Gromacs md

GROMACS Molecular Dynamics

Use GROMACS for classical biomolecular MD workflows: protein or protein-ligand systems, membrane proteins, conformational dynamics, trajectory analysis, and approximate binding free energy post-processing.

This skill is intentionally lightweight. Use it to enforce the scientific workflow and risk checks; use local gmx, official GROMACS documentation, and optional scholaraio toolref for exact command syntax and version-specific .mdp details.

Local Environment

Prefer gmx on PATH. If GROMACS is installed elsewhere, set:

$env:GMX_EXE = "C:\path\to\gromacs\bin\gmx.exe"

Optional: set PYMOL_EXE for PyMOL rendering helpers in the companion skills.

The local scholaraio toolref GROMACS index contains the official current mdp-options source as version manual-current-mdp with 323 records. Full scholaraio toolref fetch gromacs may still require GitHub access; if it fails, the installed MDP index is enough for high-risk .mdp parameter lookup.

Verify before running real work:

gmx --version

For ligand and free-energy workflows, check optional tools before promising execution:

Get-Command acpype,antechamber,gmx_MMPBSA -ErrorAction SilentlyContinue

Protocol

Classify the user's request first:

1. gmx subcommand or command-line troubleshooting
2. .mdp parameter design or review
3. topology, force field, ligand, membrane, or box-building workflow
4. production execution and performance
5. trajectory analysis, visualization, or MM/PBSA-style post-processing

Before writing .mdp files, verify high-risk parameters: thermostat, barostat, constraints, cutoffs, PME, neighbor search, output frequency, coupling groups, timestep, and production ensemble.

If scholaraio is installed, use toolref-first:

scholaraio toolref search gromacs "temperature coupling"
scholaraio toolref show gromacs mdp integrator
scholaraio toolref show gromacs mdp pcoupl
scholaraio toolref show gromacs mdp tau-t
scholaraio toolref show gromacs mdp ref-t

If scholaraio is absent or incomplete, continue with local gmx help, gmx help <command>, and current official GROMACS documentation. Mention the coverage gap, but do not make the user solve it.

Literature-Aware Workflow

Before starting a simulation, establish literature or benchmark context when the target system is specific enough:

1. Search for comparable systems, ligands, force fields, temperatures, pressures, salt concentrations, simulation lengths, and validation metrics.
2. Use those sources to justify force field, water model, ion concentration, ensemble, and analysis metrics.
3. Add short comments to generated .mdp files for non-obvious parameter choices.
4. Compare final results with literature or experimental expectations when possible.

If scholaraio usearch is available, use it for the literature search. Otherwise use web search, PubMed, Crossref, or user-provided papers.

Standard Workflow

Follow this lifecycle unless the user is only asking for a narrow command or diagnosis:

1. Prepare structures: clean PDB, choose protonation states, inspect missing residues/atoms, resolve alternate conformers, and separate protein, ligand, cofactors, waters, and ions.
2. Choose a consistent force-field family. Do not mix CHARMM protein with GAFF ligand unless the user explicitly accepts the caveat and the workflow is justified.
3. Parameterize ligands with traceable tools and charge model. For AMBER-family ligand work, prefer GAFF2 with AM1-BCC via AmberTools/ACPYPE when appropriate.
4. Generate topology, define box, solvate, and add ions with neutralization plus physiological salt when scientifically appropriate.
5. Run energy minimization and inspect maximum force, potential energy, and geometry.
6. Run restrained NVT, then restrained NPT. Use equilibration diagnostics, not only successful command exit.
7. Run production MD using production-grade coupling settings.
8. Analyze stability and mechanism: RMSD, RMSF, radius of gyration, H-bonds, contacts, distances/angles, clustering, PCA, density, or other metrics tied to the biological question.
9. For MM/PBSA-style estimates, sample equilibrated frames, report mean and uncertainty, and avoid treating a single value as a final conclusion.

Scientific Checks

Check	Preferred practice	Common failure
Force field	CHARMM36m or modern AMBER for proteins when suitable	Old tutorials copied without justification
Ligand parameters	Traceable source, compatible force-field family, documented charges	Black-box topology with unknown charges
Force-field consistency	Keep protein, ligand, water, ions, and membranes compatible	Mixing CHARMM and AMBER/GAFF casually
Production barostat	Use a production-grade barostat such as Parrinello-Rahman when appropriate	Berendsen used for production statistics
Box	Dodecahedron or other efficient shape when suitable	Oversized cube by default
Salt	Neutralize and consider 0.15 M NaCl for physiological solution	Only neutralization with no rationale
MM/PBSA sampling	Use equilibrated windows, often late trajectory frames	Whole trajectory including equilibration
Reporting	Report mean, spread/uncertainty, and interpretation	Single RMSD plot or single delta-G number

Always tie analysis back to the user's scientific question. "The trajectory ran" is not the same as "the simulation supports an interpretation."

Common Outputs

For planning tasks, return a reproducible workflow with files, commands, expected inputs, expected outputs, and validation checks.

For generated .mdp files, include concise comments for critical choices and separate minimization, NVT, NPT, and production settings.

For debugging tasks, inspect the exact error text, relevant input files, GROMACS version, force field, and command used. Fix the immediate issue and flag downstream scientific consequences.

For analysis tasks, produce plots or tables where practical and name the trajectory interval, atom selections, frame stride, and uncertainty method.

References

Read references/gromacs-mdgrok-source.md when you need the original mdgrok Chinese guidance or its exact behavior rules.