molclaw-quickvina-docking

QuickVina2 Molecular Docking

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.

step 1. Use skill molclaw-protein-structure-retrieve to get the target protein structure file. If the target protein structure file has been provided, skip this step.

step 2. If the user specifies a target chain or several chains, or if the agent autonomously identifies single-chain or multi-chain structures requiring extraction, invoke the tool extract_and_save_chains to generate and save the corresponding structure as a new PDB file. Otherwise, skip this step.

response = await tool_client.session.call_tool(
    "extract_and_save_chains",
    arguments={
        "pdb_file_path": pdb_path,
        "chain_ids": chain_ids		##Chain IDs (e.g., ["A", "C"]) 		
    }
)
result = tool_client.parse_result(response)
pdb_path = result["out_file"]

step 3. Use skill molclaw-pdbfixer to repair the protein structure file using the settings as below.

response = await client.session.call_tool(
    "fix_pdb",
    arguments={
        "input_path": pdb_path,
        "add_hydrogens": True,
        "ph": 7.0,
        "remove_heterogens": True,
        "remove_water": True,
        "replace_nonstandard": True
    }
)
result = client.parse_result(response)
fixed_pdb_path = result["output_file"]

step 4. Use skill molclaw-fpocket or molclaw-p2rank to detect binding sites on the protein structure and return pocket information of the best one. If the pocket center and box size are already known (e.g., from a co-crystal ligand), skip this step and use the known values directly.

step 5. Use tool molecule_docking_quickvina_fullprocess to perform molecular docking. This is a full-process tool — it accepts a PDB file and SMILES string directly and handles all format conversions (PDB→PDBQT, SMILES→PDBQT) internally. Do NOT manually convert to PDBQT before calling this tool.

Tool description:

Perform molecular docking using QuickVina2-GPU (Accelerated version of AutoDock Vina).
The server selects the output directory. The current QuickVina2-GPU backend accepts at most 47.625 Å per docking-box axis, 129 movable ligand atoms, and 47 ligand torsions.
Args:
    pdb_file_path (str): Path to the protein receptor file (format .pdb)
    smiles (str): Input molecule SMILES string
    pocket_center_x (float): X-coordinate of the docking pocket center
    pocket_center_y (float): Y-coordinate of the docking pocket center
    pocket_center_z (float): Z-coordinate of the docking pocket center
    pocket_size_x (float): Size of the docking pocket along the X-axis (default 25.0)
    pocket_size_y (float): Size of the docking pocket along the Y-axis (default 25.0)
    pocket_size_z (float): Size of the docking pocket along the Z-axis (default 25.0)
Return:
    status (str): success/error
    msg (str): message
    docking_affinity_value (float): Docking affinity value, unit kcal/mol
    docking_file (str): A PDBQT file contains docking poses, atom types, and charges for analyzing binding results.

Tool Usage:

for smiles in smiles_list:
    response = await client.session.call_tool(
        "molecule_docking_quickvina_fullprocess",
        arguments={
            "pdb_file_path": fixed_pdb_path,
            "smiles": smiles,
            "pocket_center_x": best_pocket["center_x"],
            "pocket_center_y": best_pocket["center_y"],
            "pocket_center_z": best_pocket["center_z"],
            "pocket_size_x": max(25.0, best_pocket.get("size_x", 25.0)),
            "pocket_size_y": max(25.0, best_pocket.get("size_y", 25.0)),
            "pocket_size_z": max(25.0, best_pocket.get("size_z", 25.0))
        }
    )
    result_data = client.parse_result(response)
    docking_affinity = result_data['docking_affinity_value']
    docking_pose_file = result_data['docking_file']