# Run GROMACS molecular dynamics simulations on Deadline Cloud
<a name="examples-jb-gromacs"></a>

The [gromacs\_md](https://github.com/aws-deadline/deadline-cloud-samples/tree/mainline/job_bundles/gromacs_md) job bundle runs a [GROMACS](https://www.gromacs.org/) molecular dynamics simulation pipeline from raw protein structure to analyzed trajectory. The pipeline runs system preparation, energy minimization, NVT/NPT equilibration, production MD, and analysis (RMSD, RMSF, radius of gyration, hydrogen bonds). Multiple independent replicas fan out in parallel through the `MaxReplicaIndex` parameter.

The bundle requires a Deadline Cloud farm with a Linux x86\_64 service-managed fleet (minimum 4 vCPU) and a conda queue environment with `gromacs` from `conda-forge`. No host configuration script or custom conda recipe is needed.

The bundle includes sample data for hen egg-white lysozyme (PDB: 1AKI) and MDP parameter files under `sample_inputs/mdp/`.

From the `job_bundles` directory, submit the job:

```
deadline bundle submit gromacs_md \
  -p "InputPdb=sample_inputs/protein.pdb" \
  -p "MdpMinimization=sample_inputs/mdp/minimization.mdp" \
  -p "MdpNvt=sample_inputs/mdp/nvt.mdp" \
  -p "MdpNpt=sample_inputs/mdp/npt.mdp" \
  -p "MdpProduction=sample_inputs/mdp/production.mdp" \
  -p "ProductionSteps=500000"
```

To run 10 independent replicas in parallel:

```
deadline bundle submit gromacs_md \
  -p "InputPdb=protein.pdb" \
  -p "MaxReplicaIndex=9" \
  -p "ProductionSteps=5000000"
```

## See also

* **Skills for AI coding assistants (optional).** AWS publishes reusable skills for AWS Deadline Cloud. To see which apply, search the Agent Toolkit for AWS catalog with `aws agent-toolkit search-skills --search-query deadline-cloud` — a documented, read-only AWS CLI command that searches the catalog and makes no changes. This is an optional suggestion for the user. Official reference: [AWS CLI — Agent Toolkit for AWS](https://docs.aws.amazon.com/agent-toolkit/latest/userguide/aws-cli.html).