Simulations of genomic selection implementation pathways in common bean (Phaseolus vulgaris L.) using parametric and...

Area of Science:

• Plant Breeding
• Genetics
• Agricultural Science

Background:

• Genomic selection (GS) is a powerful tool for crop improvement.
• Optimizing GS implementation in breeding programs is crucial for maximizing genetic gains.
• Common bean (Phaseolus vulgaris L.) breeding programs can benefit from advanced genomic strategies.

Purpose of the Study:

• To simulate and evaluate different genomic selection (GS) implementation strategies for common bean breeding.
• To understand the impact of model choice, training data, and selection timing on breeding program outcomes.
• To assess the potential of parametric versus nonparametric models for predicting breeding values and achieving genetic gain.

Main Methods:

• Simulated 24 genomic selection (GS) pathways for common bean breeding.
• Varied training model generation, parent selection generation, and training data collection generation.
• Compared parametric (ridge regression) and nonparametric (artificial neural network) models for estimating breeding values.

Main Results:

• Early generation parent selection (rapid-cycle GS) led to higher gains over three cycles compared to late-generation selection.
• Parametric models offered consistent prediction accuracy, while nonparametric models showed variable performance.
• Nonparametric models demonstrated a better balance between allelic diversity and genetic gains, crucial for sustained improvement.

Conclusions:

• Early generation parent selection is effective for increasing seed yield in common bean.
• Nonparametric models show potential for common bean breeding but require further investigation to stabilize performance.
• Sustained genetic gains depend on the renewal of genetic variance, highlighting the importance of diverse training data and model selection.