mlsanityData to 14 Oct 2026

Performance of universal machine-learned potentials with explicit long-range interactions in biomolecular simulations

arXiv 2508.10841 ↗First posted 14 Aug 2025physics.chem-ph cond-mat.soft cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Universal machine-learned potentials promise transferable accuracy across compositional and vibrational degrees of freedom, yet their application to biomolecular simulations remains underexplored. This work systematically evaluates equivariant message-passing architectures trained on the SPICE-v2 dataset with and without explicit long-range dispersion and electrostatics. We assess the impact of model size, training data composition, and electrostatic treatment across in- and out-of-distribution benchmark datasets, as well as molecular simulations of bulk liquid water, aqueous NaCl solutions, and biomolecules, including alanine tripeptide, the mini-protein Trp-cage, and Crambin. While larger models improve accuracy on benchmark datasets, this trend does not consistently extend to properties obtained from simulations. Predicted properties also depend on the composition of the training dataset. Long-range electrostatics show no systematic impact across systems. However, for Trp-cage, their inclusion yields increased conformational variability. Our results suggest that imbalanced datasets and immature evaluation practices currently challenge the applicability of universal machine-learned potentials to biomolecular simulations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 15 of its 84 references that are in the map. First posted on arXiv 14 Aug 2025.

Builds on

15 of 84 references.

References · 15

  1. 2021E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentialscited by 708
  2. 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 474
  3. 2022Learning Local Equivariant Representations for Large-Scale Atomistic Dynamicscited by 256
  4. 2021SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effectscited by 104
  5. 2023On the Expressive Power of Geometric Graph Neural Networkscited by 90
  6. 2022SPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentialscited by 82
  7. 2024Latent Ewald summation for machine learning of long-range interactionscited by 26
  8. 2025High-performance training and inference for deep equivariant interatomic potentialscited by 17
  9. 2023Uncertainty-biased molecular dynamics for learning uniformly accurate interatomic potentialscited by 16
  10. 2021Gaussian Moments as Physically Inspired Molecular Descriptors for Accurate and Scalable Machine Learning Potentialscited by 15
  11. 2023CHGNet: Pretrained universal neural network potential for charge-informed atomistic modelingcited by 13
  12. 2024Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passingcited by 13
3 more
  1. 2021Fast and Sample-Efficient Interatomic Neural Network Potentials for Molecules and Materials Based on Gaussian Momentscited by 9
  2. 2024Nutmeg and SPICE: Models and Data for Biomolecular Machine Learningcited by 7
  3. 2025Fast, Modular, and Differentiable Framework for Machine Learning-Enhanced Molecular Simulationscited by 2

Built on by

Not cited yet.