Sparse testing using genomic prediction improves selection for breeding targets in elite spring wheat.

Area of Science:

• Agricultural Science
• Genetics
• Plant Breeding

Background:

• Efficient selection is crucial in crop breeding pipelines.
• Genomic prediction offers potential for optimizing selection strategies.
• Early yield testing stages require cost-effective evaluation methods.

Purpose of the Study:

• To evaluate sparse testing strategies using genomic prediction in wheat breeding.
• To determine the impact of sparse testing on selection intensity and budget.
• To assess the accuracy of different prediction methods in multi-environment trials.

Main Methods:

• Evaluation of sparse testing designs in a CIMMYT spring wheat breeding program.
• Analysis of prediction accuracy based on environmental overlap and genomic relationships.
• Comparison of genomic best linear unbiased prediction (GBLUP) with other methods.

Main Results:

• Sparse testing effectively increases testing environments without budget escalation.
• Substantial overlap between lines across environments is key for optimal prediction accuracy.
• Genetic correlations between environments and genomic relationships are primary drivers of accuracy.

Conclusions:

• Genomic prediction-based sparse testing is efficient for early-stage yield testing.
• GBLUP is recommended as the primary metric for selection decisions in early stages.
• This approach facilitates expanded use of selection environments and maintains selection intensity.