HEALTH

AI helps researchers identify potential Parkinson’s treatments

A machine-learning model sifts though millions of known compounds to find chemicals that could be developed into new drugs
The use of the AI system makes research ten times faster and a thousand times cheaper
The use of the AI system makes research ten times faster and a thousand times cheaper

Cambridge scientists have hailed the discovery of five “highly potent” chemical compounds that could be used to create revolutionary Parkinson’s treatments.

The researchers used machine-learning to sift through a library of millions of chemicals to identify those that could block the build-up and clumping of alpha-synuclein, a protein that plays a key role in the development of Parkinson’s disease.

Identifying and testing compounds that could be effective for creating drugs can often take years, researchers said, but the new AI-driven system has “massively accelerated” the process, making it ten times faster and a thousand times cheaper.

A study published in the journal Nature Chemical Biology noted that: “This [research] is a particularly pressing need given the lack of disease-modifying therapies currently available to Parkinson’s disease