MEXCOwalk

MEXCOwalk identifies cancer driver gene modules by integrating protein-protein interaction (PPI) networks, gene mutation frequencies, and mutual exclusivity using an edge-weighted random walk.

Key Features:

• Integration of PPI networks: MEXCOwalk incorporates protein-protein interaction connectivity to assemble genes into putative functional modules.
• Incorporation of mutation data: The method uses gene mutation frequencies and mutual exclusivity patterns to prioritize candidate driver genes within modules.
• Edge-weighted random walk: It applies an edge-weighted random walk on the PPI network with edge weights derived from connectivity, mutual exclusivity, and coverage to detect modules.

Scientific Applications:

• Identification of cancer driver modules: MEXCOwalk recovers known cancer genes and uncovers putative driver genes that are infrequently mutated across cancer cohorts.
• Classification and enrichment analysis: Modules identified by MEXCOwalk can distinguish tumor versus normal samples and show enrichment for cancer-type-specific mutations.
• Patient risk stratification: Module-derived risk scores enable stratification of patients into low-risk and high-risk groups across multiple cancer types.
• Evaluation on TCGA pan-cancer data: The method has been benchmarked on The Cancer Genome Atlas (TCGA) pan-cancer datasets, demonstrating recovery of known cancer genes and identification of novel candidates.

Methodology:

MEXCOwalk computes edge weights from PPI connectivity, mutual exclusivity, and coverage, then performs an edge-weighted random walk on the PPI network while integrating gene mutation frequencies and mutual exclusivity to identify modules.