Scientists have used artificial intelligence to identify a new antibiotic that might be useful to fight a deadly drug-resistant bacteria commonly found in hospitals and medical offices.
Researchers report they used an AI algorithm to predict molecules that would neutralize the drug-resistant bacteria Acinetobacter baumannii. Researchers discovered a potential antibiotic, named abaucin, “can effectively suppress” the growth of the stubborn bacteria on the skin of mice, according to a study this week in the journal Nature Chemical Biology.
While the preliminary results on the potential new drug would need to be validated in larger studies, researchers believe the process used to winnow thousands of potential drugs to identify one that may work is an approach that can work in drug discovery.
“There’s a lot of trepidation around AI and I genuinely understand it,” said Jonathan Stokes, lead author on the paper and an assistant professor of biomedicine and biochemistry at at McMaster University in Ontario, Canada. “When I think about AI in general, I think of these models as things that are just going to help us do the thing we’re going to do better.”
Which superbug did researchers target?
Stokes teamed up with researchers from the Broad Institute of MIT and Harvard to screen for potential antibiotics to use on Acinetobacter baumannii, a superbug that can cause infections in the blood, urinary tract and lungs. This bacteria usually invades hospitals and healthcare settings, infecting vulnerable patients on breathing machines, in intensive care units and undergoing operations.
This type of bacteria, resistant to the potent antibiotic carbapenem, infected 8,500 in hospitals and killed 700 in 2017, according to the Centers for Disease Control and Prevention.
How did AI pinpoint a new antibiotic?
The researchers evaluated 7,684 drugs and the active ingredients of drugs to find out which ones would be effective against the bacteria, which was grown in the lab.
Stokes said the lab team developed AI models to predict which ones would have the highest likelihood of antimicrobial activity, narrowing the field to 240 drugs or active ingredients. Researchers then narrowed the field again through testing before discovering a molecule RS102895, renamed abaucin, that appeared to be potent against the superbug.
What’s next for abaucin?
Stokes said researchers are working to optimize the chemical structure of the potential antibiotic. Plans eventually call for doing follow-up research in larger animals, and potentially humans, if abaucin proves to be effective.
“It’s important to remember right when we’re trying to develop a drug, it doesn’t just have to kill the bacterium,” Stokes said. “It also has to be well tolerated in humans and it has to get to the infection site and stay at the infection site long enough to elicit an effect.”
Researchers said they can screen a much larger volume of potential drugs by using machine-learning techniques. The study said while existing high-throughput screening can evaluate a few million drugs or chemical ingredients at once, algorithms developed from machine learning can assess “hundreds of millions to billions” of drug molecules.
