Strategies for Selecting Crosses Using Genomic Prediction in Two Wheat Breeding Programs.

Area of Science:

• Plant breeding
• Quantitative genetics
• Genomic selection

Background:

• Cross selection is paramount in plant breeding for achieving genetic gain.
• Balancing progeny performance and genetic diversity is key for sustained improvement.
• Predictive models can optimize cross selection in wheat breeding programs.

Purpose of the Study:

• To compare cross-prediction methods using mid-parent value and progeny variance.
• To evaluate the impact of linkage disequilibrium on variance prediction.
• To assess the relative importance of mid-parent value and progeny variance for yield and quality traits in wheat.

Main Methods:

• Predicted progeny mean and variance for crosses using mid-parent value and variance prediction (with and without linkage disequilibrium).
• Selected crosses based on mid-parent value, top 10% progeny prediction, and weighted mean/variance.
• Applied methods to wheat breeding programs at INIA Uruguay and CIMMYT Mexico for grain yield and quality traits.

Main Results:

• Mid-parent values were the primary drivers of genetic gain for grain yield.
• Progeny variance had a limited impact on genetic gain for grain yield but was more influential for quality traits.
• Transgressive segregation can be enhanced by considering progeny variance.

Conclusions:

• Genomic resources and statistical models now enable accurate prediction of breeding line performance and cross potential.
• Plant breeders can leverage these tools to optimize cross selection for both yield and quality traits.
• The relative importance of progeny variance varies between yield and quality traits in wheat breeding.