Molclaw Protein Openmm

Protein OpenMM MD and Frame Extraction

Note:

• Local files are not directly accessible by the server. Please upload them to the server using molclaw-file-transfer before execution.
• For PDB file inputs, it is recommended to preprocess them using molclaw-pdbfixer before execution.
• Please refer to skill molclaw-scp-server to complete tool invocation.

Usage

1. Protein OpenMM MD

The description of tool proteinopenmmmd.

Runs OpenMM-based protein molecular dynamics preparation and simulation for structure refinement workflows.
Args:
    protein_pdb (str): Absolute or relative path to input protein PDB.
    solvent_type (str): Solvent mode, 'explicit' or 'implicit', default 'explicit'.
    gb_model (str): GB model for implicit solvent mode, default 'GBn2'.
    water_model (str): Water model for explicit solvent mode, default 'tip3p'.
    force_field (str): OpenMM force field name, default 'amber14'.
    md_time (float): Production MD time in picoseconds, default 100000.0.
    platform (str): OpenMM compute platform, default 'CUDA'.
    full_md (bool): Run full MD procedure if True, default False.
    dry_run (bool): Validate setup and produce run directory without simulation, default False.
Return:
    status (str): 'success' or 'error'.
    msg (str): Human-readable execution summary.
    output_dir (str): Run-specific directory under tool_result/openmm_md_result.
    work_dir (str | None): Final OpenMM working directory containing generated files.
    protein_pdb (str): Resolved input protein path.
    trajectory_path (str | None): Path to md_traj.dcd when available.
    topology_path (str | None): Path to a topology PDB used for frame extraction.
    generated_files (List[str]): File paths relative to work_dir.
    md_time (float): Echoed requested MD time in ps.
    solvent_type (str): Echoed solvent mode.
    force_field (str): Echoed force field.
    full_md (bool): Echoed full MD mode.

How to use tool proteinopenmmmd :

response = await client.session.call_tool(
    "protein_openmm_md",
    arguments={
        "protein_pdb": "/path/to/input.pdb",
        "solvent_type": "implicit",
        "gb_model": "OBC2",
        "water_model": "tip3p",
        "force_field": "amber14",
        "md_time": 1000.0,
        "platform": "CUDA",
        "full_md": True,
        "dry_run": False
    }
)
result = client.parse_result(response)
key_output = result["work_dir"]

Example parameter sets

# 1) Main mode
{
    "protein_pdb": "/path/to/input.pdb",
    "solvent_type": "implicit",
    "gb_model": "OBC2",
    "water_model": "tip3p",
    "force_field": "amber14",
    "md_time": 1000.0,
    "platform": "CUDA",
    "full_md": True,
    "dry_run": False
}

# 2) Variant mode
{
    "protein_pdb": "relative/path/to/protein.pdb",
    "solvent_type": "explicit",
    "water_model": "tip3p",
    "force_field": "charmm36",
    "md_time": 10000.0,
    "platform": "CUDA",
    "full_md": False,
    "dry_run": False
}

2. OpenMM Trajectory Frame Extraction

The description of tool openmmextractframes.

Extracts evenly spaced protein conformations from an OpenMM work directory for downstream screening and ensemble analysis.
Args:
    work_dir (str): OpenMM MD output directory containing topology and trajectory files.
    num_frames (int): Number of evenly spaced frames to extract, default 100.
    protein_only (bool): Keep only protein atoms in extracted frames, default False.
    align (bool): Align extracted structures to the first frame, default False.
    prefix (str): Filename prefix for extracted PDB frames, default 'frame'.
    dry_run (bool): Validate inputs and prepare output directory without extraction, default False.
Return:
    status (str): 'success', 'partial_success', or 'error'.
    msg (str): Human-readable extraction summary.
    output_dir (str): Run-specific directory under tool_result/openmm_md_result.
    work_dir (str): Resolved OpenMM working directory.
    topology_path (str | None): Resolved topology file path.
    trajectory_path (str | None): Resolved trajectory file path.
    frames_dir (str): Directory where extracted frame PDB files are saved.
    frame_count (int): Number of extracted frame files.
    frame_files (List[str]): Extracted frame file paths relative to output_dir.

How to use tool openmmextractframes :

response = await client.session.call_tool(
    "openmm_extract_frames",
    arguments={
        "work_dir": "/path/to/work_dir",
        "num_frames": 100,
        "protein_only": False,
        "align": False,
        "prefix": "frame",
        "dry_run": False
    }
)
result = client.parse_result(response)
key_output = result["frame_files"]

Example parameter sets

# 1) Main mode
{
    "work_dir": "/path/to/work_dir",
    "num_frames": 100,
    "protein_only": False,
    "align": False,
    "prefix": "frame",
    "dry_run": False
}

# 2) Variant mode
{
    "work_dir": "relative/path/to/openmm_md_output",
    "num_frames": 50,
    "protein_only": True,
    "align": True,
    "prefix": "conf",
    "dry_run": False
}

3. End-to-End Collaboration Workflow

Use the two tools in sequence via API calls:

1. Call proteinopenmmmd to generate MD outputs and get work_dir.
2. Pass that work_dir into openmmextractframes to extract evenly spaced PDB frames.

client = DrugSDAClient("http://180.184.86.2:32208/mcp")
if not await client.connect():
    print("connection failed")
    return

md_resp = await client.session.call_tool(
    "protein_openmm_md",
    arguments={
        "protein_pdb": "/path/to/input.pdb",
        "solvent_type": "implicit",
        "gb_model": "OBC2",
        "md_time": 1000.0,
        "full_md": True
    }
)
md_result = client.parse_result(md_resp)
work_dir = md_result["work_dir"]

frames_resp = await client.session.call_tool(
    "openmm_extract_frames",
    arguments={
        "work_dir": work_dir,
        "num_frames": 100,
        "prefix": "frame"
    }
)
frames_result = client.parse_result(frames_resp)
key_output = frames_result["frame_files"]

await client.disconnect() 

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: InternScience
• Source: InternScience/MolClaw
• License: MIT

Install and usage instructions live in the source repository linked above.