A data-smoothing approach to explore and test gene-environment interaction in case-parent trios

Area of Science:

• Genetics
• Epidemiology
• Biostatistics

Background:

• Complex traits arise from gene-environment interactions (G×E).
• Existing statistical tests for G×E may lack power when interaction forms deviate from assumptions.
• Understanding G×E is crucial for refining trait epidemiology.

Purpose of the Study:

• To develop a data-smoothing approach for estimating and testing G×E.
• To visualize the form of G×E between a single nucleotide polymorphism and a continuous environmental covariate.
• To provide a robust method for exploring novel interactions in complex trait research.

Main Methods:

• Utilized a generalized additive model with penalized likelihood for G×E estimation.
• Developed a permutation approach to test G×E, accounting for smoothing-induced uncertainty.
• Employed simulation studies to evaluate statistical properties and an example dataset for illustration.

Main Results:

• The proposed method provides point and interval estimates for G×E.
• A graphical display facilitates visualization and exploration of interaction forms.
• The permutation test effectively addresses uncertainty from the smoothing process.

Conclusions:

• The data-smoothing approach is effective for estimating and testing G×E.
• This method offers a valuable tool for exploring novel gene-environment interactions.
• The approach is particularly useful in data-rich settings for complex trait analysis.