Anyone suffering from inflammatory bowel disease (IBD) may tell you that antibiotics are a bit like a two-edged sword. Sure, they control the bursts of illness by destroying harmful bacteria but in this process, they also eliminate the beneficial microbes crucial for a healthy gut. The resulting imbalance can sometimes worsen the symptoms over time – it’s almost like using a sledgehammer to knock a fly.
In a remarkable discovery, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) alongside McMaster University have unveiled a new compound, enterololin. Unlike typical antibiotics, enterololin focuses on specific bacteria related to flare-ups of Crohn’s disease, leaving the rest of the microbiome unharmed. This pinpoint approach could herald a new age in gut health treatment.
But what stands out in this discovery is the method used to achieve it. Instead of depending solely on slow-paced lab experiments, the team utilised a generative AI model called DiffDock. Innovated at CSAIL, DiffDock can predict how small molecules bind to proteins- a complex task typically requiring years to decipher. Impressively, DiffDock enabled researchers to understand how enterololin works in a matter of months!
What this means is that AI is no longer just about identifying new molecules. As MIT Professor Regina Barzilay, co-developer of DiffDock, puts it, “What we’re showing here is that AI can also provide mechanistic explanations.” This could be instrumental in accelerating the development of early-stage compounds. In conventional scenarios, studies of mechanism-of-action can take upto two years and cost millions. But DiffDock slashed this timeline to about six months, massively reducing both time and expense.
Enterololin is still in its early development stage, but things are promisingly in motion. Stokes (senior author of the study and assistant professor at McMaster) has initiated a spinout company, Stoked Bio, to hone the compound for human use. In addition, the team is also exploring its potential against other antibiotic-resistant pathogens, including Klebsiella pneumoniae. If progress continues at this pace, clinical trials may spring to life within the next few years.
This breakthrough shines the spotlight beyond just Crohn’s disease. The ability to concentrate only on harmful bacteria without disrupting the microbiome has been a long-standing goal in medicine, but the discovery and validation of such drugs has been a complex task. AI tools like DiffDock could change this, paving the way for a new era of precision medicine. “What excites me isn’t just this compound, but the idea that we can now approach mechanism-of-action studies more efficiently,” says Stokes. “With the right mix of AI, scientific insight, and lab work, we’re opening new doors in drug discovery.”
The significance of this work doesn’t escape Yves Brun, a professor at the University of Montreal who wasn’t part of the study. He emphasizes, “One of the greatest challenges to our health is the rise of antimicrobial-resistant bacteria. AI is becoming a powerful tool in combating these threats. This study proves an elegant use of AI to reveal how a new antibiotic works — a crucial step in its development.”
Indeed, this exciting study redefines what’s possible in our fight against stubborn diseases. By opening their data to the public and releasing the DiffDock-L code on GitHub, the research team encourages a spirit of open science that will nurture collaborative breakthroughs. We look forward to tracking the progress as this remarkable journey unfolds.
For more details on this fascinating breakthrough, visit the original news release.
This website uses cookies.