Application of multi-locus analytical methods to identify interacting loci in case-control studies

Area of Science:

• Genetic Epidemiology
• Statistical Genetics
• Bioinformatics

Background:

• Identifying interacting genetic loci is crucial for understanding complex diseases.
• Advanced multi-locus analysis methods are needed to detect gene-gene interactions.

Purpose of the Study:

• To compare the performance of four multi-locus methods: multiple logistic regression, sum statistics, logic regression, and multifactor dimensionality reduction (MDR).
• To evaluate the strengths and weaknesses of these methods using simulated case-control data with various two-locus interaction models.

Main Methods:

• Application of multiple logistic regression.
• Utilization of sum statistics.
• Implementation of logic regression.
• Employing the multifactor dimensionality reduction (MDR) method.
• Analysis of simulated case-control datasets with diverse two-locus interaction models.

Main Results:

• Sum statistics, logic regression, and MDR demonstrated good ability in identifying interacting loci.
• Multiple logistic regression performance varied based on the underlying model and adjustment procedures.
• All tested methods showed impaired identification of interacting loci for models with two two-locus interactions involving common disease alleles with small effects.

Conclusions:

• Sum statistics, logic regression, and MDR are effective for identifying interacting loci in genetic studies.
• Further research is needed to address limitations, particularly for detecting small-effect interactions.
• Practical and methodological considerations for applying these multi-locus methods are discussed.