Network-guided sparse regression modeling for detection of gene-by-gene interactions

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genomics

Background:

• Genome-wide association studies (GWAS) explain limited heritability.
• Gene-by-gene interactions are a key source of unexplained heritability.
• Novel methods are needed to detect complex genetic interactions.

Purpose of the Study:

• To develop a novel penalized regression approach for detecting gene-gene interactions.
• To incorporate biological network information into interaction detection.
• To improve the identification of genetic factors contributing to heritability.

Main Methods:

• Utilized penalized regression and sparse estimation principles.
• Incorporated biological knowledge using a network-based penalty.
• Applied the method to simulated and real-world genetic data.

Main Results:

• The proposed method outperforms traditional stage-wise strategies in simulations, especially for weak main effects.
• Successfully identified interactions among human leukocyte antigen (HLA) genes in Framingham Heart Study data.
• These interactions are linked to total plasma immunoglobulin E (IgE) concentrations and allergy risk.

Conclusions:

• The novel network-penalized regression method is effective for detecting gene-gene interactions.
• This approach enhances the understanding of genetic architecture, particularly for complex traits like IgE levels.
• The findings have implications for allergy research and identifying genetic predispositions.