A simple loglinear model for haplotype effects in a case-control study involving two unphased genotypes

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Haplotypes are valuable for summarizing gene effects on disease.
• Existing methods for haplotype analysis in case-control studies have limitations.
• Previous methods do not adequately address Hardy-Weinberg equilibrium deviations or gene-environment interactions.

Purpose of the Study:

• To propose a novel statistical method for haplotype analysis in case-control studies.
• To generalize existing models to incorporate departures from Hardy-Weinberg equilibrium and haplotype-covariate interactions.
• To provide a robust framework for analyzing unphased genotype data.

Main Methods:

• A new method generalizing the Epstein and Satten model is introduced.
• The method utilizes a single loglinear design matrix for parameter estimation.
• It accounts for haplotype frequencies in controls, disease association, and covariate interactions.

Main Results:

• Simulations demonstrate the method's effectiveness with realistic sample sizes.
• The approach is recommended for two-genotype data, recessive/dominant models, and common haplotypes (>10%).
• Modeling departures from Hardy-Weinberg equilibrium is crucial, even without covariates.

Conclusions:

• The proposed method offers a significant advancement for haplotype-based disease association studies.
• It is particularly useful for candidate gene analysis and exploring genotype pair interactions.
• The methodology enhances the accurate inference of genetic effects in complex disease research.