AI Designs Novel Antibiotics with Potential to Combat Resistance
Generative AI now designs life-saving antibiotics, not just art and text. Penn researchers introduce AMP-Diffusion, a generative AI tool, in a new Cell Biomaterials paper.This tool creates tens of thousands of new antimicrobial peptides (AMPs) – short strings of amino acids, the building blocks of proteins – with bacteria-killing potential. in animal models, the most potent AMPs performed as well as FDA-approved drugs, without detectable adverse effects.
Previous Penn breakthroughs demonstrated AI’s ability to identify promising antibiotic candidates from large datasets. This study joins a growing number showing AI can invent candidates from scratch.
“Nature’s dataset is finite; with AI,we can design antibiotics evolution never tried,” says César de la Fuente,Presidential Associate Professor in Bioengineering (BE) and in Chemical and Biomolecular Engineering in the University of Pennsylvania School of Engineering and Applied Science (Penn Engineering),in Psychiatry and Microbiology in the Perelman School of Medicine and in Chemistry in the School of Arts & Sciences,and the paper’s senior co-author.
“We’re leveraging the same AI algorithms that generate images, but augmenting them to design potent new molecules,” adds Pranam Chatterjee, Assistant Professor in BE and in computer and Details Science within Penn Engineering, and the paper’s other senior co-author, who began work on the project while at Duke University.
Two labs, one goal
De la Fuente’s lab successfully used AI to search for molecules with antimicrobial properties in unexpected places, including proteins from woolly mammoths, animal venom, and ancient microbes called archaea. “Unfortunately,antibiotic resistance keeps increasing faster than we can discover new antibiotic candidates,” says de la Fuente.
This led his lab to collaborate with Chatterjee’s, which designs peptides using AI to treat diseases where conventional drug progress fails. “It seemed like a natural fit,” says Chatterjee. “Our lab knows how to design new molecules using AI, and the de la Fuente Lab knows how to identify strong antibiotic candidates using AI.”
Tuning out the noise
Generative AI models, like ChatGPT, predict the next word or element in a sequence. “Diffusion” models, however, start from random “noise” and iteratively refine it into a coherent output – the principle behind tools like DALL·E and Stable Diffusion.
AMP-Diffusion operates similarly, refining amino acid sequences instead of pixels.