BAGSE

BAGSE performs Bayesian gene set enrichment analysis to jointly test and quantify gene set enrichment levels using a Bayesian hierarchical model and empirical Bayes inference, particularly for studies of complex diseases.

Key Features:

• Bayesian Hierarchical Model: Accounts for uncertainty in gene-level association evidence to produce more precise enrichment estimates.
• Empirical Bayes Inference: Implements an empirical Bayes framework fitted with an efficient Expectation-Maximization (EM) algorithm.
• Quantitative Enrichment Analysis: Provides both hypothesis testing and numerical estimation of gene set enrichment levels.
• Improved Power in Gene Discovery: Leverages enrichment estimates to increase power for identifying associated genes.
• Validation via Simulation Studies: Demonstrates accurate enrichment quantification while maintaining comparable power to state-of-the-art methods in simulations.
• Application to Real Data: Applied to differential expression experiments and transcriptome-wide association studies to aid identification of potentially causal pathways and gene networks.

Scientific Applications:

• Pathway identification: Identification of biological pathways enriched for associations in genetic and transcriptomic studies.
• Gene discovery: Enhancement of gene-level discovery by incorporating enrichment information into analysis.
• Differential expression analysis: Interpretation of enriched gene sets from differential expression experiments.
• Transcriptome-wide association studies (TWAS): Detection and prioritization of pathways and gene networks implicated by TWAS.
• Complex disease research: Investigation of mechanisms and potential therapeutic targets in studies of complex diseases.

Methodology:

BAGSE fits a Bayesian hierarchical model by empirical Bayes inference using an Expectation-Maximization (EM) algorithm, with simulation studies used for validation.