Optimizing the allocation of resources for genomic selection in one breeding cycle

Area of Science:

• Quantitative genetics
• Plant breeding
• Bioinformatics

Background:

• Genomic selection (GS) enhances selection gain in plant breeding but optimal resource allocation remains unclear.
• Genotype × environment interactions (G×E) complicate GS strategies.
• Double haploid (DH) lines are crucial for biparental breeding populations.

Purpose of the Study:

• Develop a universally applicable planning tool for optimizing GS resource allocation.
• Integrate selection theory with numerical optimization to maximize selection gain under budget.
• Investigate the impact of G×E interactions on resource allocation strategies.

Main Methods:

• Combined quantitative genetic selection theory with constraint numerical optimization.
• Developed an extended selection accuracy formula accounting for G×E interactions.
• Modeled resource allocation for a single GS cycle using maize DH lines and grain yield data.

Main Results:

• The tool balances training set size, number of environments, and population size for optimal resource allocation.
• Genotyping costs are critical under small budgets; expanding prediction set size is key for large budgets.
• Using an index of phenotypic and GS values in the training set improves efficiency, especially with high G×E.

Conclusions:

• The developed framework provides a universally applicable tool for optimizing GS resource allocation.
• Effective resource allocation requires balancing training/prediction sets, environments, and G×E interactions.
• Reducing DH line production costs is vital for maximizing GS benefits with increasing budgets.