Bayesian hierarchical modeling and selection of differentially expressed genes for the EST data

Area of Science:

• Bioinformatics
• Computational Biology
• Statistical Genetics

Background:

• Expressed sequence tag (EST) sequencing provides a snapshot of gene expression.
• Analyzing EST data across multiple libraries and tissues is crucial for understanding gene function.
• Existing methods may face challenges in accurately inferring gene expression levels and detecting differential expression.

Purpose of the Study:

• To develop a novel hierarchical multinomial model with a nonlinear Dirichlet prior for analyzing EST data.
• To introduce an efficient computational algorithm for posterior inference.
• To propose a new criterion for identifying differentially expressed genes between tissue types.

Main Methods:

• Hierarchical multinomial modeling with a nonlinear Dirichlet prior.
• Development of a Markov chain Monte Carlo (MCMC) algorithm for posterior computation.
• A new gene selection criterion for differential expression analysis.

Main Results:

• The proposed model and prior demonstrate favorable properties for EST data analysis.
• The MCMC algorithm enables efficient posterior computation.
• Simulations show the new gene selection criterion achieves low false negative and false positive rates.

Conclusions:

• The developed statistical framework offers an effective approach for analyzing multi-tissue EST data.
• The novel gene selection criterion enhances the accuracy of differential gene expression detection.
• The method is validated using simulations and a real-world EST dataset.