Multi‐trait genomic prediction improves selection accuracy for enhancing seed mineral concentrations in pea

Abstract

Multi‐trait genomic selection (MT‐GS) has the potential to improve predictive ability by maximizing the use of information across related genotypes and genetically correlated traits. In this study, we extended the use of sparse phenotyping method into the MT‐GS framework by split testing of entries to maximize borrowing of information across genotypes and predict missing phenotypes for targeted traits without additional phenotyping expenditure. Using 300 advanced breeding lines from North Dakota State University (NDSU) pulse breeding program and ∼200 USDA accessions that were evaluated for 10 nutritional traits, our results show that the proposed sparse phenotyping aided MT‐GS can further improve predictive ability by >12% across traits compared with univariate (UNI) genomic selection. The proposed strategy departed from the previous reports that weak genetic correlation is a limitation to the advantage of MT‐GS over UNI genomic selection, which was evident in the partially balanced phenotyping‐enabled MT‐GS. Our results point to heritability and genetic correlation between traits as possible metrics to optimize and further improve the estimation of model parameters, and ultimately, prediction performance. Overall, our study offers a new approach to optimize the prediction performance using the MT‐GS and further highlight strategy to maximize the efficiency of GS in a plant breeding program. The sparse‐testing‐aided MT‐GS proposed in this study can be further extended to multi‐environment, multi‐trait GS to improve prediction performance and further reduce the cost of phenotyping and time‐consuming data collection process.

Effective Population Size in Field Pea

“…The breeding lines used in this experiment were carefully chosen and contain both contemporary and past elite germplasm. (Bari et al 2023;Atanda et al 2022).…”

Multi-trait multi-environment genomic prediction of preliminary yield trials in pulse crops

“…Univariate or single-trait (UNI) models have been widely employed in GS, focusing on predicting individual traits independently while assuming no correlation between traits (Atanda et al, 2022; Sandhu et al, 2022; Montesinos-López et al, 2022). Multi-trait GS (MT-GS) models integrate information from correlated traits and shared genetic information between lines to improve the accuracy.…”

“…Multi-trait GS (MT-GS) models integrate information from correlated traits and shared genetic information between lines to improve the accuracy. (Jia and Jannink, 2012; Gill et al 2021; Atanda et al, 2022; Montesinos-López et al, 2022;). As traits are genetically correlated, these MT-GS models have demonstrated their ability to enhance prediction accuracy, particularly for traits with inherently low heritability.…”

“…However, in practice, the correlation between traits ranges from positive to negative, along with varying degrees of heritability. Addressing this challenge, Atanda et al (2022) emphasized composition of traits in the training and prediction sets based on the heritability and genetic correlation between traits to enhance the prediction accuracy. Studies have also shown that the integration of genotype by environment (GxE) in the MT model further improves prediction accuracy (Gill et al, 2021; Sandhu et al, 2022).…”

A Bayesian model for genomic prediction using metabolic networks

“…The typical additional information is “omics data.” For example, transcriptome (Li et al ., 2019; Perez et al ., 2022) and metabolome (Riedelsheimer et al ., 2012; Campbell et al ., 2021) are often used together with the genome, and some studies have used both (Xu et al ., 2016; Schrag et al ., 2018). Furthermore, because data in biology are essentially multivariate (i.e., multiple traits can be measured for a genotype at multiple environments), it has been of interest to learn how to utilize multivariate information to improve prediction accuracy for target traits (e.g., Jia and Jannink, 2012; Jarquin et al ., 2016; Atanda et al ., 2022).…”