MDPFinder

MDPFinder identifies mutated driver pathways, genes, and mutations de novo from cancer mutation and gene expression data to distinguish driver elements from passenger mutations and elucidate molecular mechanisms underlying cancer.

Key Features:

• Maximum Weight Submatrix Problem: Solves the maximum weight submatrix problem to de novo identify mutated driver pathways from cancer mutation data.
• Exact Method: Provides an exact method that serves as a benchmark for evaluating approximate or heuristic algorithms and yields precise pathway identifications.
• Stochastic and Flexible Method: Implements a stochastic and flexible method that can incorporate additional information, notably integrating mutation data with gene expression profiles.
• Integrative Model: Combines mutation and gene expression data to identify more biologically relevant gene sets.
• Passenger Mutation Filtering: Distinguishes driver mutations from unfunctional and passenger mutations.
• Implementation: Implemented as a MATLAB package.

Scientific Applications:

• Head and neck squamous cell carcinoma: Applied to mutation profiles from 74 head and neck squamous cell carcinoma samples to identify driver pathways.
• Glioblastoma: Applied to 90 glioblastoma tumor samples with gene expression profiles considered for integrated analysis.
• Ovarian carcinoma: Applied to 313 ovarian carcinoma samples incorporating gene expression data.
• Clinical diagnostics and prognostics: Provides insights into molecular mechanisms relevant to diagnostics and prognostics in oncology.
• Targeted therapeutics and personalized medicine: Identifies driver pathways to support the development of targeted therapies and personalized treatment strategies.

Methodology:

Solves the maximum weight submatrix problem using two methods: an exact method and a stochastic flexible method that integrates mutation data with gene expression profiles.