Researchers at MIT have developed novel antibiotics using generative AI to combat drug-resistant bacteria, including Neisseria gonorrhoeae and Staphylococcus aureus (MRSA). These compounds, structurally distinct from existing antibiotics, disrupt bacterial cell membranes through novel mechanisms. The study, published in Cell, demonstrates AI's potential to explore vast chemical spaces for antibiotic discovery.
The team employed two AI approaches: one involved designing molecules based on a specific antimicrobial fragment, while the other allowed the AI to freely generate molecules. This led to the identification of compounds like NG1 and DN1, which effectively killed drug-resistant bacteria in laboratory and mouse models. NG1 targets the LptA protein in Gram-negative bacteria, while DN1 affects broader aspects of bacterial cell membranes.
The research highlights the potential of AI in drug discovery, enabling the design of compounds with novel mechanisms of action. This approach could lead to antibiotics effective against a wide range of bacterial pathogens. The team is collaborating with Phare Bio to further develop these compounds for preclinical testing.
This advancement marks a significant step in addressing the global challenge of antibiotic resistance, which causes nearly 5 million deaths annually. By harnessing AI, researchers can rapidly and cost-effectively identify promising antibiotic candidates, potentially revolutionizing the development of treatments for resistant infections.
