A research team led by Cornell has demonstrated how quantum computing and artificial intelligence can be used to design new peptides capable of capturing microplastics that pose serious risks to ecosystems and human health.
Peptides - chains of amino acids that can be sequenced for specific functions - can destroy microplastics by binding to their surfaces and facilitating chemical reactions that break them down. However, there are no known peptide binders for many plastics, and creating them is challenging due to the lack of data on plastic adsorption.
A study published Dec. 18 in Science Advances details a method for integrating generative AI models trained to predict peptide properties, with quantum computing techniques that optimize these designs for specific plastics. This approach allows for rapid exploration of a wide range of potential amino acid sequences, enabling the discovery of new types of peptides more quickly than the iterative process of testing each combination individually.
"Scientists don't really have the datasets to work with peptides targeting microplastics in the same way they do for some medical applications, and that's where quantum comes in," said lead author Fengqi You, the Roxanne E. and Michael J. Zak Professor in Energy Systems Engineering. "Fundamentally, we see this as an AI problem, but we use quantum as a booster. It can evaluate many possibilities of amino acid sequences simultaneously and that's different from what we can do with classical computing."
The hybrid quantum-classical framework developed in the study resulted in the discovery of peptides with high binding affinities for polyethylene terephthalate, also known as PET plastic, that maintained an environmentally friendly water solubility. The results were then validated using molecular dynamics simulations, confirming the potential for practical applications in water treatment systems, microplastic biosensors and engineered microbes designed to degrade plastics.
"This work comes at a time when AI's transformative potential has been recognized globally for applications like protein structure prediction and other biological and medical applications," said You, who is also co-director of the Cornell University AI for Science Institute. "This new hybrid quantum-classical computing framework extends AI's impact into environmental science."
You added that the framework is designed to work within the current constraints of quantum computing, leveraging simulations on classical hardware while remaining adaptable to advancements in quantum technology as it matures.
The study was a collaboration with Carol K. Hall '67, professor of chemical and biomolecular engineering at North Carolina State University. Co-first authors include Raul Conchello Vendrell, a visiting graduate student in You's lab from ETH Zurich; postdoctoral researcher Akshay Ajagekar, Ph.D. '24; and Michael T. Bergman, a doctoral student from North Carolina State University.
The research team conducted a complementary investigation, applying quantum computing, AI and biophysical modeling to design peptides for other common plastics, such as polyethylene and polypropylene. That work, published Jan. 21 in PNAS Nexus, features Jeet Dhoriyani, M.S. '24, as the first author. The study optimized amino acid sequences for strong binding and explored physicochemical diversity, enabling peptides to perform across different environmental conditions.
You said upcoming efforts will focus on synthesizing and testing the newly discovered peptides in laboratory and field settings. Additionally, the researchers plan to refine the computational frameworks to target other sustainability challenges.
The research was supported by the National Science Foundation and the Eric and Wendy Schmidt AI in Science Postdoctoral Fellowship, a program of Schmidt Sciences.
Syl Kacapyr is associate director of marketing and communications for Cornell Engineering.
