AUTHOR=Montesinos-López Osval A. , Montesinos-López Abelardo , Kismiantini  , Roman-Gallardo Armando , Gardner Keith , Lillemo Morten , Fritsche-Neto Roberto , Crossa José 

TITLE=Partial Least Squares Enhances Genomic Prediction of New Environments

JOURNAL=Frontiers in Genetics

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2022.920689

DOI=10.3389/fgene.2022.920689

ISSN=1664-8021

ABSTRACT=In plant breeding, the need to improve the prediction of future seasons or new locations and/or environments, also denoted as “leave one environment out”, is of paramount importance to be able to increase the genetic gain in breeding programs and contribute to food and nutrition security worldwide. Genomic selection (GS) has the potential to increase the accuracy of future seasons or new locations because it is a predictive methodology. However, the majority of the statistical machine learning methods used for the task of predicting a new environment or season struggle to produce moderate or high prediction accuracies. For this reason, in this paper we explore the use of the partial least square (PLS) regression methodology for this specific task, and we benchmark its performance with the Bayesian Genomic Best Linear Unbiased Predictor (GBLUP) method. The benchmarking process was done with 12 real data sets. We found that in all data sets the PLS method outperformed the popular GBLUP method by margins between 49.84% (in the EYT_3 data) and 228.28% (in the Disease data) across traits, environments and types of predictor. Our results show a lot of empirical evidence of the power of PLS methodology for the prediction of future seasons or new environments.