Researchers have utilized artificial intelligence (AI) to accelerate the discovery of new antibiotics. In their quest to find a new antibiotic, the researchers trained the AI by testing thousands of drugs on a bacterium called Acinetobacter baumannii. They manually determined which drugs could slow down or kill the bacterium and fed this information into the AI. The AI was then unleashed on a list of 6,680 compounds with unknown effectiveness. Within an hour and a half, the AI produced a shortlist of potential antibiotics.
Laboratory experiments were conducted on 240 compounds from the shortlist, leading to the discovery of nine potential antibiotics. One particularly potent antibiotic, abaucin, showed promise in treating infected wounds in mice and effectively killing A. baumannii samples from patients. Unlike many antibiotics, abaucin seemed to target A. baumannii specifically, reducing the likelihood of drug resistance and minimizing side effects.
Although this breakthrough is significant, further work is required. The researchers aim to perfect the drug in the laboratory and then proceed with clinical trials. The availability of AI-aided antibiotics for prescription is not expected until around 2030.
The researchers believe that AI can revolutionize antibiotic discovery by screening millions of potential compounds, a task that would be impractical to perform manually. By accelerating and expanding the search for novel antibiotics, AI has the potential to combat problematic pathogens such as A. baumannii and could be applied to other bacteria like Staphylococcus aureus and Pseudomonas aeruginosa.
Experts in the field, such as Professor James Collins from the Massachusetts Institute of Technology, view this work as evidence that AI can significantly contribute to the fight against antibiotic-resistant bacteria. The use of AI in antibiotic discovery holds promise for addressing the urgent need for new classes of antibiotics.
