because GS typically more accurately identifies superior parents than traditional selection methods, and decreases the amount of necessary phenotyping, it has been recognized as a way to increase the rate of genetic gain in cultivar improvement programs (Bernardo and Yu 2007), and to allow for more breeding cycles per unit time (Heffner et al. 2010).

Three extensions have been proposed to improve GS. The first extension, weighted genomic selection (WGS), was proposed to increase the frequency of rare favorable alleles in the population to maximize long-term response (Goddard 2009). In a simulation study, WGS was shown to increase response relative to GS (Jannink 2010). The second extension, optimal haploid value (OHV) selection, calculates the best possible future breeding value of an individual when doubled haploids (DHs) are produced from it. This method was shown to improve response in a simulated wheat program using DHs (Daetwyler et al. 2015). By taking the maximum haplotype GEBV at each segment, Daetwyler etai. (2015) demonstrated the utility of maintaining seemingly unfavorable genome segments, because subsequent recombination can release favorable alleles. Selecting only those individuals with the highest overall genomic estimated breeding values (GEBVs) can lead to the loss of rare favorable alleles in the population, but an individual whose GEBV falls below the truncation point may be more favorable in the long-term because it harbors rare favorable alleles.

GS, WGS, and OHV selection perform truncation selection on individual breeding values (of varying definitions). However, the genetic merit of a single individual also depends on the genetic merit of the individuals with which it may be mated. After several generations of crossing and recombination, individuals will contain genetic material from multiple founder lines. Thus, the third extension, genotype building (GB) selection, uses the best two haplotype blocks of a subset of the founder population to derive a combined fitness value for that subset (Kemper et al. 2012). As such, it represents a shift from selecting superior individuals to selecting the set of individuals that are more likely to produce superior progeny when crossed with each other. Since plant breeders often seek to develop high performing inbred lines, a genotype building strategy can be applied without the constraints on coancestry used by Kemper et al. (2012).

In this article, we propose optimal population value (OPV) selection as a combination of GBandOHVselection. LikeOHV selection, OPV considers the haploid values of individual selection candidatesbutevaluates themeritofpotentialprogeny of a subset of selection candidates like GB selection. First, we mathematically defined the OPV, GS, WGS, OHV, and GB selection approaches. Then, a simulation study based on empirical data obtained fromaset ofmaizeinbredswasusedtoanalyze OPV's relative ability to improve response. The objectives of this paper were to (i) improve genetic gain, and (ii) investigate the efficacy of population-based selection methods.

Materials and Methods

In this section, we first present OPV selection. Then, four existing selection methods (GS, WGS,OHV, andGBselection) are mathematically defined. The following definitions will be used in subsequent sections to describe the five selection methods compared in this paper:

L: the number of marker loci.

M: the ploidy of the plants.

N: the number of individuals in the population.

A ∈ {0,1}LXMXN : a binary matrix indicating whether the allele at locus l on the mth copy of a chromosome of individual n is the major (Al,m,n = 1) or minor (Al,m,n = 0) allele. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.