Molecular dynamics simulation

What is molecular dynamics simulation?

Molecular dynamics simulation is a computer method for studying how atoms and molecules move and interact over time. It repeatedly calculates forces and updates particle positions to generate a time-ordered trajectory. Conventional protein and all-atom simulations preserve atomistic detail; coarse-grained models trade detail for scale; steered methods drive selected coordinates; replica exchange couples ensembles; and enhanced-sampling methods alter sampling to cross slow barriers.

Start with the scientific observable. Local contacts and hydration favor atomistic models; membrane organization or large assemblies may justify coarse graining; forced separation calls for steered MD; rugged equilibrium landscapes may motivate replica exchange; and a known slow coordinate can support metadynamics or umbrella sampling.

ProteinIQ directly runs conventional GROMACS and OpenMM workflows for supported systems. OpenMM also supports well-tempered metadynamics and simulated tempering through its Sampling method setting. The coarse-grained, steered, and replica-exchange pages provide post-run review workflows for externally generated trajectories. The enhanced-sampling page explains the supported OpenMM methods alongside a separate trajectory-review workflow; method-specific reweighting and convergence assessment remain the researcher's responsibility.

When to use molecular dynamics simulation

• Examine stability, flexibility, interactions, transitions, and ensemble organization.
• Connect a defined system and protocol to a measurable structural or thermodynamic observable.
• Keep conventional simulation, forced pathways, coarse-graining, replica exchange, and biasing distinct.

Strengths and limitations

Connects molecular structure to motion, interactions, and ensemble behavior. Supports atomistic detail, coarse-grained organization, forced response, and advanced sampling. Preserves trajectories, settings, energies, logs, and analysis definitions.

Slow states and transitions may remain unobserved even in long trajectories. Force fields, mapping, parameters, solvent, and protonation affect results. Driven and biased trajectories require specialized interpretation and cannot be treated as conventional equilibrium MD.

Which molecular dynamics tool to use

Most runs start from a protein PDB file or an RCSB entry. You choose the engine, force field, solvent and simulated length, and ProteinIQ prepares, minimizes and equilibrates the system before production. GROMACS and OpenMM runs start at 100 credits and are billed by runtime, so system size and simulated length drive the cost; the quote appears before you submit.

What you want to do	Tool	What you get back
Simulate a solvated protein from a PDB file	GROMACS	A trajectory, final structure, and RMSD, RMSF, Rg, DSSP and energy tables
Run MD with classical or ML potentials, or metadynamics	OpenMM	Trajectories, restart files and energies
Analyze a trajectory you already have	MD Trajectory Analysis	RMSD, RMSF, Rg and other per-frame observables
Estimate a binding free energy from a complex trajectory	gmx_MMPBSA	MM/PBSA or MM/GBSA totals with a component breakdown
Sample equilibrium conformations without running MD	BioEmu	An ensemble of protein structures
Clean up a structure before simulating it	PDBFixer	A structure with missing atoms and residues restored

GROMACS is the usual choice for a conventional protein simulation: it handles setup, solvation and equilibration and runs production up to 200 ns, with the standard analyses already done. OpenMM suits runs that need a machine-learning potential or OpenMM's sampling methods. BioEmu is a generative model, not a simulation, so its ensemble carries no time ordering or kinetics.

Choosing a molecular dynamics simulation type

Choose resolution first, then sampling strategy. All-atom and coarse-grained describe representation; steered MD describes a driven perturbation; replica exchange describes interacting ensembles; enhanced sampling is a broader family that includes several biasing and generalized-ensemble strategies.

For a conventional protein run, see protein molecular dynamics simulation or all-atom molecular dynamics. For metadynamics and simulated tempering in OpenMM, see enhanced sampling. The review branches cover coarse-grained molecular dynamics, replica exchange, and steered molecular dynamics.

A project can combine categories, but the name should stay precise. Replica exchange is an enhanced-sampling method, while an all-atom system may also be steered. Report both the representation and the sampling protocol instead of replacing one with the other.

How to run molecular dynamics simulation online

A defensible trajectory begins with a defined system and ends with convergence-aware analysis, not with a single RMSD plot.

1. Define the observable. State the structural, kinetic, mechanical, or thermodynamic quantity the model should address.
2. Prepare the system. Resolve assembly, atoms, protonation, ligands, parameters, solvent, ions, and boundaries.
3. Choose the protocol. For a supported conventional run, select GROMACS or OpenMM and set force field, solvent, ensemble, duration, and output cadence. For well-tempered metadynamics or simulated tempering, select the corresponding Sampling method in OpenMM and use a fixed-step integrator. For other protocols, prepare compatible external trajectory and method records for review.
4. Run and monitor. Retain logs, energies, checkpoints, trajectories, warnings, and every independent start.
5. Validate the ensemble. Test equilibration, convergence, replicate agreement, uncertainty, and sensitivity to analysis choices.

Molecular dynamics simulation applications

Molecular dynamics supports protein stability and flexibility analysis, interaction review, conformational hypotheses, membrane and assembly modeling, forced pathway generation, and free-energy calculations. The simulation type determines which of those outputs is scientifically available.

Trajectory length alone is not a quality metric. A shorter, replicated, well-parameterized study with declared uncertainty can support a stronger claim than one long trajectory whose system, ensemble, or slow coordinates are poorly controlled.

How to interpret molecular dynamics simulation

Every observable has a definition: alignment selection for RMSD, reference and units for distance, geometric criteria for contacts, statistical weights for biased ensembles, and state reconstruction for replica exchange. Preserve those definitions with the result.

Distinguish sampled behavior from experimental truth. Force fields and coarse-grained models are approximations, trajectories are finite, driven simulations are nonequilibrium, and enhanced-sampling results require method-specific statistical treatment.

How the featured workflow works

The hub workflow prepares one protein structure with PDB Fixer and branches into independent GROMACS and OpenMM configurations.

1. Define question. Specify the observable, system, resolution, ensemble, and required timescale.
2. Prepare structure. Review assembly, missing atoms, protonation, ligands, cofactors, solvent, and ions.
3. Configure engines. Set independent GROMACS and OpenMM force-field and ensemble choices.
4. Run simulations. Retain trajectory, energy, topology, logs, warnings, and checkpoints.
5. Compare evidence. Assess equilibration, structural observables, uncertainty, and replicate behavior.

Inputs and outputs for the featured workflow

Check formats before running, then inspect and download the result from every workflow step.

Inputs

Molecular system

PDBmmCIFSDFMOL2

Prepared protein or protein–ligand coordinates with compatible parameters and system conditions.

Outputs

Simulation evidence

XTCPDBEDRTPRCSVZIP

Engine-native trajectories, energies, topologies, coordinates, analyses, logs, and archives.