CDPath: Cooperative Driver Pathways Discovery Using Integer Linear Programming and Markov Clustering.

Area of Science:

• Computational Biology
• Cancer Genomics
• Systems Biology

Background:

• Understanding cancer pathogenesis requires identifying driver pathways.
• Cooperative interactions between multiple pathways are increasingly recognized in carcinogenesis.
• Existing methods may not fully capture pathway cooperation.

Purpose of the Study:

• To develop a novel computational approach, CDPath, for discovering cooperative driver pathways.
• To identify key driver genes and biological processes involved in cancer.
• To improve the discovery of pathway cooperation compared to existing methods.

Main Methods:

• Utilized Integer Linear Programming to identify driver core modules from mutation profiles.
• Enforced co-occurrence, functional interactions, mutual exclusivity, and coverage within modules.
• Applied Markov clustering on pathway-pathway interaction networks to group pathways.
• Identified cooperative driver pathways from distinct modules within the same clusters.

Main Results:

• Applied CDPath to breast cancer (BRCA) and endometrial cancer (UCEC) TCGA datasets.
• Successfully identified known driver genes (e.g., TP53) and potential novel drivers (e.g., SPTBN2).
• Uncovered cooperative driver pathways relevant to cancer, carcinogenesis, and key biological processes.
• Demonstrated superior performance over competitive approaches in uncovering biological associations and cooperative pathways.

Conclusions:

• CDPath effectively discovers cooperative driver pathways, offering insights into cancer pathogenesis.
• The identified pathways and genes hold potential for precise cancer treatment strategies.
• CDPath significantly enhances the discovery of pathway cooperation and associated biological mechanisms.