Quick start

This minimal example follows the same stages as a production workflow. Replace the short demonstration sequence and molecule count with the system described by your study.

1. Build a single-chain model

dps pdb2dps \
  -s MSEQNNTEMTFQIQRIYTKDISFEAPNAPHVFQKDW \
  -oc chain.pdb -op chain.itp -on PROT --seed 1215

The explicit seed records the model-building random state. The default is also 1215, but writing it in a workflow script makes the provenance visible.

2. Pack the simulation box

dps genmesh -f chain.pdb -p chain.itp -n 64 \
  -mesh 4 4 4 -oc system.pdb -op system.top --seed 1215

3. Prepare a portable run input

Copy and edit a packaged MDP template, then validate and compile the inputs:

cp dropps/share/templates/md.mdp md.mdp
dps grompp -f system.pdb -p system.top -m md.mdp -o run.tpr

The generated TPR v2 file stores a serialized OpenMM System, coordinates, topology data, parameters, and source checksums. It is a DROPPS format rather than a GROMACS TPR.

4. Run and validate

dps mdrun -s run.tpr -o run
dps check -s run.tpr -f run.xtc

A production GPU run should select the platform explicitly:

dps mdrun -s run.tpr -deffnm run \
  --platform CUDA -gpu_id 0 --precision mixed -cpt 10

5. Create groups and analyse

dps make_ndx -s run.tpr -o index.ndx
dps density -s run.tpr -f run.xtc -o density.xvg -x z

See the Command reference for the exact options of all 28 commands. For slab simulations, dps editconf can extend a compact box:

dps editconf -f compact.pdb -o slab.pdb -mz 10