Genomic prediction of grain yield in commercial Finnish oat (Avena sativa) and barley (Hordeum vulgare) breeding programmes

Abstract

Genomic selection has been adopted in many plant breeding programmes. In this paper, we cover some aspects of information necessary before starting genomic selection. Spring oat and barley breeding data sets from commercial breeding programmes were studied using single, multitrait and trait‐assisted models for predicting grain yield. Heritabilities were higher when estimated using multitrait models compared to single‐trait models. However, no corresponding increase in prediction accuracy was observed in a cross‐validation scenario. On the other hand, forward prediction showed a slight, but not significant, increase in accuracy of genomic estimated breeding values for breeding cohorts when a multitrait model was applied. When a correlated trait was used in a trait‐assisted model, on average the accuracies increased by 9%–14% for oat and by 11%–28% for barley compared with a single‐trait model. Overall, accuracies in forward validation varied between breeding cohorts and years for grain yield. Forward prediction accuracies for multiple cohorts and multiple years’ data are reported for oat for the first time.

Mitigating the impact of selective phenotyping in training populations on the prediction ability by multi‐trait pedigree and genomic selection models

“…The usage of molecular marker data for implementing predictive breeding methods to support the development of new and improved varieties has strongly increased during recent years (Haikka et al., 2020; Juliana et al., 2019; Schrag, Schipprack, & Melchinger, 2019), as such a strategy can result in higher selection gains than those possible by relying merely on observed phenotypic performance of quantitatively inherited traits like yield (Rife, Graybosch, & Poland, 2018; Sleper et al., 2020). Aside from introducing genomic relationships into various prediction models, the possibility to utilize relationships based on pedigree records has been the subject of past (Bauer, Reetz, & Léon, 2006) and more recent investigations (Cappa et al., 2019).…”

“…It has been furthermore shown that a large phenotypic variance is desirable in such training populations (Isidro et al., 2015), which can be of advantage when studying specific traits like abiotic stress tolerance that can be difficult to phenotype for larger panels and for which model updating is accordingly rather irregularly done. However, a large part of prediction model updating in applied breeding programmes is done with superior germplasm, which is advanced to thorough multi‐environment testing without conducting additional field trials for specific training populations (Haikka et al., 2020; Schrag et al., 2019; Sleper et al., 2020).…”

Multi-Year Dynamics of Single-Step Genomic Prediction in an Applied Wheat Breeding Program

“…The more cost-efficient single-step genomic prediction with a non-genotyped validation population (SSG-BLUP) showed an increase in predictive performance for grain yield and protein content when ω → 1 in H −1 [6] in comparison to the baseline P-BLUP model (Figure 3 central + right columns), but it resulted generally in lower prediction accuracies than the mentioned pedigree-genomic prediction (PG-BLUP) models. Blending of the pedigree and genomic relationship matrices by αG adj + βA 22 appeared not to be favourable in SSG-BLUP as higher prediction accuracies were achieved when α → 1 and β → 0 .…”

“…These combinations have thus been investigated in various studies, which found that the prediction of tested genotypes in already tested environments results in the highest accuracy followed by the prediction of untested genotypes in tested environments [1,2]. The most challenging scenario is though given by predicting untested genotypes in untested environments, which corresponds to, among others, genomic-based prediction across multiple years in applied breeding programs [3][4][5][6]. This prediction problem can however be simplified in some cases to predicting already tested genotypes in untested environments or years by including preexisting information e.g., from preliminary yield trials for the trait of interest or of correlated secondary traits into prediction models [7][8][9].…”

Genome-wide Association Study and Genomic Prediction for Fusarium graminearum Resistance Traits in Nordic Oat (Avena sativa L.)

“…To our knowledge, validated genomic predictions have not been reported before for FHB resistance in oat. GS studies in oat include traits, like grain yield [98,99], which imply that the use of GS in breeding for quantitative traits may be beneficial in oats.…”