MIT Researchers Uses Generative AI to Target Deadly Superbugs with New Molecules

The MIT team applied generative AI models to explore vast chemical spaces and design entirely new molecules with novel mechanisms of action.

Researchers at the Massachusetts Institute of Technology (MIT) have developed new antibiotics using generative artificial intelligence, marking a breakthrough in the fight against antibiotic-resistant bacteria.

The study, published in Cell on August 14, 2025, details how AI-designed compounds were able to kill methicillin-resistant Staphylococcus aureus (MRSA) and Neisseria gonorrhoeae, the bacteria that causes gonorrhea.

Antibiotic resistance is a growing global challenge, with MRSA and N. gonorrhoeae increasingly difficult to treat with existing drugs. Traditional antibiotic discovery has been slow and costly, limiting progress. The World Health Organization has warned that antimicrobial resistance could cause up to 10 million deaths annually by 2050.

The MIT team, led by James Collins, Termeer Professor of Medical Engineering and Science at MIT, along with lead authors Aarti Krishnan, Melis Anahtar, and Jacqueline Valeri, applied generative AI models to explore vast chemical spaces and design entirely new molecules with novel mechanisms of action.

For N. gonorrhoeae, the researchers began with a library of 45 million chemical fragments, using machine-learning models to identify a candidate foundation.

Generative AI tools then produced 7 million variations, ultimately yielding a compound named NG1. In lab tests, NG1 targeted LptA, a protein essential for bacterial membrane synthesis, making it more difficult for the pathogen to develop resistance.

In the case of MRSA, the team used a different approach, allowing AI to generate 29 million potential compounds. After screening, 22 candidates were synthesized, with six showing strong antibacterial activity. One compound, DN1, demonstrated significant effectiveness by disrupting bacterial membranes in a way distinct from existing antibiotics. DN1 successfully cleared MRSA infections in mouse models.

“This approach allows us to explore chemical spaces that are impractical to search with conventional methods,” said Collins. “We’re discovering new classes of antibiotics that bacteria haven’t seen before.”

The researchers emphasized that NG1 and DN1 remain early-stage discoveries. Further optimization and clinical trials are required before the compounds can be considered for human use.

The team, working with Phare Bio, plans to refine the compounds and extend their AI platform to other pathogens such as Mycobacterium tuberculosis and Pseudomonas aeruginosa.

“We’re just beginning to tap into AI’s potential,” said Krishnan. “This platform could accelerate the discovery of antibiotics for a wide range of resistant bacteria.”

The study was supported by the National Institutes of Health, the Broad Institute, and other partners. Researchers believe the work could pave the way for a new generation of antibiotics at a time when resistance is outpacing traditional discovery methods.

Stay tuned for more such updates on Digital Health News