News

Cambridge is using Machine Learning to discover new drugs

Machine Learning
Researchers have designed a machine learning algorithm for drug discovery which has been shown to be twice as efficient as the industry standard

Researchers at the University of Cambridge have built a Machine Learning algorithm which could accelerate the process of discovering and developing new pharmaceutical drugs.

The researchers, led by Dr Alpha Lee, used their algorithm to discover four molecules that can help treat Alzheimer’s and Schizophrenia.

The ability to fish out four active molecules from six million is like finding a needle in a haystack
Alpha Lee, Lead Researcher

The algorithm works by predicting whether a molecule is capable of activating a particular protein that is thought to be helpful in treating the diseases, the prediction model works by creating a statistical model by searching for patterns that are known to be present in such molecules.

Machine Learning will revolutionalize drug discovery

The method discovered by the researchers will not only help make it easy and cost-effective to discover new drugs. However, even in cases where its algorithm identifies a molecule that doesn’t match with the pattern it is looking for, the failed step will also help it add more information to its database for future use.

The researchers used 222 active molecules and computationally screened six million molecules to identify about 100 molecules that seemed most viable. From the shortlisted 100, the researchers identified 4 that had the ability to activate the protein that the researchers were looking at.

Essentially, the process helped researchers to discard almost 6 million molecules which would have taken months through conventional methods. The process according to Dr. Lee is twice as efficient as industry methods.

Read more here at the University of Cambridge.

The original research is published here.

vote
Article Rating
ALSO READ  Convervationists are using Intel AI to stop animal poaching
Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments