Identifying gene-environment and gene-gene interactions using a progressive penalization approach.

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Identifying gene-environment (GxE) and gene-gene (GxG) interactions is crucial for understanding complex diseases.
• Traditional statistical methods and penalization techniques often fail to respect the hierarchical structure of biological interactions.
• Existing methods that accommodate hierarchical structures are computationally intensive and limited in scalability.

Purpose of the Study:

• To develop a computationally efficient penalization method for detecting significant GxE and GxG interactions.
• To propose a method that inherently respects the "main effect, interaction" hierarchical structure in statistical models.
• To provide a scalable and versatile tool applicable to various data types and regression models.

Main Methods:

• A stagewise penalization approach that progressively expands the optimization domain to capture hierarchical interactions.
• Utilizing a coordinate descent algorithm to efficiently compute the entire regularized solution path.
• The method is designed to be applicable across multiple data types and statistical models.

Main Results:

• The proposed method demonstrates superior performance in simulation studies compared to existing approaches.
• It successfully identifies important GxE interactions in a lung cancer prognosis study using gene expression data.
• The method is computationally fast and respects the hierarchical nature of gene interactions.

Conclusions:

• The developed stagewise penalization method offers an efficient and effective solution for identifying complex gene interactions.
• This approach overcomes the limitations of existing methods regarding computational complexity and scalability.
• The method has practical implications for genomic research, particularly in disease prognosis and etiological studies.