ProTDyn: A Foundation Protein Language Model for Thermodynamics and Dynamics Generation
Abstract
Molecular dynamics (MD) simulation has long been the principal computational tool for exploring protein conformational landscapes, but its application is limited by high computational cost. We present ProTDyn, a foundation protein language model that unifies conformational ensemble generation and multi-timescale dynamics modeling within a single framework. Unlike prior approaches that treat these tasks separately, ProTDyn allows flexible i.i.d ensemble sampling and dynamic trajectory simulation. Across diverse protein systems, ProTDyn yields thermodynamically consistent ensembles, faithfully reproduces dynamical properties over multiple timescales, and generalizes to proteins beyond its training data—offering a scalable and efficient alternative to conventional MD simulations.
Cite
Text
Liu et al. "ProTDyn: A Foundation Protein Language Model for Thermodynamics and Dynamics Generation." International Conference on Learning Representations, 2026.Markdown
[Liu et al. "ProTDyn: A Foundation Protein Language Model for Thermodynamics and Dynamics Generation." International Conference on Learning Representations, 2026.](https://mlanthology.org/iclr/2026/liu2026iclr-protdyn/)BibTeX
@inproceedings{liu2026iclr-protdyn,
title = {{ProTDyn: A Foundation Protein Language Model for Thermodynamics and Dynamics Generation}},
author = {Liu, Yikai and Zheng, Haoyang and Mao, Lining and Wang, Yanbin and Chen, Ming and Lin, Guang},
booktitle = {International Conference on Learning Representations},
year = {2026},
url = {https://mlanthology.org/iclr/2026/liu2026iclr-protdyn/}
}