Answers are grounded in this paper and cite their sources.
Biomolecular thermodynamics and spectroscopy depend on relative conformer energies, local curvatures, and collective dipole fluctuations on the potential-energy surface. Conventional molecular mechanics force fields enable large-scale simulations, but their fixed functional forms can misrepresent infrared intensities, mode character, and environment-dependent vibrational response. Here we assess general-purpose machine-learned force fields by using finite-temperature infrared spectra across small molecules, gas-phase peptides, and monomeric, oligomeric, and solvated protein assemblies. To enable this analysis, we introduce QVib, a dataset of 293 molecules and 1366 conformers, together with peptide amide-band benchmarks and p53 oligomerization-domain models, to evaluate vibrational transferability from DFT references to experimental spectra. Across these systems, machine-learned force fields improve over molecular mechanics in reproducing DFT-level forces, vibrational frequencies, densities of states, mode eigenvectors, conformational energetics, and experimental infrared spectra. Among models with explicit long-range electrostatics, SO3LR provides the best accuracy–cost balance of the models tested here. These results show that machine-learned force-field dynamics can recover collective, environment-dependent vibrational landscapes at near-DFT fidelity, enabling spectroscopically validated biomolecular simulations at force-field-like cost.
No takes yet. Share an insight, caveat, or question.
Dou et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: