AI discovers new superbug-killing antibiotic

This AI-driven process efficiently sifted through thousands of potential chemicals to identify a select few for laboratory testing. The development comes as antibiotic resistance becomes a pressing global health concern. The team's focus centred on Acinetobacter baumannii, a highly problematic bacterium classified as a "critical" superbug by the World Health Organization due to its resistance to multiple antibiotics. This superbug poses a substantial threat, particularly in healthcare settings.

Researchers trained the AI in the search for a new antibiotic through manual tests on known drug structures against Acinetobacter baumannii, providing the AI with insights into the chemical features necessary to combat it. The AI then evaluated the effectiveness of 6,680 compounds with unknown efficacy. The results, published in Nature Chemical Biology, revealed that the AI could produce a shortlist in just an hour and a half.

Laboratory experiments identified nine potential antibiotics, including the highly potent abaucin, capable of treating infected wounds in mice and eradicating A. baumannii samples from patients.
Notably, abaucin exhibits precision, selectively targeting A. baumannii while sparing other bacteria, which could reduce the development of drug resistance and side effects. This eliminates the laborious need to manually screen potential compounds.

Lead author of the paper Dr Jonathan Stokes of McMaster University said of the discovery: "AI enhances the rate, and in a perfect world decreases the cost, with which we can discover these new classes of antibiotic that we desperately need".

While the researchers initially tested AI-aided antibiotic discovery in E. coli in 2020, they have since focused on addressing more challenging pathogens, with plans to investigate Staphylococcus aureus and Pseudomonas aeruginosa. Prof Dame Sally Davies, former Chief Medical Officer for England and government envoy on anti-microbial resistance said of the discovery: "We're onto a winner."

This discovery highlights AI's potential in the healthcare field, especially with the expedition of antibiotic development. Further refinement in the laboratory and subsequent clinical trials will be undertaken before AI antibiotics become available, with dates penned for 2030.