Optimizing the selection of quantitative traits in plant breeding using simulation.

Abstract

This review summarizes findings from simulation studies on quantitative traits in plant breeding and translates these insights into practical schemes. As agricultural productivity faces growing challenges, plant breeding is central to addressing these issues. Simulations use mathematical models to replicate biological conditions, bridging theory and practice by validating hypotheses early and optimizing genetic gain and resource use. While strategies can improve trait value, they reduce genetic diversity, making a combination of approaches essential. Studies emphasize the importance of aligning strategy with trait heritability and selection timing and maintaining genetic diversity while considering genotype-environment interactions to avoid biases in early selection. Using markers accelerates breeding cycles when marker placement is precise, foreground and background selection are balanced, and QTL are effectively managed. Genomic selection increases genetic gains by shortening breeding cycles and improving parent selection, especially for low heritability traits and complex genetic architectures. Regular updates of training sets are critical, regardless of genetic architecture. Bayesian methods perform well with fewer genes and in early breeding cycles, while BLUP is more robust for traits with many QTL, and RR-BLUP proves flexible across different conditions. Larger populations lead to greater gains when clear objectives and adequate germplasm are available. Accuracy declines over generations, influenced by genetic architecture and population size. For low heritability traits, multi-trait analysis improves accuracy, especially when correlated with high heritability traits. Updates including top-performing candidates, but conserving variability enhances gains and accuracy. Low-density genotyping and imputation offer cost-effective alternatives to high-density genotyping, achieving comparable results. Targeting populations optimizes genetic relationships, further improving accuracy and breeding outcomes. Evaluating genomic selection reveals a balance between short-term gains and long-term potential and rapid-cycling genomic programs excel. Diverse approaches preserve rare alleles, achieve significant gains, and maintain diversity, highlighting the trade-offs in optimizing breeding success.