The logistic regression model for gene-environment interactions using both case-parent trios and unrelated...

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Gene-environment interactions are crucial for understanding complex diseases but are difficult to study.
• Existing methods often lack statistical power or face computational challenges, especially with mixed study designs.

Purpose of the Study:

• To propose a novel logistic regression framework for analyzing gene-environment interactions in mixed study samples (case-parent trios and unrelated subjects).
• To address convergence issues encountered in previous two-stage association analyses.

Main Methods:

• Developed a logistic regression model based on the combined haplotype relative risk (CHRR) method.
• Integrated case-parent trios and unrelated subjects into a unified analysis framework using a 2x2 table.
• Evaluated the model's performance and compared it with a two-stage approach via computer simulations.

Main Results:

• The proposed logistic regression model offers effortless adjustment for covariates.
• Both methods showed similar performance when the two-stage test converged in larger sample sizes.
• The two-stage test was more powerful for dominant and additive disease models, while the extended CHRR model was more powerful for recessive models.

Conclusions:

• The logistic regression framework provides a robust alternative for gene-environment interaction analysis, particularly when computational convergence is a concern.
• The choice between the two-stage test and the extended CHRR model depends on the specific genetic model (dominant, additive, or recessive).