IHRW

IHRW predicts efficacious three-drug combinations by performing random walks on hypergraphs to model higher-order drug interactions for triple-drug therapy discovery.

Key Features:

• Higher-Order Information Utilization: Uses hypergraphs to capture interactions among three drugs simultaneously, representing higher-order relationships beyond pairwise graphs.
• Novelty in Three-Drug Prediction: Targets prediction of triple drug therapies, addressing a gap where approximately 21% of known effective drug combinations involve three drugs.
• Enhanced Predictive Power: Demonstrated the ability to identify promising triple-drug options in case studies involving breast cancer, lung cancer, and colon cancer.
• Random Walk-Based Inference: Employs random walks on hypergraph structures to explore potential synergistic effects among drug triplets.

Scientific Applications:

• Combination Therapy Discovery: Identification and prioritization of three-drug combinations for complex diseases.
• Cancer Combination Strategy: Applied to breast cancer, lung cancer, and colon cancer to identify promising triple-drug therapy options.

Methodology:

Constructs hypergraphs to represent higher-order drug interactions and performs random walks across these hypergraphs to explore and predict synergistic effects among triplets of drugs.