Optimising desired gain indices to maximise selection response.

Area of Science:

• Plant breeding
• Quantitative genetics
• Genomic selection

Background:

• Improving multiple traits simultaneously is crucial in plant breeding.
• Desired gain selection indices offer a way to prioritize trait improvement without economic weights.
• Traditional indices may not maximize selection response or correlate optimally with genetic merit.

Purpose of the Study:

• To develop an iterative desired gain selection index method.
• To optimize selection response for multiple traits, with user-specified constraints.
• To maximize selection response and improve correlation with net genetic merit.

Main Methods:

• Developed an iterative method for desired gain selection indices.
• Optimized desired gain values for targeted or user-specified selection response.
• Applied the method to genomic estimated breeding values (GEBVs) for seven traits in bread wheat.

Main Results:

• Achieved prediction accuracies between 0.29 and 0.47 for seven traits in wheat.
• The iterative method matched user-specified responses when constraining traits.
• Unconstrained iterative indices maximized selection response and shifted GEBVs towards desired directions, outperforming traditional methods.

Conclusions:

• The iterative desired gain selection index is effective when economic weights are unknown.
• This method provides an optimal approach for plant breeding, especially when constraining selection response is not feasible.
• It enhances the ability to maximize selection response and achieve targeted genetic improvements.