Detection of epistatic effects with logic regression and a classical linear regression model

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Traditional methods like Cockerham's model are used to identify interacting quantitative trait loci (QTLs).
• Cockerham's model defines interactions as effects beyond additive gene contributions.
• This model struggles to detect complex Boolean interactions influencing phenotypes, such as diseases.

Purpose of the Study:

• To introduce and evaluate a logic regression framework for more efficient detection of multiple interacting QTLs.
• To address the limitations of existing models in identifying complex, non-additive genetic interactions.

Main Methods:

• Development and application of a logic regression framework to model gene-gene interactions.
• Analytical demonstration of increased power for a two-way interaction model.
• Validation through simulation studies and real-world data analysis.

Main Results:

• Logic regression models efficiently represent higher-order logic interactions.
• This approach significantly increases statistical power for detecting complex QTL interactions compared to Cockerham's model.
• The enhanced power was confirmed analytically, in simulations, and with real data.

Conclusions:

• Logic regression provides a superior framework for detecting complex genetic interactions.
• It offers increased power and efficiency, especially for Boolean-type gene effects.
• This method enhances the ability to identify multiple interacting QTLs influencing traits.