A partially linear tree-based regression model for assessing complex joint gene-gene and gene-environment effects.

Area of Science:

• Genetic Epidemiology
• Statistical Genetics
• Computational Biology

Background:

• Analyzing complex diseases requires understanding joint gene effects, which is challenging for standard regression models.
• Higher-order gene-gene interactions are difficult to detect with traditional statistical methods.
• Existing tree-based methods struggle to model additive main effects.

Purpose of the Study:

• To propose a novel statistical framework, partially linear tree-based regression (PLTR) models, for analyzing complex genetic data.
• To combine the strengths of generalized linear regression and tree models for enhanced interaction analysis.
• To develop methods for fitting and validating the proposed PLTR models.

Main Methods:

• Introduced partially linear tree-based regression (PLTR) models combining linear main effects and non-parametric tree structures.
• Developed an iterative algorithm for fitting PLTR models.
• Proposed a unified resampling approach for optimal tree selection and significance testing.

Main Results:

• Simulation studies confirmed the resampling procedure's ability to maintain correct type I error rates.
• The PLTR model successfully summarized joint effects of 53 single nucleotide polymorphisms (SNPs) associated with biliary stone risk.
• The model demonstrated utility in exploring gene-environment interactions.

Conclusions:

• PLTR models offer a powerful approach for dissecting complex disease genetics by effectively modeling joint gene effects.
• The proposed statistical framework and validation methods advance the application of tree methodology in genetic epidemiology.
• This work provides a parsimonious yet comprehensive tool for understanding genetic and environmental risk factors in complex diseases.