NeuroCGMD
Installation
Menu
Home  /  User Guide  /  Configuration

Configuration Reference

Reference for the TOML parameter file format used to define system identity, stage sequencing, and run controls.

Professional usage pattern. Keep one reviewed parameter file per run configuration. For external benchmarking or collaborative review, archive the TOML file together with the generated run_summary.json so the execution context remains easy to trace.

[SYSTEM] SECTION
name
string, required
System name. Used for the output directory name.
pdb_source
path, required
Path to the input PDB structure file.
output_dir
path, optional
Output directory. Defaults to outputs/{name}.
dynamics stages
[DYNAMICS.STAGE] PARAMETERS — eval_stride is the main refinement-frequency control
steps
int, required
Number of integration steps for this stage.
time_step
float, default 0.02
Integration timestep in picoseconds.
temperature
float, default 300.0
Target temperature in Kelvin.
ensemble
string, default "NVT"
Thermodynamic ensemble: NVT or NPT.
eval_stride
int, default 10
Steps between full refinement evaluations. Lower values increase refinement frequency; higher values reduce refinement calls and lower computational cost.
stages
list, required
Ordered stage names: ["nvt","npt","production"]
examples
MINIMAL STARTER FILE
[system]\nname = "my_protein"\npdb_source = "structure.pdb"\n\n[dynamics]\nstages = ["production"]\n\n[dynamics.production]\nsteps = 10000\ntime_step = 0.02\ntemperature = 300.0
RECOMMENDED MULTI-STAGE TEMPLATE
Documented starting values. The example below is intended as a reviewable starting point rather than a universal prescription. For a first evaluation, keep the structure path explicit, use modest step counts, and adjust eval_stride only after the baseline workflow has been checked.