Running Simulations

Recommended command-line workflow for first evaluation runs, controlled parameter changes, and reproducible output generation.

Operational guidance. For external evaluation, start with short NVT and NPT preparation stages followed by a modest production segment. Confirm the output structure, energy series, and analysis figures before extending trajectory length or adjusting refinement frequency.

Primary commands

neurocgmd run run.toml

Standard end-to-end workflow

Prepares the system, runs the configured stages, performs back-mapping, and writes analysis products.

neurocgmd prepare run.toml

Preparation only

Parses the input structure, builds the coarse-grained representation, and writes the prepared bundle used by downstream stages.

neurocgmd analyze run.toml

Analysis only

Rebuilds analysis products from an existing run context without repeating the dynamics stages.

neurocgmd info

Environment and capability check

Reports the installed command-line capabilities and is useful for installation validation prior to a formal evaluation run.

Recommended practice. Archive the exact run.toml used for each reported result together with the generated run_summary.json and key plots. That simple pairing makes internal review and external discussion much easier.

Example run sequence

cd project
neurocgmd info
neurocgmd run run.toml
neurocgmd analyze run.toml

Recommended project layout

project/
  structure.pdb
  run.toml
  outputs/
    my_system/
      prepared_bundle.json
      energies.csv
      traj.jsonl
      cg_trajectory.pdb
      aa_backmapped_trajectory.pdb
      run_summary.json
      plots/

Suggested execution sequence

1. Validate the environment

Confirm installation and command availability

2. Review the TOML file

Check stage names, step counts, and output paths

3. Run a short evaluation job

Begin with conservative equilibration and production settings

4. Inspect outputs

Confirm the trajectory products, energy series, and figure set

5. Scale parameters deliberately

Increase production length or adjust eval_stride after verification

6. Re-analyze when needed

Regenerate figures after parameter or analysis adjustments

Key run-control parameters

Parameter	Role	Practical guidance
`steps`	Length of each stage	Use shorter values for evaluation runs and extend only after workflow validation.
`time_step`	Integrator time step	Keep aligned with the documented force-field assumptions unless you are benchmarking method sensitivity.
`temperature`	Thermostat target	Set explicitly for reproducibility across runs and comparisons.
`ensemble`	Thermodynamic stage type	Use NVT and NPT deliberately as preparation stages before production when applicable.
`eval_stride`	Frequency of higher-cost refinement calls	This is the main performance-control knob. Larger values reduce refinement frequency; smaller values increase it.

← Return to Home

Open Quickstart Request Evaluation