Harvard scientists, led by Dr. Marinka Zitnik, have introduced TxGNN, an AI-based tool designed to identify drug repurposing opportunities for diseases where effective treatments are lacking. This innovative model adopts a unique approach by simultaneously analyzing multiple diseases, enabling insights from well-documented conditions to be applied to rare diseases characterized by sparse data. This approach addresses a critical gap, as traditional drug-repurposing models typically concentrate on individual diseases, which limits their broader applicability.
TxGNN's Methodology
TxGNN operates within a "zero-shot" framework, allowing it to predict therapeutic candidates for diseases with limited or no known treatments. This is particularly relevant given that there are approximately 7,000 rare diseases affecting populations worldwide. The model focuses on U.S. FDA-approved drugs and aims to target disease-related networks, identifying both direct and indirect therapeutic effects. TxGNN comprises two primary modules: one that suggests drug indications and contraindications, and another that explains the knowledge graphs linking drugs to diseases.
Performance and Validation
In a recent study published in Nature Medicine, TxGNN ranked approximately 8,000 drugs, assessing their treatment potential for over 17,000 diseases. The model's top predictions frequently aligned with existing off-label drug prescriptions, demonstrating its ability to outperform other drug-repurposing AI models. The "Explainer" module provides insights into how medical knowledge influences the AI’s recommendations, which is particularly useful for researchers evaluating drug efficacy and patient safety. In real-world trials, the Harvard team observed that this feature raised confidence levels among clinicians and researchers assessing TxGNN’s predictions.
Clinical Trials and Accessibility
With $48.3 million in funding from the Advanced Research Projects Agency for Health, further developments are underway. In partnership with BioPhy, TxGNN’s results are being tested in clinical trials. Every Cure, a nonprofit organization dedicated to advancing drug repurposing efforts, is playing a key role in deploying these tools for high-value drug-disease matches. TxGNN is currently available to the public, enabling researchers and clinicians to input disease and drug queries for immediate AI-driven predictions. The tool aims to democratize access to AI resources in drug development, addressing the critical needs of underserved patient groups with rare diseases. Harvard’s collaboration with patient organizations, including the Chan Zuckerberg Initiative, highlights the social impact of TxGNN by supporting foundations dedicated to advancing treatments for rare conditions.
