BAGSE: a Bayesian hierarchical model approach for gene set enrichment analysis

Area of Science:

• Computational Biology
• Statistical Genetics
• Bioinformatics

Background:

• Gene set enrichment analysis (GSEA) identifies biological pathways in complex diseases.
• Current GSEA methods struggle with accurate enrichment quantification, limiting downstream analyses.

Purpose of the Study:

• To introduce Bayesian Analysis of Gene Set Enrichment (BAGSE), a novel computational method.
• To provide a rigorous approach for GSEA emphasizing hypothesis testing and enrichment estimation.

Main Methods:

• Developed a Bayesian hierarchical model accounting for gene association uncertainty.
• Employed an empirical Bayes inference framework with an efficient EM algorithm.
• Validated through simulation studies and real-world data from differential expression and TWAS.

Main Results:

• BAGSE provides accurate enrichment quantification with power comparable to state-of-the-art methods.
• BAGSE enhances gene discovery power by leveraging enrichment information.
• Demonstrated effectiveness in identifying potentially causal pathways and gene networks in real data.

Conclusions:

• BAGSE offers a robust statistical framework for gene set enrichment analysis.
• The method accurately quantifies enrichment and aids in discovering biologically relevant pathways.
• BAGSE is freely available, promoting reproducibility and further research.