Network-Based Drug Repositioning

Drug repositioning -- identifying new therapeutic indications for approved drugs -- offers a faster and cheaper path totreatment than de novo drug discovery, reducing development timelines from 12-15 years to 3-5 years and costs from$2.6 billion to $300-500 million per indication. Network-based approaches that model drugs, targets, diseases, and genesas interconnected nodes in biological knowledge graphs can systematically predict repositioning candidates byidentifying drugs proximal to disease modules in the interactome. However, existing network methods achieve hit rates ofonly 5-15% in experimental validation, limited by incomplete knowledge graphs, simplistic proximity metrics, and inabilityto capture polypharmacology (multi-target drug effects). This study developed and benchmarked six drug repositioningapproaches -- literature-based similarity (DrugBank indication overlap), transcriptomic signature matching (CMap/L1000),network proximity (shortest path in PPI), graph embedding (node2vec on drug-target-disease graph), heterogeneousGNN link prediction (RGCN), and a proposed Multi-Evidence Network Repositioning Engine (MENRE) combiningheterogeneous knowledge graph embedding (drug-target-disease-pathway-side effect, 248,000 nodes, 2.4 millionedges), transcriptomic perturbation signatures, molecular structural similarity, and clinical evidence from EHR-deriveddrug-disease co-occurrence via a multi-relational graph transformer -- for predicting novel drug-disease associationsacross 1,248 approved drugs and 862 diseases. MENRE achieved the highest prediction accuracy: AUROC 0.942 onheld-out known indications (vs network proximity 0.768, RGCN 0.884, CMap 0.812; p < 0.001), with precision at top-50predictions of 38.4% (vs RGCN 24.6%, CMap 18.2%). Experimental validation of MENRE's top 20 novel predictions bycell-based phenotypic assays confirmed 7 of 20 (35%) as biologically active in the predicted disease context. Clinicalvalidation using retrospective EHR analysis (n = 4.2 million patients) identified 3 repositioning candidates with significantprotective associations (HR 0.62-0.78; p < 0.01). These results establish multi-evidence network repositioning as themost effective computational approach for systematic drug repurposing.