What is it about?
Proteins must fold into specific 3D shapes to function, but finding their fleeting intermediate steps within massive computer simulations is difficult. We developed a new Artificial Intelligence tool called Conditional Transition Clustering (CTC). Instead of grouping shapes by geometric similarity—which often introduces bias—CTC learns how shapes transition over time, automatically discovering hidden, short-lived states in a protein's folding journey.
Featured Image
Photo by Shubham Dhage on Unsplash
Why is it important?
We introduce a "dynamics-centric" paradigm that changes how complex molecular simulations are analyzed. Previously, scientists had to make pre-defined, biased assumptions about protein states to build kinetic models. Our AI framework removes these rigid constraints, allowing the molecule's true kinetic behavior to emerge naturally. This objective approach is crucial for mapping complex folding pathways and understanding misfolding-related diseases.
Perspectives
As a computational biophysicist, I was often frustrated by the amount of human bias traditionally required to just interpret simulation data. I wanted a framework where the data could speak for itself. I hope this "dynamics-first" philosophy not only transforms how we study protein folding, but also empowers researchers to decipher other complex molecular systems with unprecedented objectivity.
Xuyang Liu
Nankai University
Read the Original
This page is a summary of: Unveiling hidden intermediate states in protein folding with AI-based conditional transition clustering, Proceedings of the National Academy of Sciences, March 2026, Proceedings of the National Academy of Sciences,
DOI: 10.1073/pnas.2531221123.
You can read the full text:
Contributors
The following have contributed to this page
