Random regression models for detection of gene by environment interaction.

Area of Science:

• Quantitative genetics
• Statistical genomics
• Bioinformatics

Background:

• Quantitative trait loci (QTL) analysis is crucial for understanding genetic contributions to complex traits.
• Gene by environment (GxE) interactions significantly influence phenotypic expression.
• Accurate QTL detection and characterization require robust statistical models.

Purpose of the Study:

• To evaluate two random regression models for QTL detection and GxE interaction analysis.
• To compare random regression models with a standard model lacking GxE interaction terms.
• To assess the impact of model parameter constraints on QTL detection power.

Main Methods:

• Development and application of two random regression models for QTL analysis.
• Regression of putative QTL effects on a simulated environmental gradient.
• Simulation study using a granddaughter design with varying QTL effect models.
• Comparison of model performance in QTL detection, positioning, and GxE interaction identification.

Main Results:

• Random regression models demonstrated suitability for detecting QTL effects under various environmental conditions.
• The model restricting the correlation between intercept and slope showed increased power when QTL effects were re-ranked across environments.
• Accurate QTL positioning and GxE interaction detection were improved by the random regression approaches.

Conclusions:

• Random regression models are effective tools for dissecting genetic architecture influenced by environmental factors.
• Constraining the correlation parameter in random regression models enhances power for detecting dynamic QTL effects.
• These models provide a flexible framework for integrating genotype and environmental data in genetic studies.