Functional logistic regression approach to detecting gene by longitudinal environmental exposure interaction in a...

Area of Science:

• Genetics
• Epidemiology
• Biostatistics

Background:

• Complex diseases arise from gene-environment (G × E) interactions, crucial for understanding disease mechanisms and improving risk prediction.
• Identifying G × E interactions is challenging due to limited statistical power and static exposure measurements, despite many environmental factors changing over time.
• Existing methods lack the capacity to model interactions between genes and time-varying environmental exposures.

Purpose of the Study:

• To propose a novel statistical method for detecting gene by time-varying environmental exposure interactions.
• To develop a model that accommodates longitudinal environmental exposures, irregular measurements, and measurement errors.
• To improve the understanding of disease mechanisms and risk prediction by incorporating dynamic G × E interactions.

Main Methods:

• Development of a functional logistic regression (FLR) approach utilizing functional data analysis.
• Modeling the time-varying effect of longitudinal environmental exposure and its interaction with genetic factors on disease risk.
• Extensive simulations to evaluate the method's statistical power and Type I error control, and application to a pancreatic cancer case-control study.

Main Results:

• The proposed FLR method demonstrates robust control of Type I error and superior power compared to alternative methods in simulations.
• The study successfully identified critical time windows for body mass index (BMI) exposure influencing pancreatic cancer risk.
• Specific genes that may modify the association between lifetime BMI and pancreatic cancer risk were identified.

Conclusions:

• The functional logistic regression (FLR) approach is a powerful tool for analyzing gene by time-varying environmental exposure interactions.
• This method enhances the ability to identify dynamic G × E interactions, offering new insights into disease etiology.
• The findings provide a foundation for more accurate disease risk prediction and targeted interventions by considering the temporal nature of environmental exposures.