Roger-Luc Chayer, from CNN (Photo: Matt Clarke/McMaster University)
Thanks to artificial intelligence, researchers say they’ve discovered an entirely new type of antibiotic that can take on a particularly dangerous bacteria—one that’s resistant to almost all known treatments.
When they tested the drug on mice infected with this “superbug,” the results were promising: the bacteria’s growth was significantly slowed. This suggests the method could help create other tailor-made antibiotics to fight equally tough pathogens.
Even more impressively, the new compound — named abaucin — proved effective against 41 different strains of Acinetobacter baumannii, a hospital-acquired bacterium notorious for its resistance. Of course, before it can be used in humans, it will need further refinement and clinical trials.
But there’s something else that makes abaucin stand out: it has a rare superpower. Unlike broad-spectrum antibiotics that wipe out both harmful and helpful bacteria, abaucin only targets its specific enemy, leaving the rest of the microbiome intact. Think of it more as a precision sniper than a bulldozer.
According to the researchers, this kind of surgical precision could help slow down the rise of antibiotic resistance — one of the biggest challenges in modern medicine.
The study was published in the journal Nature Chemical Biology. So far, abaucin has only been tested on animals, but the method used to discover it could change the way we develop antibiotics going forward.
It all started with a particularly aggressive bacterium: Acinetobacter baumannii. Commonly found in hospitals, it loves to linger on doorknobs, metal surfaces, and medical equipment. It’s also a master of genetic recycling, snatching up DNA from other bacteria to arm itself against drugs. In labs, scientists call it a “professional pathogen.”
It can cause serious skin, blood, and lung infections, and according to the U.S. Centers for Disease Control and Prevention, it’s one of the top-priority bacteria in need of new treatments. A recent study found that among hospitalized patients infected with strains resistant even to powerful carbapenem antibiotics, one in four died within a month of diagnosis.
To find a way to stop it, researchers at McMaster University in Canada, working with the Broad Institute of MIT and Harvard, screened more than 7,500 molecules on lab-grown samples of the bacterium. Of those, 480 were found to inhibit its growth.
They then fed all that data into an AI model trained to recognize the most promising compounds. Once trained, the model scanned over 6,000 new molecules in just a few hours — a task that would have taken months by traditional methods.
Next, the researchers tested 240 of these compounds in the lab, narrowing it down to 9. One stood out: a molecule called RS102895, originally developed as a potential treatment for diabetes. It works in a totally new way — by blocking an internal mechanism that the bacteria use to move key components to their surface.
And most importantly, it only works on Acinetobacter baumannii. Not on other bacteria. Unlike traditional antibiotics, it doesn’t trigger a survival-of-the-fittest arms race among all the bacteria in the body — a process that often leads to widespread resistance.
In short, artificial intelligence may have just given the fight against superbugs a major boost. And abaucin, this futuristic little molecule, could be the first of many more to come.
ADVERTISING
