Stochastic search gene suggestion: a Bayesian hierarchical model for gene mapping

Area of Science:

• Genetics and Genomics
• Statistical Genetics
• Computational Biology

Background:

• Mapping complex disease genes involves identifying multiple genetic loci, posing challenges with multiple testing and sparse data when analyzing haplotypes.
• Existing methods struggle with the complexity of genetic dependencies, particularly in regions like the human leukocyte antigen (HLA) complex.

Purpose of the Study:

• To develop a robust statistical framework for identifying genetic loci contributing to complex diseases.
• To address the challenges of multiple testing and sparse data in genetic association studies.
• To model genetic dependencies within the human leukocyte antigen (HLA) region.

Main Methods:

• Proposed a hierarchical Bayesian model utilizing case-parent triad data and a conditional logistic regression likelihood.
• Incorporated hierarchical prior distributions for allele effects to capture genetic dependencies in the HLA region.
• Implemented Bayesian variable selection for locus and allele selection, and included linkage disequilibrium as a covariance structure.

Main Results:

• The hierarchical Bayesian model effectively performs variable selection for both genetic loci and alleles.
• The model accounts for linkage disequilibrium, improving the accuracy of genetic marker identification.
• Simulations demonstrated the procedure's performance, and it was successfully applied to identify RA-associated genetic markers in the HLA region.

Conclusions:

• The proposed hierarchical Bayesian model offers a powerful approach for mapping genes involved in complex diseases.
• This method enhances the identification of relevant genetic markers by managing multiple testing and genetic dependencies.
• The developed software is publicly available for broader research application.