About This Project
Project Background
MyTxGNN is a drug repurposing research platform based on Harvard University's TxGNN model published in Nature Medicine. It predicts potential new indications for drugs registered with Malaysia's National Pharmaceutical Regulatory Agency (NPRA), integrating evidence from ClinicalTrials.gov, PubMed, and other authoritative sources.
Team & Attribution
| Item | Information |
|---|---|
| Project Maintained By | Yao.Care |
| Model Foundation | Harvard TxGNN (Zitnik Lab) |
| Data Source | NPRA Malaysia, data.gov.my |
| Last Updated | March 2026 |
Academic Reference
This project’s AI prediction model is based on:
Huang, K., et al. (2023). A foundation model for clinician-centered drug repurposing. Nature Medicine. DOI: 10.1038/s41591-023-02233-x
What is Drug Repurposing?
Drug Repurposing (also known as drug repositioning) is the process of finding new therapeutic uses for existing drugs. Compared to developing new drugs from scratch, which takes 10-15 years and costs $1-2 billion, drug repurposing can be completed in 3-5 years at a fraction of the cost, with lower failure risk since safety data already exists.
| Comparison | New Drug Development | Drug Repurposing |
|---|---|---|
| Development Time | 10-15 years | 3-5 years |
| Development Cost | $1-2 billion | $100-300 million |
| Safety Data | Must establish new | Already available |
| Failure Risk | Very high (>90%) | Lower |
What is TxGNN?
TxGNN is a deep learning model developed by Harvard Medical School's Zitnik Lab, published in Nature Medicine. It predicts new drug-disease relationships and is the first foundation model designed specifically for clinician-centered drug repurposing.
"TxGNN integrates a knowledge graph of 17,080 biomedical entities, using graph neural networks to learn complex relationships between nodes, enabling prediction of drug efficacy for rare diseases." — Huang et al., Nature Medicine (2023)
Technical Features
- Knowledge Graph: Integrates 17,080 nodes including drugs, diseases, genes, and proteins
- Graph Neural Network: Learns complex relationships between nodes
- Prediction Capability: Predicts which diseases a drug may be effective for
Data Sources
This platform integrates multiple public data sources including AI predictions, clinical trials, academic literature, drug information, Malaysia regulatory data, and drug interaction databases.
| Data Type | Source | Description |
|---|---|---|
| AI Prediction | TxGNN | Harvard knowledge graph prediction model |
| Clinical Trials | ClinicalTrials.gov | Global clinical trial registry |
| Literature | PubMed | Biomedical literature database |
| Drug Information | DrugBank | Drug and target database |
| Malaysia Regulatory | NPRA | National Pharmaceutical Regulatory Agency |
| Open Data | data.gov.my | Malaysia Government Open Data |
Project Scale
| Item | Count |
|---|---|
| NPRA Registered Products | 27,938 |
| Drugs with DrugBank Mapping | 508 |
| KG Predictions | 41,560 |
| DL High-Score Predictions | 176,021 |
| Unique Diseases | 17,041 |
How to Cite
If using this platform’s data, please cite:
TxGNN Model
@article{huang2023txgnn,
title={A foundation model for clinician-centered drug repurposing},
author={Huang, Kexin and others},
journal={Nature Medicine},
year={2023},
doi={10.1038/s41591-023-02233-x}
}
Contact & Feedback
For questions or suggestions, please contact us through:
- GitHub Issues: https://github.com/yao-care/MyTxGNN/issues
- Project Homepage: https://mytxgnn.yao.care/
This report is for academic research purposes only and does not constitute medical advice. Drug use should follow physician guidance. Do not self-adjust medications. Any drug repurposing decisions require complete clinical validation and regulatory review.
Last reviewed: 2026-03-03 | Reviewer: MyTxGNN Research Team