Tests for candidate-gene interaction for longitudinal quantitative traits measured in a large cohort

Area of Science:

• Genetics
• Biostatistics
• Cardiovascular Disease Research

Background:

• Gene-gene interactions significantly influence complex quantitative traits.
• Accurate statistical methods are crucial for identifying these interactions in large cohort studies.

Purpose of the Study:

• To compare the effectiveness of a longitudinal nonparametric association test (LNPT) and survival analysis for detecting gene-gene interactions in quantitative traits.
• To identify specific gene-gene interactions affecting body mass index (BMI) and coronary artery calcification (CAC) in Framingham Heart Study (FHS) data.

Main Methods:

• Application of LNPT and survival analysis to Framingham Heart Study (FHS) and simulated FHS (FHSsim) datasets.
• Testing for pairwise gene-gene interactions on quantitative traits, including body mass index and coronary artery calcification.
• Validation of type I error rates and comparison of statistical power between the two methods.

Main Results:

• Both LNPT and survival analysis demonstrated well-controlled type I error rates.
• The methods identified significant pairwise non-additive effects for genes FTO, PON1, and PFKP on body mass index in the FHS Offspring Cohort.
• Comparative power analysis using FHSsim data indicated the performance of both methods in detecting gene interactions for coronary artery calcification.

Conclusions:

• LNPT and survival analysis are reliable methods for detecting gene-gene interactions in quantitative traits.
• Identified interactions involving FTO, PON1, and PFKP highlight their roles in the genetic architecture of body mass index.
• These findings contribute to understanding the genetic basis of cardiovascular risk factors.