Researchers at the University of Pennsylvania, United States, managed to develop an artificial intelligence system capable of designing completely new antibiotics and who have managed to eliminate drug -resistant bacteria in mice.
The work of researchers, whose results have been published today in Cell Biomaterials magazine, led to the creation of a generative artificial intelligence tool (which baptized as “AMP-Diffusion”) capable of studying gigantic collections of synthetic antimicrobial molecules and proposing candidates to become new antibiotics.
The investigation was directed by the Spanish César de la Fuente, a professor at the University of Pennsylvania – where the machine’s biology group and Pranam Chatterjee, a professor of bioengineering and computer sciences, and the scientists verified that in mouse models with skin infection the most effective compounds designed by artificial intelligence showed a performance comparable to antibiotics already approved by the agency already approved by the agency already approved by the agency of the medicine and did not present adverse effects.
The tool they developed uses the “grammar” learned from proteins to propose biologically plausible sequences, while optimizing antimicrobial activity and other relevant properties for drug development, explained the University of Pennsylvania in a press release.
A new border in drug development
“The nature set is finite, but with artificial intelligence we can design antibiotics that evolution never tried,” said De la Source, and valued that this project demonstrated that machines can invent new antibiotics from scratch, opening a new border in drug development.
The researchers created approximately 50,000 candidate molecules with this system and selected them through AI -based filters.
Of the 46 main candidates who were synthesized and undergoing experimental tests, two proved to be able to reduce drug -resistant skin infections in mice, with an effectiveness comparable to that of current clinical antibiotics, without detecting signs of toxicity in studies.
“This is a proof of principle that generative AI can create new effective antibiotics,” said Chatterjee, and has celebrated that the results “point to a way to quickly develop therapies that address the growing threat of antibiotic resistance.”
You may be interested: Are you really allergic to penicillin? Specialist explains why you are not very likely and how to make sure
Reduce the discovery of antibiotics from years to days
The study adds to a small but growing number of demonstrations that artificial intelligence can invent antibiotic candidates from scratch.
For years, the Laboratory of De la Fuente successfully took the AI to look for molecules with antimicrobial properties in unlikely places, from the proteins of the woolly mammoths to those of the animal venom and the old microbes (called arches).
“Unfortunately, antibiotic resistance increases at a greater pace than we can discover new antibiotic candidates,” the Spanish researcher observed.
For researchers, the current study constitutes a proof of principle: generative AI can go beyond the extraction of what evolution has already created, until it reaches the design of new antibiotics.
“Ultimately, our goal is to shorten the antibiotic discovery time from years to days,” said De la Fuente.
With EFE information
Follow us on Google News to always keep you informed
