MD-LLM-2: A Transferable Language Model of Molecular Dynamics with Physical Conditioning and Explicit Path Probabilities
Abstract
Molecular dynamics (MD) simulations model the motion of proteins between conformational states, but crossing free energy barriers and sampling multiple thermodynamic conditions remain computationally costly. We recently introduced MD-LLM-1, which showed that language models can generate conformational trajectories at lower cost, but required a separate model for each protein and did not explicitly model thermodynamics or kinetics. Here we introduce MD-LLM-2, a transferable language model that generates trajectories conditioned on sequence, temperature and molecular history. Because each step defines a normalized distribution over structural tokens, the model assigns an explicit likelihood to every generated or simulated trajectory. Trained across mdCATH domains, MD-LLM-2 reproduces the median radius of gyration of MD within 1 Angstrom in 10 of 12 unseen domains, predicts reduced Trp-cage compaction at higher temperature, and generates L99A T4 lysozyme trajectories that reach published transition-region backbone structures. In alanine dipeptide simulations, conditioning on a torsional scaling parameter λ predicts basin populations at unseen couplings with a mean absolute error of 0.09 versus 0.35 for MBAR, and first-passage probabilities with an error of 0.08 versus 0.16 for a reversible Markov state model. Frame-isoentropic sampling further limits the accumulation of errors during autoregressive trajectory generation. Together, these results extend MD-LLM from protein-specific trajectory generation to transferable, physically conditioned modelling with explicit path likelihoods, and show that the framework can predict equilibrium populations for force-field parameters not seen during training, with an initial indication that first-passage kinetics can also be captured.
Related
- Bolek: A Multimodal Language Model for Molecular Reasoning
- Structural insights into an aptainhibitor of the SARS-CoV-2 E protein using molecular docking, molecular dynamics, and biophysical analyses
- Allosteric Communication in the Tubulin Dimer Revealed through Molecular Dynamics and Network Analysis
- Edge-Aware Graph Attention Networks for Interpreting Biophysical Mechanisms from Molecular Dynamics Simulations
- MDigest: A comprehensive toolkit for detection of allosteric communication from molecular dynamics simulations of biochemical systems
BibTeX
@misc{murtada2026mdllm2,
title = {{MD-LLM-2: A Transferable Language Model of Molecular Dynamics with Physical Conditioning and Explicit Path Probabilities}},
author = {Mhd Hussein Murtada and Z. Faidon Brotzakis and Michele Vendruscolo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.10879},
eprint = {2610.10879},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.10879},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - MD-LLM-2: A Transferable Language Model of Molecular Dynamics with Physical Conditioning and Explicit Path Probabilities
AU - Mhd Hussein Murtada
AU - Z. Faidon Brotzakis
AU - Michele Vendruscolo
PY - 2026
DA - 2026/10/07
PB - arXiv
DO - 10.48550/arxiv.2610.10879
UR - https://arxiv.org/abs/2610.10879
AB - Molecular dynamics (MD) simulations model the motion of proteins between conformational states, but crossing free energy barriers and sampling multiple thermodynamic conditions remain computationally costly. We recently introduced MD-LLM-1, which showed that language models can generate conformational trajectories at lower cost, but required a separate model for each protein and did not explicitly model thermodynamics or kinetics. Here we introduce MD-LLM-2, a transferable language model that generates trajectories conditioned on sequence, temperature and molecular history. Because each step defines a normalized distribution over structural tokens, the model assigns an explicit likelihood to every generated or simulated trajectory. Trained across mdCATH domains, MD-LLM-2 reproduces the median radius of gyration of MD within 1 Angstrom in 10 of 12 unseen domains, predicts reduced Trp-cage compaction at higher temperature, and generates L99A T4 lysozyme trajectories that reach published transition-region backbone structures. In alanine dipeptide simulations, conditioning on a torsional scaling parameter λ predicts basin populations at unseen couplings with a mean absolute error of 0.09 versus 0.35 for MBAR, and first-passage probabilities with an error of 0.08 versus 0.16 for a reversible Markov state model. Frame-isoentropic sampling further limits the accumulation of errors during autoregressive trajectory generation. Together, these results extend MD-LLM from protein-specific trajectory generation to transferable, physically conditioned modelling with explicit path likelihoods, and show that the framework can predict equilibrium populations for force-field parameters not seen during training, with an initial indication that first-passage kinetics can also be captured.
ER -