PARTNERSHIP: LEVERAGING CO-EXPRESSION NETWORKS TO UNDERSTAND AND IMPROVE GENOTYPE-BY-ENVIRONMENT INTERACTION IN GENOMIC PREDICTION

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: ACTIVE
• Funding Source: AFRI COMPETITIVE GRANT
• Reporting Frequency: Annual
• Accession No.: 1032133
• Grant No.: 2024-67013-42588
• Cumulative Award Amt.: $719,731.00
• Proposal No.: 2023-11077
• Project Start Date: Jun 1, 2024
• Project End Date: May 31, 2027
• Grant Year: 2024
• Program Code: [A1141]- Plant Health and Production and Plant Products: Plant Breeding for Agricultural Production

Recipient Organization
UNIVERSITY OF ILLINOIS
2001 S. Lincoln Ave.
URBANA,IL 61801

Performing Department
(N/A)

Non Technical Summary
Genomic prediction and doubled haploids transformed hybrid maize breeding by increasing selection intensity and speeding up generation cycles. However, genomic prediction, which traditionally uses SNP markers, falls short in accurately modeling hybrid performance in specific environments. As a response to environmental stresses, plants dynamically modulate gene expression. We hypothesize that integrating gene expression modulation into genomic prediction will increase the accuracy of these models across environments. To test this hypothesis, we will evaluate gene expression modulation of maize inbreds and their hybrids responding to temperature (heat and cold) stress. Leveraging genotypic and phenotypic data previously generated from the Genomes2Fields collaborative Genotype×Environment project, we will determine how incorporating this information into genomic prediction models impacts prediction accuracy. Stochastic simulations will then be used to effectively find operational strategies to obtain and use gene expression information in breeding programs. We envision that the results of this project will shed light on the contribution of gene expression modulation to plasticity and environmental responsiveness and will be relevant to other crops. Ultimately, we will explore genomic selection methodologies that allow breeders to improve hybrid yield and agronomic performance more efficiently in specific environments to respond quickly to new stresses. This project will help guarantee adequate agricultural production despite a growing world population, pressure on natural resources, and climate change.

Animal Health Component

(N/A)

Research Effort Categories

Basic

70%

Applied

(N/A)

Developmental

30%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
203	1510	1081	50%
201	1510	1080	50%

Knowledge Area
203 - Plant Biological Efficiency and Abiotic Stresses Affecting Plants; 201 - Plant Genome, Genetics, and Genetic Mechanisms;

Subject Of Investigation
1510 - Corn;

Field Of Science
1081 - Breeding; 1080 - Genetics;

Keywords

maize

genomic

prediction

gene

co-expression

networks

genotype-by-environment

interactions

plasticity

Goals / Objectives
The changing agricultural landscape necessitates breeding crops that can thrive in diverse environments to ensure a sustainable food, feed, and fiber supply. While genomic prediction has revolutionized crop and animal breeding, current methods struggle to predict genotype performance across different environments (genotype-by-environment interaction), partly due to the limitations of genomic markers like SNPs. Understanding how gene expression varies in response to the environment (transcriptome variation) is crucial, as it influences an individual's performance within a specific environment. Previous research suggests that transcript abundance can predict performance as well as or better than SNP markers. Building on this, we will incorporate transcriptome profiles' variation in response to the environment, essential for improving genotype performance prediction across environments. The specific objectives are:Objective 1 - Assess variation in gene expression modulation across different environmental conditions.Objective 2- Determine the usefulness of information from transcriptional modulation and gene co-expression networks for genomic prediction.Objective 3 - Investigate the advantages and challenges of utilizing gene expression modulation information in maize breeding programs.

Project Methods
We hypothesize that integrating gene expression modulation into genomic prediction will increase the accuracy of these models across environments.To test this hypothesis, we will evaluate gene expression modulation of maize inbreds and their hybrids responding to temperature (heat and cold) stress. Leveraging genotypic and phenotypic data previously generated from the Genomes2Fields collaborative Genotype×Environment project, we will determine how incorporating this information into genomic prediction models impacts prediction accuracy. Stochastic simulations will then be used to effectively find operational strategies to obtain and use gene expression information in breeding programs.

Progress 06/01/24 to 05/31/25

Outputs
Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project has created training opportunities for a graduate student and a postdoctoral researcher in advanced statistical genetics and simulation-based breeding design. They gained experience in estimating variance components, developing and testing multi-kernel genomic prediction models, and conducting large-scale phenotyping. Both also worked in the summer nurseries, gaining practical experience in corn breeding. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?While the germplasm is being phenotyped, we will continue developing analysis and simulation pipelines. Publicly available datasets will be used to test and refine these pipelines. Plant material will be sampled and processed for RNAseq and transcriptome analysis. We expect that these efforts will provide the first expression data needed to integrate gene expression modulation into our genomic prediction models and will generate the foundation for the simulation studies outlined in Objectives 2 and 3.

Impacts
What was accomplished under these goals? For the first objective, we selected 21 ex-PVP inbreds and 75 elite hybrids that had been evaluated across 25 environments for yield and agronomic traits through the G2F project. While seed from those experiments was still available, we produced new seed in Summer 2024 to avoid potential problems with seed quality. Large-scale phenotyping of these inbreds and hybrids is underway and will provide the basis for transcriptome profiling under heat and cold stress. For the second objective, we worked with a larger panel of 162 single-cross hybrids derived from Stiff Stalk (SS) and Non-Stiff Stalk (NSS) inbred lines using a North Carolina Design II (NCII) mating scheme. The set of inbreds and hybrids used in Objective 1 is a subset of this population. We developed GBLUP-based multi-kernel models to partition additive variance from the general combining ability of SS and NSS lines and dominance variance from the specific combining ability of their interaction. We also used these models to predict the performance of untested single-crosses under different training set configurations. This work provided baseline estimates of genetic variance and genotype by environment interaction variances, which are critical for the simulations planned in Objectives 2 and 3. A manuscript describing this work is currently in preparation for submission The third objective focuses on building a simulation framework to test how incorporating gene expression data might improve breeding pipelines. Using the variance estimates from Objective 2, we started to set up stochastic simulations that compare conventional breeding strategies with approaches that make use of transcriptional modulation data.

Publications