AI-designed peptides break through the outer defenses of MRSA bacteria

A newly-designed AI-powered tool is effective in developing treatments to attack antibiotic-resistant bacteria by breaking down their outer defenses, according to new research from Houston Methodist.

The study, published in Nature Communications and led by Eleftherios Mylonakis, M.D., Ph.D., chair, Houston Methodist Charles W. Duncan Jr. Department of Medicine, details how researchers used the tool to identify antimicrobial peptides - small proteins that are part of the body's natural immune system – that effectively targeted bacteria like methicillin-resistant Staphylococcus aureus (MRSA) in lab tests.

"Antibiotic-resistant bacteria represent a major global health threat, with an estimated 2.8 million infections and more than 35,000 deaths annually in the U.S. Addressing this challenge is critical," Mylonakis said.

Antimicrobial peptides offer a promising approach to target difficult-to-treat bacteria while reducing the likelihood of resistance. However, designing these molecules with precision has traditionally been complex and time intensive. To overcome this, we developed an AI-powered platform that enables the identification and design of peptides most effective against MRSA and other pathogens."

Eleftherios Mylonakis, Houston Methodist

First authors Fadi Shehadeh and Biswajit Mishra, together with their collaborators, designed CAMPER (Constraint-driven AMP Engineering with Ranking), an AI-based platform that integrates machine learning with biologically informed features. CAMPER evaluates and ranks libraries of candidate peptides based on their physical and chemical properties and predicted performance. Using this approach, the team identified a promising candidate, WP-CAMPER1, which showed potent activity against MRSA at low concentrations and revealed its potential for treating antibiotic-resistant infections.

"Ultimately, our study reports and validates the CAMPER methodology, demonstrating its ability to generate peptides that shows effectiveness against difficult-to-treat and persistent infections. It represents an important step toward a scalable platform for developing therapeutics targeting complex pathogens," Mylonakis said.

Other collaborators on this study include LewisOscar Felix, Charilaos Dellis, Narchonai Ganesan and Liyang Zhang from Houston Methodist; Raquel Ferrer-Espada, Andrew Martens, Youlian Goulev and Johan Paulsson from Harvard Medical School; Anindya Basu from Rajiv Gandhi Technological University; Michael B. Sherman from The University of Texas Medical Branch at Galveston; Mandar Naik from Brown University and Paul Sotiriadis from National Technical University of Athens and Archimedes-Athena Research Center.