Leveraging probability concepts for cultivar recommendation in multi-environment trials

Area of Science:

• Agricultural Science
• Biotechnology
• Statistical Genetics

Background:

• Understanding genotype-by-environment interaction (GEI) is vital for successful plant breeding.
• Phenotypic plasticity influences genotype performance across diverse environments.
• Accurate cultivar recommendation requires robust statistical approaches.

Purpose of the Study:

• To integrate Bayesian probability concepts into stability analysis for GEI.
• To develop a framework for informed cultivar recommendation in multi-environment trials.
• To quantify adaptation and stability probabilities using Bayesian methods.

Main Methods:

• Utilized Bayesian models and probability methods for stability analysis.
• Employed the No-U-Turn sampler for Hamiltonian Monte Carlo estimation.
• Applied posterior distributions to estimate adaptation and stability probabilities.
• Tested the models on two empirical tropical datasets.

Main Results:

• Successfully untangled genotype-by-environment interaction (GEI) using Bayesian probability.
• Generated estimates for adaptation and stability probabilities.
• Demonstrated the utility of the approach in tropical agricultural datasets.
• Provided a basis for considering recommendation uncertainty.

Conclusions:

• Bayesian probability methods offer a powerful tool for unraveling GEI.
• This framework enhances decision-making for cultivar recommendation in multi-environment trials.
• The approach accounts for uncertainty in genotype adaptation and stability.