Strategies to Increase Prediction Accuracy in Genomic Selection of Complex Traits in Alfalfa (Medicago sativa L.)

Abstract

Agronomic traits such as biomass yield and abiotic stress tolerance are genetically complex and challenging to improve through conventional breeding approaches. Genomic selection (GS) is an alternative approach in which genome-wide markers are used to determine the genomic estimated breeding value (GEBV) of individuals in a population. In alfalfa (Medicago sativa L.), previous results indicated that low to moderate prediction accuracy values (<70%) were obtained in complex traits, such as yield and abiotic stress resistance. There is a need to increase the prediction value in order to employ GS in breeding programs. In this paper we reviewed different statistic models and their applications in polyploid crops, such as alfalfa and potato. Specifically, we used empirical data affiliated with alfalfa yield under salt stress to investigate approaches that use DNA marker importance values derived from machine learning models, and genome-wide association studies (GWAS) of marker-trait association scores based on different GWASpoly models, in weighted GBLUP analyses. This approach increased prediction accuracies from 50% to more than 80% for alfalfa yield under salt stress. Finally, we expended the weighted GBLUP approach to potato and analyzed 13 phenotypic traits and obtained similar results. This is the first report on alfalfa to use variable importance and GWAS-assisted approaches to increase the prediction accuracy of GS, thus helping to select superior alfalfa lines based on their GEBVs.

Alfalfa genomic selection for different stress‐prone growing regions

“…Our study confirmed the greater predictive ability of the WG‐BLUP model over other statistical models as put forward for alfalfa biomass yield by Medina et al. (2021), although its current advantage over other models was not as large as that reported by these authors. Our extension of the WG‐BLUP model to incorporate GEI effects confirmed the advantage of this model over other models even under this scenario.…”

“…(2018) and Montesinos López et al. (2022), and WG‐BLUP (Medina et al., 2021). Ridge regression BLUP assumes a linear mixed additive model where each marker is assigned an effect as a solution of the following equation: $$$$\begin{equation*}{y}=1{{\mu}}+{\bf{Zu}}+ {{\bf \varepsilon}}\end{equation*}$$$$where y is the vector of observed phenotypes, μ is the mean of y , Z is the genotype matrix (e.g., {0,1,2} for biallelic SNPs), u ∼ N (0, Iσ 2 u ) is the vector of marker effects, and ε is the vector of residuals.…”

“…We tested five single-environment models: ridge regression BLUP (RR-BLUP), genomic BLUP (G-BLUP), BayesCπ and Bayesian reproducing kernel Hilbert space (RKHS) as described by Wang et al (2018) and Montesinos López et al (2022), and WG-BLUP (Medina et al, 2021). Ridge regression BLUP assumes a linear mixed additive model where each marker is assigned an effect as a solution of the following equation:…”

“…The adopted GS statistical model among the several ones available (Crossa et al., 2017; Montesinos López et al., 2022; Wang et al., 2018) may affect the model predictive ability for alfalfa biomass yield (Annicchiarico, Nazzicari, et al., 2015; Hawkins & Yu, 2018). In a recent study, a weighted genomic best linear unbiased prediction (WG‐BLUP) GS model exploiting genome‐wide association study results to assign a weight to SNP markers exhibited a sharp prediction accuracy advantage over other popular statistical models (Medina et al., 2021). Another relevant factor that may influence genome‐enabled predictions in alfalfa is the SNP calling procedure for GBS‐generated markers, which may rely on (a) the genome of the model species Medicago truncatula (Li, Wei et al., 2014), (b) a mock reference genome based on the linkage disequilibrium of SNP markers observed in a small subset of test genotypes (Puritz et al., 2014), or (c) the alfalfa genome recently sequenced by Chen et al.…”

KBeagle: An Adaptive Strategy and Tool for Improvement of Imputation Accuracy and Computing Efficiency

“…The full name of rrBLUP (Medina et al 2021) is 'ridge regression best linear unbiased prediction', which is the best linear unbiased prediction of ridge regression. It is one of the most commonly used models for genome selection and also represents the indirect method models.…”