A comparison of restricted selection index and linear programming in sire selection

Area of Science:

• Animal Breeding and Genetics
• Quantitative Genetics
• Mathematical Modeling in Biology

Background:

• Restricted selection index aims to improve one trait while maintaining another at zero change.
• Linear programming offers a different approach to optimize functions under constraints.

Purpose of the Study:

• To compare the efficacy of restricted selection index and linear programming.
• To evaluate their ability to maximize genetic gain in one trait while restricting change in a second trait to zero.

Main Methods:

• A numerical study was conducted to simulate genetic selection scenarios.
• Both restricted selection index and linear programming were applied to the same dataset.
• Performance was evaluated based on maximizing response in the target trait and minimizing deviation in the restricted trait.

Main Results:

• Linear programming demonstrated superior effectiveness in limiting correlated responses compared to restricted selection index.
• While both methods achieved near-zero change in the restricted trait, linear programming resulted in a smaller squared deviation.
• Restricted selection index yielded a greater response in the unrestricted trait than linear programming.

Conclusions:

• Linear programming is a more robust method for controlling correlated genetic responses in breeding programs.
• The choice between methods depends on the priority: precise control of a secondary trait (linear programming) or maximizing response in the primary trait (restricted selection index).
• Further research could explore hybrid approaches or applications in specific livestock or crop breeding contexts.