Domain 4 of 10
Force Fields and Dynamics
Force fields, conformers and molecular dynamics from gas-phase molecules to periodic water boxes with particle mesh Ewald.
4 Checked1 Checked (exact)10 Partial2 UnverifiedEngine 1.10.12 live
What You Can Compute
Small-Molecule Force Fields
Complete MMFF94 with conformers, strain and frequencies, plus UFF.
Biomolecular Force Fields
AMBER Cornell 1995, the ff99SB family, an ff14SB composite, OPLS-AA (unverified), GAFF ligands, CHARMM22 (CMAP not applied) and CGenFF.
Periodic Dynamics
NVE, NVT and NPT in triclinic cells with PME, Bussi, MTTK and Parrinello-Rahman, six water models.
Ab Initio Dynamics
Born-Oppenheimer dynamics on Hartree-Fock or DFT gradients up to 30 atoms, and builders for solvent boxes, bilayers, polymers, slabs and nanoparticles. Isolated MM dynamics runs up to 600 atoms, and 5000 atoms with the biomolecular force fields.
Every Method, with Its Level of Support
Generated from the engine's own capability report (engine 1.10.12), so this list never runs ahead of the code.
Checked reproduces printed values Checked (exact) no printed value exists; checked against exact mathematics or brute force Partial code exists, unchecked parts badged in the app Unverified built, no printed check yet
| Method | Support | Since | Primary Sources |
|---|---|---|---|
| Polarisable and materials force fields, ion setsDrude and Thole polarisation, Stillinger-Weber, Tersoff and EAM potentials, Joung-Cheatham and Li-Song-Merz ion sets. Tersoff silicon c11 and c12 reproduce Tersoff 1988; c44 is not computed. Drude, Thole, Stillinger-Weber, EAM and the ion sets are not yet checked against printed values. | Checked | 1.9.0 | Joung, Cheatham 2008; Li, Song, Merz 2015; Tersoff 1988 |
| System builders and trajectory analysisSolvent boxes with ions, lipid bilayers, polymers, amorphous cells, crystal supercells, slabs and nanoparticles; radius of gyration, SASA, DSSP, PCA, clustering and NetCDF trajectories. Lipid, ion and area-per-lipid parameters are yours. No space-group generation from a symbol, no Wulff shapes and no surface relaxation. | Unverified | 1.9.0 | |
| MMFF94Complete MMFF94 energies, minimisation, conformers, strain and frequencies. | Checked | 0.3 | Halgren 1996 |
| UFFUniversal force field across the periodic table. | Partial | 0.3 | Rappe et al. 1992 |
| AMBER Cornell 1995 and ff14SB compositeBiomolecular energies, optimisation, scans and MD. Parameters are those of Cornell et al. 1995 Table 14 (row counts and spot values compared with the printed table); energies and forces are checked by finite differences and invariance on N-methylacetamide with test charges. No printed energy or geometry is reproduced yet. The ff14SB option is a composite (Cornell 1995 base with the side-chain and psi terms of the Maier et al. 2015 Supporting Information), not the distributed ff14SB, and has no printed check. Missing parameters are refused, never invented. | Partial | 0.9.5 | Cornell et al. 1995; Maier et al. 2015 |
| CHARMM22 and CGenFFCHARMM functional form with residue typing from RTF records. CMAP is listed separately below. | Partial | 1.1.0 | MacKerell et al. 1998; Vanommeslaeghe et al. 2010 |
| Molecular dynamics (NVE, Nose-Hoover, Langevin)Lennard-Jones, MM, Born-Oppenheimer and MLP dynamics with exact restarts. Printed-value check covers the Lennard-Jones liquid only (Verlet 1967 Table I): the internal energy per atom at four state points within 0.02 and beta P/rho at five within 0.10, from Nose-Hoover chain runs of 256 or 864 atoms where Verlet ran 864 atoms at constant energy. The printed internal energy at density 0.88 and temperature 1.095 (-5.66) is not reproduced (ours -5.87). Thermostats are checked against exact canonical moments, and molecular mechanics, Born-Oppenheimer and MLP dynamics by energy conservation, time reversal and step-size scaling, not against printed values. Up to 600 atoms for MM in the isolated-molecule module, 5000 with the biomolecular force fields and 20000 for Lennard-Jones. | Checked | 0.7.40 | Martyna, Klein, Tuckerman 1992; Leimkuhler, Matthews 2013; Verlet 1967 |
| Born-Oppenheimer molecular dynamicsAb initio molecular dynamics on analytic Hartree-Fock or DFT gradients in NVE, Nose-Hoover chain, Langevin and Bussi ensembles, and metadynamics on ab initio forces. Up to 30 atoms and 20000 steps. No Car-Parrinello or extended-Lagrangian dynamics, and no NPT for isolated molecules. | Partial | 1.9.0 | |
| Periodic MD with Ewald and PMENVE, NVT and NPT boxes of TIP3P, SPC, SPC/E, TIP4P-Ew, TIP4P/2005 and OPC water. Printed-value checks: the Ewald sum reproduces the Wigner lattice constant 2.837297 (Yeh, Hummer 2004 Eq. 12) to six decimals; a Lennard-Jones NPT run at the pressure of the Verlet 1967 Table I state point (density 0.85, temperature 0.88) returns the density 0.849; the monomer dipole moments of SPC, SPC/E, TIP3P and OPC and the dipole and quadrupole moments of TIP4P-Ew and TIP4P/2005 equal the printed values. Liquid water: short NPT runs (216 molecules, PME, 12 ps) give 1.000 g/cm3 for SPC/E and 0.980 for TIP3P against the printed simulation values 0.998 and 0.982, which were obtained with plain cut-offs; this comparison is not like for like and is still marked pending. PME is checked against the Ewald sum, not against print, and no liquid property is reproduced for SPC, TIP4P-Ew, TIP4P/2005 or OPC. Triclinic cells are listed separately below. | Partial | 0.8.0 | Essmann et al. 1995; Miyamoto, Kollman 1992; Horn et al. 2004 |
| Conformer searchSystematic and stochastic conformers with MMFF94 ranking. MMFF94 conformer energies reproduce printed values; ring flips are not sampled. | Checked | 0.3 | |
| GAFF, AM1-BCC and RESP ligand chargesGAFF atom typing, ligand charges and protein plus ligand plus TIP3P systems. Parameters not printed in the paper are refused unless you supply them. | Partial | 1.4.0 | Wang et al. 2004; Jakalian, Jack, Bayly 2002; Bayly et al. 1993 |
| QM-derived ligand parameters (Seminario)Bond and angle force constants from the QM Hessian, Seminario and modified Seminario. Fit corrections are engineering choices. | Unverified | 1.4.0 | Seminario 1996; Allen et al. 2018 |
| ff99SB family, a99SB-disp, OPLS-AA and OpenFFAMBER ff99SB, ff99SB-ILDN, ff99SB*-ILDN, a99SB-disp with its water, TIP4P-D, OPLS-AA, OPLS-AA/M and OpenFF Parsley 1.0.0 parameters. DES-Amber refused with the reason. The AMBER sets are checked against their source tables; OPLS-AA and OPLS-AA/M are built and unverified. | Partial | 1.6.0 | Hornak et al. 2006; Lindorff-Larsen et al. 2010; Robustelli, Piana, Shaw 2018; Jorgensen, Maxwell, Tirado-Rives 1996; Robertson, Tirado-Rives, Jorgensen 2015; Qiu et al. 2021 |
| CHARMM22 CMAP and OPC waterCMAP backbone correction and the OPC four-site water model. CHARMM36 and ff19SB refused: parameter files not held. | Partial | 1.5.0 | MacKerell, Feig, Brooks 2004; Izadi, Anandakrishnan, Onufriev 2014 |
| Bussi thermostat, MTTK and Parrinello-Rahman barostats, r-RESPAStochastic velocity rescaling, MTTK and Parrinello-Rahman pressure coupling, multiple time step dynamics. No barostat with constraints. | Partial | 1.4.0 | Bussi, Donadio, Parrinello 2007; Martyna et al. 1996; Parrinello, Rahman 1981; Tuckerman, Berne, Martyna 1992 |
| Triclinic cells, LINCS and hydrogen mass repartitioningTriclinic water, Lennard-Jones and MM cells with PME, anisotropic and flexible NPT, LINCS constraints. Triclinic Lennard-Jones and MM cells and anisotropic or flexible NPT came with 1.5.0. Anisotropic MTTK reduction Unverified. | Partial | 1.4.0 | Hess et al. 1997; Essmann et al. 1995 |
| Steered MD, Jarzynski and trajectory analysisSteered MD with Jarzynski and Crooks work estimators, RDF, MSD, VACF, RMSD and H-bond analysis. No paper prints a value for the work estimators. Checked against exact properties: the Jarzynski and Crooks free energies of a pulled harmonic particle and of a distance pull agree with the exact free energy differences within 0.2 kT (400 trajectories); neighbour-list energies, forces and virials equal the all-pairs sums within a relative 1e-11; the radial distribution function of an ideal gas, mean-square displacement, velocity autocorrelation, superposition RMSD and hydrogen-bond lifetimes are checked on cases with known answers. Printed-value side check: the Lennard-Jones liquid energy and pressure from the radial distribution function agree with Verlet 1967 Table I within 0.02 and 0.07. Not checked: steered runs on molecular force fields beyond finishing with a finite estimate; no printed potential of mean force is reproduced. | Checked (exact) | 1.4.0 | Jarzynski 1997; Shirts et al. 2003; Yeh, Hummer 2004; Verlet 1967 |
Printed Value, Our Value
| Quantity | Source | Printed | Ours | Result |
|---|---|---|---|---|
| MMFF94 Validation Suite | Halgren 1996 | 761 molecules | none outside rounding | Match |
| MMFF94 Conformer Energies: Butane, Ethanol and Methyl Ethyl Ether Gauche Minus Anti, Kcal/mol | Halgren, Nachbar 1996 Table II | 0.78, 0.18, 1.50 | 0.78, 0.18, 1.50 | Match |
| Tersoff Silicon Elastic Constants c11, c12, GPa | Tersoff 1988, doi 10.1103/PhysRevB.37.6991 | 121, 86 | 121.7, 85.8 | Match |
| Lennard-Jones Liquid, Internal Energy per Atom at Rho 0.85 T 0.880; 0.45, 1.744; 0.75, 1.069; 0.75, 1.071 (Reduced Units) | Verlet 1967 Table I, doi 10.1103/PhysRev.159.98 | -5.94, -2.90, -5.19, -5.17 | -5.955, -2.909, -5.174, -5.166 (Nose-Hoover chain, 256 or 864 atoms; Verlet ran 864 at constant energy) | Match |
Exact Checks
Our output against exact mathematics or a recomputation inside the engine, never another program. These are not printed-value checks and are counted separately.
| Quantity | Checked Against | Tolerance | Ours | Result |
|---|---|---|---|---|
| Jarzynski and Crooks (BAR) Free Energy of a Pulled Harmonic Particle, 400 Trajectories, kT | Exact 3.6 from the Gaussian integral | 0.2 and 0.1 (three standard errors) | 3.63 and 3.58 | Within tolerance |
| Jarzynski Free Energy of a Distance Pull, 400 Trajectories, kT | 0.531 by quadrature | 0.1 (three standard errors) | 0.541 | Within tolerance |
| Neighbour-List Energy, Force and Virial in Lennard-Jones and Periodic MMFF94 Runs | All-pairs sums over the same pairs | relative 1e-12 (periodic virial 1e-11) | largest 1.4e-15 | Within tolerance |
Not Offered Today
- CHARMM36 and ff19SB (parameter files not held)
- Barostats with constraints
- Car-Parrinello and extended-Lagrangian dynamics
In the Chemistry Lab
Start, then Large systems and dynamics. Pick a force field, an ensemble and a length; runs checkpoint and resume exactly.
Molecular Mechanics and Dynamics Questions
What happens when a force field parameter is missing?
The run stops and names the missing parameter. Parameters are never invented or guessed.
Can I restart a long run?
Yes. Molecular dynamics writes checkpoints and resumes exactly from them.
See It on a Molecule of Your Own
Describe a molecule and your question; the assistant drafts the plan and you approve it. Or see it first in a 30 minute demo.