Scientists have discovered a new antibiotic with the help of artificial intelligence, which can kill a deadly species of superbugs, the BBC reports.
Artificial intelligence helped researchers select a few substances out of thousands of possibilities to test in the lab.
The result was a powerful experimental antibiotic called abaucin. However, it will need further testing before it can be used by humans.
Researchers from Canada and the US say that artificial intelligence has the ability to significantly speed up the process of discovering new drugs.
This is the latest example of how AI tools can be a revolutionary force in science and medicine.
Antibiotics were revolutionary drugs when they were discovered because they kill bacteria. But new classes of antibiotics haven’t appeared in decades, and bacteria are becoming harder to treat as they develop resistance.
It is estimated that more than one million people die annually from antibiotic-resistant infections.
The researchers focused on one of the most problematic species of bacteria – Acinetobacter baumannii, which can infect wounds and cause pneumonia. It is one of three bacteria that the World Health Organization considers a “critical” threat.
They are often resistant to several antibiotics and are a problem in hospitals and nursing homes because they can survive on medical surfaces and equipment.
Dr. Jonathan Stokes of McMaster University says this superbug is “public enemy number one” because there are many cases where it is “resistant to almost every antibiotic.”
To find a new antibiotic, researchers first had to “train” artificial intelligence. They took thousands of drugs whose precise chemical structure was known and manually tested Acinetobacter baumannii to see which ones could slow or kill it.
This information was fed to artificial intelligence tools to learn the chemical characteristics of drugs that might attack the problematic bacteria.
She was then made to study a list of 6,680 compounds whose efficacy was unknown. The results – published in Nature Chemical Biology – showed that it took the AI an hour and a half to produce a short list of substances.
The researchers tested 240 compounds in the laboratory and found nine potential antibiotics. One of them, which turned out to be very strong, was called abaucin.
Laboratory experiments showed that it could treat infected wounds in mice and was able to kill Acinetobacter baumannii samples from patients.
But it won’t hit the market anytime soon.
“This is when the work starts,” Stokes said.
The next step is to refine the drug in the laboratory and then conduct clinical trials. The first antibiotics created with the help of artificial intelligence would be able to be prescribed to patients only in 2030, he added.
Curiously, this experimental antibiotic had no effect on other species of bacteria and only works on Acinetobacter baumannii.
Many antibiotics kill bacteria indiscriminately. The researchers believe that abaucin’s precision will make it harder for drug resistance to emerge and could lead to fewer side effects.
In principle, AI could analyze tens of millions of potential compounds – something that would be impossible to do manually.
“AI is increasing the rate and, in a perfect world, decreasing the cost at which we can discover these new classes of antibiotics that we desperately need,” Dr Stokes said.
Researchers tested the principles of AI-assisted antibiotic discovery on E. coli in 2020, but have now used that knowledge to focus on superbugs. They plan to continue looking for antibiotics to fight Staphylococcus aureus and Pseudomonas aeruginosa.
“This discovery further supports the premise that artificial intelligence can significantly accelerate and expand the search for new antibiotics,” said Professor James Collins, from the Massachusetts Institute of Technology.
“I’m excited that this work shows that we can use AI to help fight problematic pathogens,” he added.
Editor : M.B.