➕ AI Will Help Treat Rare Diseases
There are more than 7,000 rare and difficult-to-diagnose diseases affecting about 300 million people globally. Many of them are genetic. Only 5-7% of these diseases have cures — the rest remain primarily untreated. AI may help fix this situation.
💊 Researchers from Harvard University have developed an AI model called TxGNN that identifies drug candidates for rare diseases from existing medicines.
They trained the model on a dataset that includes DNA information, gene activity, and more than 1.2 million patient records. The algorithms identify common gene mechanisms of well-known and treatable diseases and extrapolate this data to poorly understood or untreatable ones.
Results
⚛️ TxGNN found drug candidates for more than 17,000 diseases, many without existing treatments. The new tool performed nearly 50% more efficiently than other AI models.
This is precisely where we see the promise of AI in reducing the global disease burden, in finding new uses for existing drugs, which is also a faster and more cost-effective way to develop therapies than designing new drugs from scratch.
Marinka Zhitnik, lead researcher and professor of biomedicine at Harvard
The use
➡️ AI algorithms will speed up the search for cures: the team is already collaborating with rare disease foundations to help identify treatments.
➡️ The model predicts side effects and contraindications — the data is usually only obtainable in the early stages of clinical trials.
➡️ Search for more effective and safer drugs and also for common diseases.
