Training Set Optimization for Sparse Phenotyping in Genomic Selection: A Conceptual Overview.

Area of Science:

• Agricultural Science
• Genetics
• Plant Breeding

Background:

• Genomic selection (GS) is vital for increasing genetic gain per unit time in breeding programs.
• The training reference set (TRS) is a critical component of GS implementation.
• Optimizing TRS is essential for the efficiency, effectiveness, and cost reduction of breeding programs.

Purpose of the Study:

• To review lessons learned from training reference set optimization studies in crop breeding.
• To discuss key features for successful TRS optimization under various scenarios.
• To provide guidelines for improving genetic improvement rates using TRS optimization in GS.

Main Methods:

• Literature review of training reference set optimization studies in genomic selection.
• Analysis of challenges in GS, including population size, training-test set relationships, and TRS updates.
• Discussion of different packages and algorithms for TRS implementation.

Main Results:

• Optimally selected TRS yield higher prediction accuracies than arbitrary TRS.
• TRS optimization can significantly reduce the number of genotypes needing phenotyping, lowering costs and time.
• Lessons learned highlight the importance of TRS design for effective GS.

Conclusions:

• Training reference set optimization is a key strategy for enhancing genomic selection efficiency and effectiveness.
• Addressing challenges like population size and TRS updates is crucial for successful optimization.
• Implementing TRS optimization guidelines can maximize genetic improvement in breeding programs.