Fitting a regression model for genotype-by-environment data on heading dates in grasses by methods for nonlinear...

Area of Science:

• Agricultural Science
• Biometrics
• Plant Breeding

Background:

• Genotype-by-environment interaction complicates agricultural crop variety trial analysis.
• The Yates and Cochran regression approach is a common method for analyzing these interactions.
• This approach involves regressing genotype performance on environmental means.

Purpose of the Study:

• To modify the existing regression model by exchanging the roles of genotypes and environments.
• To assess the adequacy of the modified model for analyzing crop trial data.
• To discuss parameter estimation methods for the modified model within the framework of nonlinear mixed models.

Main Methods:

• Modification of the Yates and Cochran regression model by swapping genotype and environment roles.
• Application of diagnostic plots to evaluate model fit.
• Consideration of the model as a nonlinear mixed model with environments as random and genotypes as fixed factors.
• Discussion of parameter estimation using maximum likelihood and Taylor series expansion methods.

Main Results:

• The modified regression model demonstrated adequacy for analyzing heading dates in Dactylis glomerata.
• Diagnostic plots supported the suitability of the proposed model modification.
• The study explored the connection between the modified model and nonlinear mixed models.

Conclusions:

• The modified regression model offers a viable approach for analyzing genotype-by-environment interactions in crop trials.
• The model's flexibility extends to nonlinear mixed model frameworks.
• Effective parameter estimation is achievable through established statistical methods like maximum likelihood.