Artificial intelligence discovers new antibiotic to treat “superbugs”

The discovery, made by McMaster University and the Massachusetts Institute of Technology (MIT), could be key in the search for other preparations to treat these pathogens known as “superbugs”, which have developed immunity to all known treatments and would become a common cause of death more than cancer in just a few decades.

Specifically, researchers have discovered a new drug to treat the bacterium Acinetobacter baumannii, identified by the World Health Organization as one of the most dangerous in the world because it is extremely difficult to eradicate. Usually contracted in hospital, this pathogen can cause pneumonia, meningitis or seriously infect wounds, putting life at risk.

The researchers published the results of their work in the journal Nature Chemical Biology, where they explain that they used an artificial intelligence algorithm to predict new antibacterial molecular structures, which led them to identify a new compound, which they named abaucin.

Unlike traditional, expensive and very limited-scope methods, new algorithmic approaches have access to millions of molecules with antibacterial properties and their combinations, greatly speeding up the research process.

“This work validates the benefits of machine learning in the search for new antibiotics,” said Jonathan Stokes, lead author of the paper and professor of biomedicine and biochemistry at McMaster University, who also pointed out that thanks to artificial intelligence, “It can now explore rapidly vast regions of chemical space, significantly increasing the chances of discovering new antibacterial molecules”.

Artificial intelligence training

To find the new antibiotic, the researchers trained the artificial intelligence by providing it with thousands of molecules effective against the bacterium Acinetobacter baumannii. Because they have a precise chemical structure, the algorithm learned the features that work best against the pathogen and provided a list of more than 6,500 previously unknown compounds. The scientific team tested the 240 most promising ones in the laboratory and found nine potential antibiotics, including abaucin.

Laboratory experiments have shown its productivity in treating infected skin wounds in mice and in killing samples of Acinetobacter baumannii in humans. The next step now will be to perfect the drug in the laboratory, before capitalizing on clinical trials with patients. The research team does not believe that the first antibiotics discovered with the help of artificial intelligence can be commercialized before 2030. Source: rtve.es