In a groundbreaking study, researchers from MIT and McMaster University have leveraged the power of generative AI to dramatically speed up the discovery and understanding of a new antibiotic. The drug, named enterololin, shows promise for treating inflammatory bowel diseases (IBD), including Crohn’s disease, by selectively targeting harmful gut bacteria while preserving beneficial microbes.
Typically, discovering how an antibiotic interacts with bacterial proteins is a lengthy process, taking up to 18-24 months. However, using the AI model called DiffDock, the research team was able to predict, in a mere 100 seconds, how enterololin attacks gut bacteria. The AI suggested that enterololin specifically targets the LolCDE protein complex in harmful strains of E. coli, which was subsequently confirmed through laboratory experiments.
This approach not only reduces the drug discovery timeline from years to six months but also significantly cuts the associated costs. Traditionally costing up to $2 million, the AI-driven method managed to achieve this milestone for just $60,000.
The successful prediction and validation of enterololin’s mechanism represent a breakthrough in drug development, demonstrating how AI can move beyond identifying candidate molecules to offering mechanistic insights much earlier in the process. This has closed significant bottlenecks in the field and is paving the way for enterololin to potentially enter human trials within the next three years, under the license of Stoked Bio.
This case highlights the transformative potential of AI in revolutionizing the efficiency of antibiotic development, offering a promising future for treating complex diseases with precision.

Collections

Researchers from MIT and McMaster University used generative AI model DiffDock to accelerate antibiotic discovery from 18-24 months to 100 seconds. The AI predicted how enterololin targets harmful gut bacteria linked to inflammatory bowel diseases while preserving beneficial microbes. This approach reduced development time from years to six months and costs from $2 million to $60,000, with the drug potentially entering human trials within three years.

AI Accelerates Discovery of New Antibiotic Targeting Gut Bacteria Linked to IBD