Utilizing Technology in a Hybrid Sorghum Breeding Program with the Implementation of Enviromics and Genomic Prediction

No Thumbnail Available

Date

2022-08-25

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Grain sorghum [Sorghum bicolor (L.) Moench] is an important crop native to Africa and grown in many subtropical and temperate regions worldwide. Its tropical origin means that most diverse sorghum is photoperiod sensitive and not suited to temperate grain production. The first chapter of this thesis evaluated two populations consisting of diverse introgression material derived from a modified backcross program that utilized an elite recurrent parent to determine if useful genetic variation resided in this material for sorghum improvement. These two populations were selected for agronomic relevance at the population level and the results revealed that both populations harbored lines that outperformed their recurrent parent. While there is material that outperformed the recurrent parent, most of the material did not. As such, the utility of genomic information to predict the hybrid performance of population members was evaluated prior to entering these lines into a breeding program. It was revealed that the use of genetic distance based on genomic information was not an effective way to screen material for hybrid performance, however genomic selection appeared to be an effective screening method to eliminate inferior-performing lines that will not be relevant for a hybrid breeding program. The second chapter of this thesis evaluated the inclusion of envirotypic data in genomic prediction models for multi-environment trials. Reaction norms informed by envirotype data may aid in modeling the differential responses of genotypes across multi-environment trials, and ultimately increase prediction accuracies for hybrid trials. A combination of genomic and enviromic information was applied to predict grain sorghum hybrid performance across standard U.S. production environments. Five different models were tested under three cross validation schemes that simulate challenges encountered by breeding programs. Of these models, the non-reaction norm GxE model produced the highest prediction accuracies, and the envirotype-informed reaction norm produced prediction accuracies comparable to the standard GxE model. The use of envirotype information can also characterize the environments targeted by the breeding program, thus assisting with the allocation of resources. Based on these observations, the utilization of envirotype data can expand the information available to a sorghum breeding program to create more robust genomic prediction models and for characterizing the target population of environments to aid in the effective selection of environments for hybrid evaluation.

Description

Keywords

Sorghum, Genomic Prediction, Introgression, Mega-Environments, Enviromics, GxE Effects,

Citation