OMEN

OMEN identifies driver genes and driver modules within biological interaction networks by leveraging mutual exclusivity and a logic programming framework grounded in random walk semantics to integrate gene-specific and gene-set properties in their network context.

Key Features:

• Mutual exclusivity via functional impact scores: Uses functional impact scores of mutations to implement mutual exclusivity constraints for driver detection.
• Logic programming framework: Employs a logic programming framework as the core computational formalism.
• Random walk semantics: Grounds the framework in random walk semantics to model information propagation on networks.
• Integration of gene-specific and gene-set properties: Combines gene-level attributes and broader gene-set characteristics within their network contexts.
• Dynamic gene-set adjustment: Dynamically adjusts gene-set properties based on network context to refine specificity.
• Sensitivity to rare drivers: Avoids restrictive a priori filtering of genes to increase sensitivity for detecting rare driver genes.
• Mitigation of propagation artifacts: Accounts for network topology and high-scoring nodes to reduce spurious predictions from propagation-based approaches.

Scientific Applications:

• Identification of individual driver genes: Detects candidate driver genes, including rare drivers, within complex biological networks.
• Discovery of driver modules and pathways: Identifies clusters and modules that act as proxies for broader driver pathways.
• Benchmarking on TCGA datasets: Has been applied to The Cancer Genome Atlas (TCGA) benchmark datasets to evaluate performance in driver and module identification.

Methodology:

Uses a logic programming framework grounded in random walk semantics; integrates gene-specific properties and gene-set characteristics within network contexts; applies mutual exclusivity using functional impact scores; dynamically adjusts gene-set properties and avoids restrictive a priori gene filtering.