Â鶹ÒùÔº

A study published in reveals a powerful new use for artificial intelligence: designing small, drug-like molecules that can stick to and break down harmful proteins in the body—even when scientists don't know what those proteins look like.

The breakthrough could lead to new treatments for diseases that have long resisted traditional drug development, including certain cancers, brain disorders, and viral infections.

The study was published by a multi-institutional team of researchers from McMaster University, Duke University, and Cornell University. The AI tool, called PepMLM, is based on an algorithm originally built to understand human language and used in chatbots, but was trained to understand the "language" of proteins.

In 2024, the Nobel Prize in Chemistry was awarded to researchers at Google DeepMind for developing AlphaFold, an AI system that predicts the 3D structure of proteins—a major advance in drug discovery. But many disease-related proteins, including those involved in cancer and neurodegeneration, don't have stable structures.

That's where PepMLM takes a different approach—instead of relying on structure, the tool uses only the protein's sequence to design peptide drugs. This makes it possible to target a much broader range of disease proteins, including those that were previously considered "undruggable."

"Most drug design tools rely on knowing the 3D structure of a protein, but many of the most important disease targets don't have stable structures," said Pranam Chatterjee, senior author of the study who led the work at Duke and is now a faculty member at the University of Pennsylvania. "PepMLM changes the game by designing peptide binders using only the protein's amino acid sequence," said Chatterjee.

In lab tests, the team showed that PepMLM could design peptides—short chains of amino acids—that stick to disease-related proteins and, in some cases, help destroy them. These included proteins involved in cancer, reproductive disorders, Huntington's disease, and even live viral infections.

"This is one of the first tools that can design these kinds of molecules directly from the protein's sequence," said Chatterjee. "It opens the door to faster, more effective ways to develop new treatments."

The study included major contributions from McMaster University, where Christina Peng, a Ph.D. student in the Truant Lab, led the Huntington's disease experiments.

"It's exciting to see how these AI-designed peptides can actually work inside cells to break down toxic proteins," said Peng. "This could be a powerful new approach for diseases like Huntington's, where traditional drugs haven't been effective."

Other parts of the study were carried out at Cornell, where Matthew DeLisa and Hector Aguilar's labs constructed and tested the peptides on viral proteins, and at Duke, where Chatterjee's team developed the AI model and ran early validation experiments. The study also included contributions from Ray Truant at McMaster.

"This work shows we can now bind any protein to any other protein," said Truant, a professor in the Department of Biochemistry & Biomedical Sciences. "We can degrade harmful proteins, stabilize beneficial ones, or control how proteins are modified—depending on the therapeutic goal."

The team is already working on next-generation AI algorithms, like PepTune and MOG-DFM, to improve how these peptides behave in the body—making them more stable, more targeted, and easier to deliver.

"Our ultimate goal is a general-purpose, programmable peptide therapeutic platform—one that starts with a sequence and ends with a real-world drug," said Chatterjee.