Linkage analysis with gene-environment interaction: model illustration and performance of ordered subset analysis

Area of Science:

• Genetics and Bioinformatics
• Statistical Genetics
• Computational Biology

Background:

• Complex human diseases often involve genetic heterogeneity, including gene-environment (G x E) interactions.
• Ordered subset analysis (OSA) is a method to incorporate covariates into linkage analysis, aiming to reduce heterogeneity.
• Evaluating OSA's performance in the presence of G x E interactions is crucial for understanding its utility in genetic studies.

Purpose of the Study:

• To evaluate the performance of ordered subset analysis (OSA) in simulation studies where the disease model includes gene-environment (G x E) interactions.
• To assess OSA's power to detect linkage and improve efficiency in identifying disease-associated genetic regions under G x E interaction scenarios.

Main Methods:

• Simulation study designed to model complex diseases influenced by G x E interactions, with and without main genetic or environmental effects.
• Graphical illustration of complex simulation models to clarify disease-generating mechanisms.
• Application of OSA to simulated datasets to assess its power and efficiency compared to standard linkage analysis on the entire dataset.

Main Results:

• OSA demonstrated underpowered performance for detecting small to moderate G x E interaction effects.
• When interaction effects were large, standard linkage methods showed sufficient power, and OSA's ability to improve linkage scores depended on the genetic model and allele frequency.
• For strong G x E interactions (RR(G x E) = 10) in ≥70% of families, OSA achieved >70% power to detect subsets with significantly greater linkage evidence, capturing most linked families and increasing per-genotype efficiency by 20-30%.

Conclusions:

• Ordered subset analysis (OSA) is underpowered for detecting weak to moderate gene-environment interactions in genetic linkage analysis.
• OSA can effectively identify informative subsets of families for follow-up studies when gene-environment interactions are strong and prevalent.
• The utility of OSA in improving linkage analysis efficiency is contingent upon the strength of G x E interactions and the underlying genetic architecture.