eQtlBma

eQtlBma applies Bayesian Model Averaging to detect expression quantitative trait loci (eQTLs) across multiple tissues or cell types and to estimate their tissue-specific and shared activity.

Key Features:

• Bayesian Model Averaging (BMA): Uses Bayesian Model Averaging to evaluate alternative models of eQTL activity across subgroups.
• Joint detection across subgroups: Performs joint statistical detection of eQTLs across multiple tissues or cell types rather than a tissue-by-tissue analysis.
• Modeling tissue-specific activity: Explicitly models each eQTL's active or inactive status per tissue to infer tissue-specific regulation.
• Formal estimation of shared eQTLs: Estimates model parameters representing the proportion of eQTLs shared among tissues.
• Increased detection power: In a re-analysis of transformed B cells, T cells, and fibroblasts, identified 63% more genes with eQTLs at a false discovery rate (FDR) of 0.05.

Scientific Applications:

• Cross-tissue eQTL discovery: Detect eQTLs that are shared across tissues or cell types.
• Characterization of tissue-specific regulation: Determine in which tissues or cell types individual eQTLs are active or inactive.
• Interpretation of genetic basis of complex traits: Provide insights into how genetic variation affects gene expression across tissues relevant to complex traits and diseases.

Methodology:

Implements Bayesian Model Averaging to compare models of eQTL activity across subgroups, explicitly models active/inactive status per tissue, and estimates parameters for shared eQTLs; reported performance includes a 63% increase in detected eQTL-containing genes at FDR 0.05 in transformed B cells, T cells, and fibroblasts.