AI designed antibiotics show promise

MIT researchers use generative AI to design novel antibiotics that fight drug resistant gonorrhoeae and MRSA.

AI designed antibiotics target drug resistant bacteria

MIT researchers used generative AI to design antibiotics that target two hard-to-treat infections: drug resistant Neisseria gonorrhoeae and MRSA. They built and screened millions of compounds and identified two promising candidates with novel modes of action that seem to disrupt the bacteria’s outer membranes. NG1 interacts with LptA, a protein involved in membrane synthesis, and showed strong activity in lab tests and a mouse model of drug resistant gonorrhea.

In a second approach, the team conducted unconstrained design, generating more than 29 million molecules for Staphylococcus aureus. After filtering and synthesis checks, 22 molecules were tested, and six showed potent activity against multi-drug resistant S. aureus in the lab, with the top candidate DN1 clearing a MRSA skin infection in mice. The candidates appear to target membranes broadly rather than a single protein, and the work underscores AI’s potential to explore vast chemical space. The study was published in Cell as part of MIT’s Antibiotics AI Project, with Phare Bio helping advance preclinical work.

AI lets scientists imagine molecules that do not exist today, expanding the toolkit beyond current libraries. But moving from dish to patient raises questions about safety, regulatory review, and how quickly new drugs can reach patients. The promise is clear, yet the path is long. The mix of science and policy means funding, safety checks, and clear guidelines will shape whether this becomes a real tool in the fight against antibiotic resistance.

Highlights

The road from lab to clinic will test both science and policy.