In a big step forward for drug discovery, US and Canadian scientists have used artificial intelligence (AI) to discover a new drug capable of killing one of the world's deadliest superbugs.
Researchers from McMaster University and Massachusetts Institute of Technology (MIT) published the study on Thursday in the science journal Nature Chemical Biology.
The team used machine learning to help identify new antibiotics for the bacteria acinetobacter baumannii - a superbug that survives most treatments and causes most hospital-acquired infections.
MacMaster University scientist Gary Liu trained the AI model.
"All that this model was going to be doing is telling us essentially if new molecules will have antibacterial properties or not," he said.
After Liu trained the algorithm, it took 90 minutes for the AI to screen about 7,500 new compounds for their bacteria-killing effectiveness.
In the lab, they tested 240 of the most promising chemicals - resulting in nine new potential antibiotics, including one they named abaucin.
Abaucin was then tested on A. baumannii infections in mice, and the team found it could prevent infection.
The purpose of the study was to find new structural classes of antibiotics - those chemically different enough to known drugs, Liu said.
"We're able to just increase the efficiency of the drug discovery pipeline," he added.
The World Health Organisation says A. baumannii is one of the top three "critical" pathogen threats to human health.
The superbug can cause infections in the blood, urinary tract, lungs, and wounds including from surgery - so it's a growing risk for hospital patients and aged care facilities.
A. baumannii also lives for long periods on surfaces, equipment, and can be spread via contaminated hands.
A 2022 study in Te Ara Tika o te Hauora Hapori / New Zealand Medical Journal says A. baumanii "poses a threat to healthcare" in Aotearoa, due to limited treatments and the tendency for it to cause hospital outbreaks.
The study's lead, Jonathan Stokes from McMaster University, says the work validates machine learning's application to medical sciences.
"Using AI, we can rapidly explore vast regions of chemical space, significantly increasing the chances of discovering fundamentally new antibacterial molecules," he added.
Stokes also said superbugs will continue to evolve and adjust, so AI's potential uses here are exciting.
"AI methods afford us the opportunity to vastly increase the rate at which we discover new antibiotics, and we can do it at a reduced cost. This is an important avenue of exploration for new antibiotic drugs," he added.
A. baumannii can "colonise" or live in someone's body without causing symptoms, according to the US' Centers for Disease Control and Prevention.
The WHO adds infections such as pneumonia, tuberculosis, blood poisoning (sepsis), and gonorrhoea are becoming hard to treat - sometimes impossible - as antibiotics become less effective over time.
It says when antibiotics are unnecessarily prescribed for humans and animals, it's "accelerating the progress" of antibiotic resistance.