Drug Repositioning with GraphSAGE and Clustering Constraints Based on Drug and Disease Networks

The study proposed a novel method called DRGCC, which utilized GraphSAGE and clustering constraints to explore the therapeutic properties of drugs for drug repositioning. By extracting drug structure and disease symptom features, and constructing drug–drug interaction and disease similarity networks, the method aimed to predict new drug-disease associations. The approach demonstrated reliable predictive performance, outperforming six other methods in benchmark comparisons across two datasets. Additionally, case studies focused on predicting drugs potentially effective against COVID-19, with some identified candidates already under clinical investigation and showing binding to COVID-19-related protein receptors.