Analysis of human mini-exome sequencing data from Genetic Analysis Workshop 17 using a Bayesian hierarchical mixture...

Area of Science:

• Genetic Epidemiology
• Statistical Genetics
• Genomics

Background:

• Next-generation sequencing (NGS) advances genetic epidemiology, enabling analysis of the full allele frequency spectrum in complex diseases.
• Traditional statistical methods struggle to estimate effects of low minor allele frequency (MAF) variants, necessitating new analytical approaches.

Purpose of the Study:

• To apply a Bayesian hierarchical mixture model for identifying genes associated with a simulated binary phenotype.
• To evaluate the model's utility in detecting both rare and common variants contributing to complex diseases.

Main Methods:

• Utilized the Genetic Analysis Workshop 17 dataset (697 unrelated individuals, 24,487 autosomal variants).
• Employed a Bayesian hierarchical mixture model with a transformed genotype design matrix weighted by allele frequencies.
• Implemented a Metropolis-Hastings algorithm and Gibbs sampling for parameter estimation.

Main Results:

• Identified 58 genes with a posterior probability > 0.8 for association with the simulated binary phenotype.
• Correctly identified the PIK3C2B gene as associated with affected status, validating the model's performance.
• Demonstrated the model's capability to detect genes with both rare and common variants.

Conclusions:

• Bayesian hierarchical mixture models offer a powerful approach for genetic association studies, particularly for complex diseases.
• The transformed genotype matrix effectively incorporates allele frequencies, improving the detection of low MAF variants.
• This methodology enhances the exploration of the genetic architecture underlying complex traits.