Transferable Graph Neural Network Surrogates for Molecular Dynamics Across Crystal Symmetries

Abstract

We present a transferable graph neural network (GNN) surrogate framework for molecular dynamics (MD) that directly predicts atomic displacements and propagates atomistic configurations without explicit force evaluation or numerical time integration. The central objective of this work is to establish whether a common GNN formulation can represent atomic dynamics across materials with fundamentally different crystal symmetries and coordination environments. The same network architecture, feature representation, graph construction, and training protocol are applied without symmetry-specific modification to face-centered cubic (FCC) aluminum, body-centered cubic (BCC) iron, and hexagonal close-packed (HCP) magnesium. Across these distinct elemental and crystallographic systems, the framework achieves position-prediction errors on the order of 10−4 Å 2 and supports stable autoregressive propagation to nanosecond time scales. The predicted trajectories preserve thermodynamic stability, characteristic coordination-shell structure in the radial distribution functions, and temperature-dependent mean-squared-displacement behavior. In particular, the framework captures the closely spaced coordination shells of BCC iron and the anisotropic coordination environment of HCP magnesium without introducing lattice-specific representations. These results demonstrate that direct GNN-based atomic propagation can be formulated as a transferable framework across different elements, lattice symmetries, and coordination geometries, providing a pathway toward generalizable surrogate models for accelerated molecular dynamics.

SupraTITO: Transferable Generative Molecular Dynamics for Supramolecular Systems
Weilong Chen, Nuno Costa, Julija Zavadlav
arXiv1 Oct 2026PreprintCC BY 4.0Full textPhysical chemistry Biological & medical physicsarXiv
Out-of-Distribution Inverse Design of Elastic Networks with Differentiable Graph Neural Network Molecular Dynamics
Sergey A. Shteingolts, Salman N. Salman, Ron Levie +1
arXiv6 Sep 2026PreprintCC BY-NC-ND 4.0Full textPhysical chemistryarXiv
Elucidating the MultiPathway Therapeutic Potential of Trichosanthes kirilowii Bioactives in Intestinal Obstruction via Network Pharmacology and Molecular Dynamics
Tianying Wu, Xiaofen Lin, Wencong Lin +2
ChemistryOpen5 Oct 2026Peer-reviewedCC BY 4.0Full textGeneral chemistryCrossref
Comparative Anti-Inflammatory Evaluation of Lamiaceae Essential Oils and Mechanistic Insights into Mentha haplocalyx Essential Oil via Network Pharmacology and Molecular Dynamics
Jiabao Cui, Junmei Pu, Hui Yang +2
Journal of Oleo Science7 Oct 2026Peer-reviewedCC BY-SAFull textGeneral chemistryJ-STAGE
Network toxicology, molecular docking and molecular dynamics simulations for bisphenol A neurotoxicity in depression pathogenesis
Jinlian Zhang, Yuanhao Ma, Miguo Lu +7
PLOS One5 Oct 2026Peer-reviewedCC BY 4.0Full textEnvironmental chemistry Theoretical & computational chemistryPLOS

Cite

BibTeX RIS

BibTeX

@misc{immanuel2026transferable,
  title = {{Transferable Graph Neural Network Surrogates for Molecular Dynamics Across Crystal Symmetries}},
  author = {Judah Immanuel and Avik Mahata and Aniruddha Maiti},
  howpublished = {ChemRxiv},
  year = {2026},
  month = sep,
  doi = {10.26434/chemrxiv.15009166/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009166/v1},
  note = {CC BY},
}

RIS

TY  - UNPB
TI  - Transferable Graph Neural Network Surrogates for Molecular Dynamics Across Crystal Symmetries
AU  - Judah Immanuel
AU  - Avik Mahata
AU  - Aniruddha Maiti
PY  - 2026
DA  - 2026/09/21
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15009166/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009166/v1
AB  - We present a transferable graph neural network (GNN) surrogate framework for molecular dynamics (MD) that directly predicts atomic displacements and propagates atomistic configurations without explicit force evaluation or numerical time integration. The central objective of this work is to establish whether a common GNN formulation can represent atomic dynamics across materials with fundamentally different crystal symmetries and coordination environments. The same network architecture, feature representation, graph construction, and training protocol are applied without symmetry-specific modification to face-centered cubic (FCC) aluminum, body-centered cubic (BCC) iron, and hexagonal close-packed (HCP) magnesium. Across these distinct elemental and crystallographic systems, the framework achieves position-prediction errors on the order of 10−4 Å 2 and supports stable autoregressive propagation to nanosecond time scales. The predicted trajectories preserve thermodynamic stability, characteristic coordination-shell structure in the radial distribution functions, and temperature-dependent mean-squared-displacement behavior. In particular, the framework captures the closely spaced coordination shells of BCC iron and the anisotropic coordination environment of HCP magnesium without introducing lattice-specific representations. These results demonstrate that direct GNN-based atomic propagation can be formulated as a transferable framework across different elements, lattice symmetries, and coordination geometries, providing a pathway toward generalizable surrogate models for accelerated molecular dynamics.
ER  -

Abstract

Related

Cite