UPenn Researchers Use AI to Discover Antibiotic Alternatives
Using AI, researchers were able to find patterns in peptides containing antimicrobial properties, inspired by our ancient ancestors!
Created on August 1|Last edited on August 1
Comment
Bioengineers at the University of Pennsylvania have leveraged artificial intelligence (AI) to resurrect molecules from long-extinct human relatives, such as Neanderthals and Denisovans. This revolutionary process, published on July 28 in Cell Host & Microbe, is aimed at addressing the increasingly urgent problem of antibiotic-resistant bacteria.
Stagnation
Traditional antibiotic development has stagnated, with most drugs on the market being over three decades old. With antibiotic-resistant bacteria becoming more common, there is a dire need for new treatments.
Leveraging the Past
The team took inspiration from the concept of bringing back molecules from the past. The idea initially sparked from a notion connected to the film "Jurassic Park," but the scientists turned their attention to the more attainable goal of resurrecting molecules rather than dinosaurs.
Predicting Peptides
Using AI, the researchers were able to identify specific sites on human proteins where peptides, short protein subunits with antimicrobial (capable of killing microorganisms) properties, are formed. By applying this AI algorithm to available protein sequences from modern humans, Neanderthals, and Denisovans, they could predict peptides that might be effective against bacteria.
This approach dramatically accelerates the process of finding and testing drug candidates. Traditional methods take three to six years to discover a new antibiotic, whereas AI can accomplish this in mere weeks.
Interesting Results
Dozens of peptides were tested against bacteria in the laboratory, with six potent peptides selected for further study. While they successfully halted the growth of the common bacterium Acinetobacter baumannii in mice, none managed to kill the bacteria, and the doses required were extremely high.
Despite these challenges, the research team believes that tweaking the molecules and altering the algorithm could lead to more successful outcomes. Cesar de la Fuente, a co-author of the study, asserts that this concept and framework open up a new path for drug discovery.
Criticisms
While some experts are optimistic about this new angle in antibiotic development, others, like Nathanael Gray, a chemical biologist at Stanford University, remain cautious. He argues that until the algorithm can predict clinically relevant peptides with greater success, this method may not significantly impact drug discovery.
Euan Ashley, a genomics and precision-health expert at Stanford University, sees potential in exploring the archaic human genome for antibiotic development and acknowledges the team's work as an interesting and potentially useful approach.
An AI-driven Medical Future
This novel utilization of AI to "de-extinct" ancient molecules represents an intriguing direction in the search for new antibiotics. While the initial results may not have yielded groundbreaking molecules, the project's methodology has created an entirely new perspective on how we might tackle modern health issues by looking to our distant past. It also underlines the significant role AI can play in accelerating the discovery of life-saving medications, a notion that will likely continue to shape scientific research in the coming years.
The article: https://www.nature.com/articles/d41586-023-02403-0
Add a comment
Tags: ML News
Iterate on AI agents and models faster. Try Weights & Biases today.