Semiparametric odds ratio model for case-control and matched case-control designs.

Area of Science:

• Biostatistics
• Genetic Epidemiology
• Statistical Modeling

Background:

• Case-control studies are crucial for investigating disease etiology.
• Estimating gene-environment interactions is complex due to data limitations.
• Existing methods may lack efficiency or struggle with high-dimensional data.

Purpose of the Study:

• To develop an efficient semiparametric odds ratio model for gene-environment interaction analysis.
• To extend existing methods to accommodate both discrete and continuous data.
• To address challenges in estimating disease risk models by directly modeling gene-environment associations.

Main Methods:

• Proposed a semiparametric odds ratio model extending Umbach and Weinberg's approach.
• Directly modeled gene-environment association in controls to bypass difficult intercept estimation.
• Introduced a novel permutation-based approach for high-dimensional nuisance parameters in matched case-control designs.
• Demonstrated the model's reduction to conditional logistic regression under specific conditions.

Main Results:

• Simulation studies confirmed the proposed approach's good performance.
• The method effectively handles discrete and continuous gene-environment data.
• The permutation-based approach successfully eliminates nuisance parameters.

Conclusions:

• The proposed semiparametric model offers an efficient and robust method for gene-environment interaction studies.
• The approach is applicable to various data types and complex study designs.
• The method was successfully applied to a coronary artery disease gene-environment interaction study.