Researchers have identified peptides that can help remove microplastics from the environment by combining biophysical modeling, molecular dynamics, quantum computing, and reinforcement learning. The ultimate goal of the work is peptide-based technologies that can find, capture, and destroy microscopically tiny plastic particles.
Microplastics, plastic particles smaller than 5 mm, are ubiquitous pollutants, found everywhere from human breastmilk to Antarctic snow. Fengqi You and colleagues used a range of tools to identify peptides able to capture and hold microplastics, which could be used to remove the tiny particles from various environments. The authors used biophysical modeling to predict peptide-plastic interactions at atomic resolution, then validated the results with molecular dynamics simulations. The process was optimized with the addition of quantum annealing and reinforcement learning—specifically a method known as proximal policy optimization. Using these tools, the authors identified a set of plastic-binding peptides with high affinities for polyethylene and polypropylene. According to the authors, the method, when paired with experimental approaches, could be used to develop peptide-based tools for detecting, capturing, and degrading microplastic pollution.
