Skip to main content
ARS Home » Northeast Area » Ithaca, New York » Robert W. Holley Center for Agriculture & Health » Plant, Soil and Nutrition Research » Research » Publications at this Location » Publication #308927

Title: Increased prediction accuracy in wheat breeding trials using a marker x environment interaction genomic selection model

Author
item CRUZ, MARCO LOPEZ - International Maize & Wheat Improvement Center (CIMMYT)
item CROSSA, JOSE - International Maize & Wheat Improvement Center (CIMMYT)
item BONNETT, DAVID - International Maize & Wheat Improvement Center (CIMMYT)
item DREISIGACKER, SUSANNE - International Maize & Wheat Improvement Center (CIMMYT)
item Poland, Jesse
item Jannink, Jean-Luc
item SINGH, RAVI - International Maize & Wheat Improvement Center (CIMMYT)
item DE LOS CAMPOS, GUSTAVO - University Of Alabama

Submitted to: Genes, Genomes, Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/3/2015
Publication Date: 4/1/2015
Publication URL: http://DOI: 10.1534/g3.114.016097
Citation: Cruz, M., Crossa, J., Bonnett, D., Dreisigacker, S., Poland, J.A., Jannink, J., Singh, R., De Los Campos, G. 2015. Increased prediction accuracy in wheat breeding trials using a marker x environment interaction genomic selection model. Genes, Genomes, Genetics. 5(4):569-582.

Interpretive Summary: Genomic selection (GS) models use genome-wide genetic information to predict genetic values of candidates for selection. Originally these models were developed without considering genotype ' environment interaction (GE). In this study, we introduce a simple marker-by-environment interaction model where correlation between environments is modeled using the variance of the main effects of the markers. The proposed model is conceptually simple and can be implemented using existing software. We discuss two ways the model can be implemented. We illustrate the use of the interaction model using three CIMMYT wheat data sets whose lines were evaluated over three years under controlled environmental conditions with different irrigation levels and planting systems. Our results show substantial gains in prediction accuracy whenever the environments analyzed were positively correlated. The interaction model presented here should be useful when selecting for stability as well as when selecting for targeted environments.

Technical Abstract: Genomic selection (GS) models use genome-wide genetic information to predict genetic values of candidates for selection. Originally these models were developed without considering genotype ' environment interaction (GE). Several authors have proposed extensions of the cannonical GS model that accommodate GE using either co-variance functions or environmental data. In this study, we introduce a simple marker-by-environment interaction model where co-variance between environments is modeled using the variance of the main effects of the markers. The proposed model is conceptually simple and can be implemented using existing software for GS. We discuss how the model can be implemented by using explicit regression of phenotypes on markers or a GBLUP-type model. We illustrate the use of the interaction model using three CIMMYT wheat data sets whose lines were genotyped using genotyping-by-sequencing and evaluated over three years (W1, W2 and W3) under controlled environmental conditions with different irrigation levels, planting types and systems. Our results show substantial gains in prediction accuracy whenever the environments analyzed were positively correlated. The model decomposes genomic values into components that are stable across environments and others that are environment-specific; therefore, the interaction model presented here should be useful when selecting for stability as well as when selecting for targeted environments.