The Role of Explainable AI in Knowledge Graph-Based Drug Repurposing

Drug repurposing, the process of identifying new therapeutic uses for existing drugs, has emerged as a cost-effective and time-efficient alternative to traditional drug development. Recently, Knowledge Graphs (KGs) have become a valuable resource for drug repurposing due to their ability to integrate diverse biomedical data into a structured format. However, the complexity of KG-based models and recommendations often limits clinical adoption due to reduced interpretability. This study addresses this challenge by integrating Explainable AI (XAI) techniques into the drug repurposing pipeline, enhancing transparency in prediction models. We explore SHAP, LIME, GNNExplainer, PGExplainer, and attention mechanisms to provide insights into how predictions are made within KG-based systems.