AutoSolvate 2.0: Automated Preparation of Chemically Diverse Explicit-Solvent Systems for Multiscale Molecular Simulation
Abstract
Advanced quantum mechanics/molecular mechanics (QM/MM), machine-learned interatomic potentials (MLIPs), and hybrid methods provide powerful engines for modeling solution-phase chemistry. Yet before production begins, chemically diverse explicit-solvent systems must be parameterized, assembled at the desired composition, relaxed, equilibrated, and sampled. This preparatory “first mile” remains labor-intensive, requiring multiple software packages and manual customization, particularly for transition-metal complexes, solvent mixtures, and multicomponent electrolytes. Here, we present AutoSolvate 2.0, a modular, open-source framework that converts molecular structures and solution compositions into equilibrated, simulation-ready explicit-solvent ensembles. The new AutoMCPB module automates Metal Center Parameter Builder (MCPB.py)-based parameterization of transition-metal complexes (TMCs), while generalized system representations support multiple solutes and solvents at arbitrary molar, weight, or volume ratios. Standardized wrappers integrate force-field generation, box construction, minimization, heating, equilibration, classical sampling, and microsolvated-cluster extraction. Native classical molecular dynamics (MD) and TeraChem-based QM/MM workflows execute directly; equilibrated coordinates, velocities, trajectories, and clusters provide starting points for external quantum mechanical (QM), other QM/MM, MLIP training, and hybrid machine learning (ML)/classical workflows. JSON, guided command-line, and Model Context Protocol interfaces support high-throughput and agent-assisted use. We demonstrate these capabilities by reproducing the solvent-dependent absorption spectrum of an aqueous transition-metal photosensitizer, the composition-dependent microsolvation of water–acetonitrile mixtures, and characteristic ion-pairing features of a published multicomponent electrolyte simulation. Together, these applications establish AutoSolvate 2.0 as a preparation and initial-sampling layer connecting chemical system specification to classical, quantum, and machine-learning simulations. Rather than competing with higher-accuracy methods, AutoSolvate provides the on-ramp needed to apply them efficiently to complex explicit-solvent systems.
Related
- Fault-tolerant resource estimation for ground-state preparation via Lindblad simulation
- Multiscale Quantum-Classical Simulation of Protein Autoproteolysis in a Million-Atom Membrane Environment
- Unified Gas-Kinetic Wave-Particle Method for Multiscale Simulation of Vlasov-Poisson-Fokker-Planck System
- State preparation and detection for quantum simulation of particle collisions
- Rigorously justified local time stepping in the unified gas-kinetic wave-particle method for steady multiscale flow simulation
BibTeX
@misc{xun2026autosolvate,
title = {{AutoSolvate 2.0: Automated Preparation of Chemically Diverse Explicit-Solvent Systems for Multiscale Molecular Simulation}},
author = {Sangni Xun and Fangning Ren and Pinyuan Li and Xu Chen and Lechen Dong and Bushra Alam and Fang Liu},
howpublished = {ChemRxiv},
year = {2026},
month = oct,
doi = {10.26434/chemrxiv.15007305/v2},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15007305/v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - AutoSolvate 2.0: Automated Preparation of Chemically Diverse Explicit-Solvent Systems for Multiscale Molecular Simulation
AU - Sangni Xun
AU - Fangning Ren
AU - Pinyuan Li
AU - Xu Chen
AU - Lechen Dong
AU - Bushra Alam
AU - Fang Liu
PY - 2026
DA - 2026/10/13
PB - ChemRxiv
DO - 10.26434/chemrxiv.15007305/v2
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15007305/v2
AB - Advanced quantum mechanics/molecular mechanics (QM/MM), machine-learned interatomic potentials (MLIPs), and hybrid methods provide powerful engines for modeling solution-phase chemistry. Yet before production begins, chemically diverse explicit-solvent systems must be parameterized, assembled at the desired composition, relaxed, equilibrated, and sampled. This preparatory “first mile” remains labor-intensive, requiring multiple software packages and manual customization, particularly for transition-metal complexes, solvent mixtures, and multicomponent electrolytes. Here, we present AutoSolvate 2.0, a modular, open-source framework that converts molecular structures and solution compositions into equilibrated, simulation-ready explicit-solvent ensembles. The new AutoMCPB module automates Metal Center Parameter Builder (MCPB.py)-based parameterization of transition-metal complexes (TMCs), while generalized system representations support multiple solutes and solvents at arbitrary molar, weight, or volume ratios. Standardized wrappers integrate force-field generation, box construction, minimization, heating, equilibration, classical sampling, and microsolvated-cluster extraction. Native classical molecular dynamics (MD) and TeraChem-based QM/MM workflows execute directly; equilibrated coordinates, velocities, trajectories, and clusters provide starting points for external quantum mechanical (QM), other QM/MM, MLIP training, and hybrid machine learning (ML)/classical workflows. JSON, guided command-line, and Model Context Protocol interfaces support high-throughput and agent-assisted use. We demonstrate these capabilities by reproducing the solvent-dependent absorption spectrum of an aqueous transition-metal photosensitizer, the composition-dependent microsolvation of water–acetonitrile mixtures, and characteristic ion-pairing features of a published multicomponent electrolyte simulation. Together, these applications establish AutoSolvate 2.0 as a preparation and initial-sampling layer connecting chemical system specification to classical, quantum, and machine-learning simulations. Rather than competing with higher-accuracy methods, AutoSolvate provides the on-ramp needed to apply them efficiently to complex explicit-solvent systems.
ER -