Why AMMI is better ober regression based method
Pro Research Analysisby 
Searched over 200M research papers
Consensus Meter
Why AMMI is Better than Regression-Based Methods for Analyzing Genotype × Environment Interactions
Introduction to Genotype × Environment Interactions
Genotype × Environment (G×E) interactions are critical in agricultural research as they help in understanding how different genotypes perform across various environments. Two primary methods for analyzing these interactions are Joint Regression Analysis (JRA) and Additive Main Effects and Multiplicative Interaction (AMMI) models. This article explores why AMMI is often considered superior to regression-based methods.
Higher Interaction Sum of Squares Explained by AMMI
One of the key advantages of AMMI over JRA is its ability to explain a higher percentage of the interaction sum of squares (SS). In a study comparing the two methods, AMMI accounted for an average of 37% of the interaction SS, whereas JRA only explained about 11%. This significant difference indicates that AMMI provides a more detailed and accurate representation of G×E interactions.
Robustness Across Diverse Environments
AMMI's effectiveness remains consistent regardless of environmental diversity or the number of sites involved in the study. This robustness makes it particularly useful for large regional or international trials where environmental conditions can vary widely. In contrast, the performance of JRA can be inversely related to the number of sites, making it less reliable in diverse settings.
Superior Description of Genotype-Environment Effects
AMMI has been shown to be more valuable in describing genotype-environment (GE) effects across multiple datasets. In a study involving various cereals in Italy, AMMI outperformed JRA in six out of seven datasets for describing GE effects and in four datasets for genotype-location (GL) effects. This superiority was not crop-specific and was more pronounced when distinct environmental constraints affected genotype responses.
Consistent Ordination and Stability Measures
When both methods are appropriate, they often provide similar ordination of sites and genotypes for GL effects. However, AMMI tends to offer more consistent and repeatable stability measures across different cropping seasons. Measures like the Euclidean distance from the origin of significant interaction principal component (PC) axes (D) were found to be more repeatable than those provided by JRA.
Complementary Use with Other Models
AMMI can also be effectively used in conjunction with other statistical models like Partial Least Squares (PLS) and Factorial Regression (FR). These models incorporate external environmental and cultivar variables, which can be regressed on the environmental or genotypic scores estimated from AMMI. This complementary use allows for a more comprehensive understanding of G×E interactions.
Conclusion
In summary, AMMI models offer several advantages over regression-based methods like JRA for analyzing G×E interactions. They explain a higher percentage of interaction sum of squares, are robust across diverse environments, provide superior descriptions of GE effects, and offer consistent and repeatable stability measures. Additionally, AMMI can be effectively combined with other models for a more holistic analysis. Therefore, AMMI is recommended for detailed studies of G×E effects, especially in large and diverse trials.
Sources and full results
Most relevant research papers on this topic
Regression and AMMI analyses of genotype × environment interactions : an empirical comparison
Joint regression vs AMMI analysis of genotype-environment interactions for cereals in Italy
Using Partial Least Squares Regression, Factorial Regression, and AMMI Models for Interpreting Genotype × Environment Interaction
A comparison between joint regression analysis and the AMMI model: a case study with barley
Alternative regression models to assess increase in childhood BMI
A Comparison between Multiple Regression Models and CUN-BAE Equation to Predict Body Fat in Adults
Classification of Obesity among South African Female Adolescents: Comparative Analysis of Logistic Regression and Random Forest Algorithms
A Bayesian Shrinkage Approach for AMMI Models
Varietal stability analysis in Guangxi maize regional trial using AMMI model.
Physiological models of body composition and human obesity
Try another search
Why do we have to eat food that is not genetically modified?
What is the psychology of financial decisions?
What percentage of global carbon emissions does the electric industry account for?
what is time dilation
What is the potential of citizen science projects in contributing to scientific research and environmental monitoring?
Where does the moon go when it's not out at night?