Marginal interaction test for detecting interactions between genetic marker sets and environment in genome-wide...

Area of Science:

• Genetics
• Environmental Health
• Biostatistics

Background:

• Complex human diseases arise from gene-environment interactions (G×E), necessitating advanced analytical tools for risk prediction and mechanistic understanding.
• Current G×E analysis methods often focus on single genetic variants (common or rare) and a single environmental factor, limiting comprehensive investigation.
• Developing robust statistical methods is crucial for dissecting the complex interplay between genetic and environmental factors in disease etiology.

Purpose of the Study:

• To develop and validate novel statistical methods, MAGEIT_RAN and MAGEIT_FIX, for identifying gene-environment interactions (G×E).
• To enable the analysis of G×E involving both rare and common genetic variants simultaneously with an environmental factor.
• To apply these methods to identify novel G×E associated with hypertension and blood pressure in a multiethnic population.

Main Methods:

• Development of MAGEIT_RAN and MAGEIT_FIX, which utilize the MinQue statistic to detect G×E.
• MAGEIT_RAN models genetic main effects as random effects, while MAGEIT_FIX models them as fixed effects.
• Simulation studies were conducted to assess type I error control and statistical power, followed by genome-wide application to gene-alcohol interactions in the Multiethnic Study of Atherosclerosis.

Main Results:

• Both MAGEIT_RAN and MAGEIT_FIX demonstrated controlled type I error rates in simulations.
• MAGEIT_RAN exhibited superior statistical power compared to MAGEIT_FIX.
• Genome-wide analysis identified significant gene-alcohol interactions for hypertension and seated systolic blood pressure, highlighting genes such as EIF2AK2, CCNDBP1, and EPB42.

Conclusions:

• MAGEIT_RAN and MAGEIT_FIX are effective tools for identifying G×E, accommodating both rare and common variants.
• The application of MAGEIT_RAN revealed specific genes and pathways (apoptosis, signal transduction) involved in alcohol-related hypertension.
• These findings underscore the utility of MAGEIT_RAN in uncovering biologically relevant G×E for complex diseases.