Researchers from the University of Cambridge have designed and used an artificial-intelligence (AI)-based approach to advance drug design and accelerate the search for Parkinson’s disease (PD) treatments.
Published in the journal Nature Chemical, researchers used AI to identify compounds that block the clumping or aggregation of alpha-synuclein, the key protein that characterises PD.
Affecting more than six million people worldwide, PD is a progressive neurological condition that slowly deteriorates parts of the brain.
As well as motor symptoms, PD can also affect the gastrointestinal system, nervous system, sleeping patterns, mood and cognition and can contribute to a reduced quality of life and significant disability.
Researchers developed and used a machine learning technique to screen a chemical library that contained millions of entries to identify small molecules that bind to the amyloid aggregates and block their proliferation.
The process of screening for drug candidates among large chemical libraries can be time-consuming, expensive and often unsuccessful.
The research team successfully sped up the initial screening process ten-fold and identified five highly potent compounds to be further investigated while reducing the cost by a thousand-fold, meaning that potential treatments for PD could reach patients much faster.
“Using the knowledge… gained from the initial screening…, we were able to train the model to identify the specific regions on these small molecules responsible for binding,… [to] re-screen and find more potent molecules,” explained Michele Vendruscolo, co-director of the Centre for Misfolding Diseases, University of Cambridge.
In doing so, researchers developed compounds to target pockets on the surfaces of the aggregates responsible for the exponential proliferation of the aggregates themselves, which were hundreds of times more potent and cheaper to develop.
“Machine learning is… speeding up the whole process of identifying the most promising candidates,” said Vendruscolo. “This means we can start work on multiple drug discovery programmes instead of just one…, massively reducing… both time and cost.”
