This research introduces K-Paths, a framework that harnesses biomedical knowledge graphs to accelerate drug discovery and repurposing by identifying meaningful relationships between drugs and diseases.

• Extracts biologically significant paths from complex knowledge graphs without requiring specialized graph neural networks
• Demonstrates superior performance in predicting drug-disease relationships and potential drug interactions
• Provides interpretable results that align with known biological mechanisms
• Achieves state-of-the-art results on benchmark datasets while maintaining compatibility with various model architectures

With pharmaceutical R&D costs exceeding $2.6B per drug, K-Paths offers a cost-effective alternative to traditional drug discovery by identifying new therapeutic uses for existing medications and predicting potential drug interactions before clinical trials.

K-Paths: Reasoning over Graph Paths for Drug Repurposing and Drug Interaction Prediction