Powerful New Antibiotic Identified by Artificial Intelligence

MIT researchers recently employed a machine-learning algorithm to identify a new powerful antibiotic. This new drug has proven quite effective in treating many of the world’s most troublesome disease-causing bacteria, and this includes several strains that are known to be antibiotic-resistant. This has been observed both in lab tests and using mice.

Amazing Power of Artificial Intelligence

Yet again, we see artificial intelligence playing a huge role in modern medicine. This deep learning computer model is capable of sifting through over a hundred million various chemical compounds in just a few days. And it has been specifically designed to screen out antibiotics showing the potential of killing bacteria with mechanisms that differ from drugs that currently exist in the market.

“We wanted to develop a platform that would allow us to harness the power of artificial intelligence to usher in a new age of antibiotic drug discovery,” said James Collins, who is a Termeer Professor of Medical Engineering & Science at the MIT’s Institute for Medical Engineering and Science and Depart of Biological Engineering. “Our approach revealed this amazing molecule which is arguably one of the more powerful antibiotics that has been discovered.”

More Antibiotic Compounds Show Promise

Powerful New Antibiotic Identified by Artificial IntelligenceIn this amazing new study, scientists have also screened out a number of other potential antibiotic compounds which will be tested further. They also feel that this excited new model will also design new drugs in the future because of how quickly it has learned which chemical structures are most capable of killing bacteria.

“The machine learning model can explore, in silico, large chemical spaces that can be prohibitively expensive for traditional experimental approaches,” claims Regina Barzilay, who is the Delta Electronics Professor of Electrical Engineering & Computer Science at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL).

Barzilay and Collins are the faculty co-leaders for the MIT Abdul Latif Jameel Clinic for Machine Learning in Health (J-Clinic), are also the authors of the study’s final report, which was recently posted in Cell.

A Brand Antibiotic Pipeline

Amazingly, medicine has not seen very many new antibiotic drugs over the last few decades. And most of those were simply spin-offs of drugs that already exist. One big reason for this is that existing methods of screening out new antibiotics have been very expensive. Not only do existing methods come with a huge cost, they are also very time-consuming, and are also limited to a very narrow spectrum of drug classes.

“We’re facing a growing crisis around antibiotic resistance, and this situation is being generated by both an increasing number of pathogens becoming resistant to existing antibiotics, and an anemic pipeline in the biotech and pharmaceutical industries for new antibiotics,” Collins points out.

In order to identify completely novel compounds, this research team has created an artificial intelligence that analyzes the molecular structure of compounds and then correlates them with traits that have the potential to kill off bacteria.

This method of employing predictive computer models to screen drugs has actually been around for a while.  However, the older models weren’t accurate enough to create new drug discoveries. In these previous methods, molecules were treated as vectors that determined the absence or presence of certain groups of chemicals.

These new neural networks can identify these group representations automatically, and they can map molecules straight into continuous vectors that will accurately predict their properties. Watching this antibiotic identified by artificial intelligence is very exciting.