Using AI to discover a new super-strong bacteria-killing antibiotic

Scientists have harnessed artificial intelligence (AI) to discover a new antibiotic that can kill deadly superbugs.

AI has helped narrow thousands of potential chemicals down to a handful that can be tested in the lab.

The result is a powerful experimental antibiotic called Abaucin, which needs further testing before use.

Researchers in Canada and the United States say AI has the potential to greatly accelerate drug discovery.

This is the latest example of how artificial intelligence tools can be a revolutionary force in science and medicine.

Antibiotics kill bacteria. But new drugs have been scarce for decades, and bacteria have evolved resistance to the bacteria we have, making it difficult to treat.

Researchers focused on one of the most problematic bacterial species. Acinetobacter baumanniiIt can infect wounds and cause pneumonia.

You may not have heard of it, but it’s one of three superbugs that the World Health Organization has identified as a “major” threat.

It is a problem in hospitals and nursing homes because it can often defy multiple antibiotics and survive on surfaces and medical devices.

Dr. Jonathan Stokes of McMaster University described the bug as “public enemy number one” because cases were “very common” that the bug was “resistant to nearly all antibiotics.” there is

To find new antibiotics, researchers first had to train an AI. They took thousands of drugs with known exact chemical structures and manually tested them. Acinetobacter baumannii See which ones can slow you down or stop you.

This information is fed into AI, which can now learn the chemical signature of drugs that might attack the bacteria in question.

The AI ​​was then published based on a list of 6,680 compounds of unknown potency. Results published in Nature Chemical Biology showed that it took the AI ​​an hour and a half to create the shortlist.

Researchers tested 240 people in the lab and found nine potential antibiotics. One of them was the incredibly powerful antibiotic Abaucin.

Laboratory experiments have shown that it can heal and kill infected wounds in mice. A. Baumani A sample taken from a patient.

But Dr. Stokes told me: “This is where the work begins.”

The next step is to perfect the drug in the lab and conduct clinical trials. He expects it could take as long as 2030 before the first AI antibiotics can be prescribed.

Curiously, this experimental antibiotic has no effect on other bacterial species and only acts on bacteria. A. Baumani.

Many antibiotics kill bacteria indiscriminately. Researchers believe that Abaucin’s precision may make drug resistance less likely to emerge, leading to fewer side effects.

In principle, AI can screen tens of millions of potential compounds, but doing this manually is impractical.

“AI will increase that rate and reduce costs in a perfect world. This will allow us to discover new classes of antibiotics that we desperately need,” Dr. Stokes told me. .

Researchers tested the principles of AI-powered antibiotic discovery in E. coli in 2020, and are now using that knowledge to focus on a critical problem.they plan to consider Staphylococcus aureus and Pseudomonas aeruginosa Next.

“This finding further supports the premise that AI can greatly accelerate and scale the search for novel antibiotics,” said Professor James Collins of the Massachusetts Institute of Technology.

He added, “We are excited that this study shows that AI can be used to combat problematic pathogens such as A. baumannii.”