Structural Biology MCP Agent

Provides MCP-driven access to PDB search, metadata, downloads, and structural modifications for protein structures.

python

0

GitHub Stars

python

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "sanatava-structural-biology-mcp": {
      "command": "python",
      "args": [
        "pdb_mcp_server.py"
      ]
    }
  }
}

You can use the Structural Biology MCP Agent to search, analyze, modify, and visualize protein structures from the RCSB PDB through a standardized MCP interface. This setup lets an AI client discover available tools, invoke them, and render results in an interactive workflow without hardcoding tools into your application.

How to use

You connect an MCP client to the Structural Biology MCP Server to perform a sequence of tools that start with locating structures and can proceed to modify and visualize them. Typical workflows include searching for structures by keyword, retrieving metadata, downloading coordinate files, mutating residues, swapping metals, removing ligands, and rendering the final structure in a 3D viewer. The server exposes all actions as MCP tools that your client can discover and invoke automatically, enabling flexible, composable analysis.

How to install

Prerequisites: ensure you have Python 3.10 or newer, and access to install Python packages.

Step 1: Install the MCP command line tools and Python dependencies.

pip install "mcp[cli]" requests streamlit anthropic py3Dmol
pip install -r requirements.txt

Step 2: Obtain the server code for the PDB MCP server. Clone the repository and navigate into it.

git clone https://github.com/sanatava/structural-biology-mcp
cd structural-biology-mcp

Step 3: Run the MCP server script that exposes the 11 tools for structural biology. Use the standard Python runtime as shown.

python pdb_mcp_server.py

Additional notes

You can also run the client side to interact with the server via an MCP client such as Claude Code or a Streamlit-based chat app. For example, you can start a chat interface that sends requests to the MCP server and receive structured tool results and visualizations.

What you can do with the 11 tools

The server provides a comprehensive set of tools to work with PDB structures. You can search, retrieve metadata, download coordinates, and perform modifications or cleanups to prepare structures for analysis or visualization.

Security and best practices

Limit access to the MCP server to trusted clients. Use API keys or authentication where supported by your MCP client, and avoid exposing the server to untrusted networks without protection.

Available tools

search_structures

Search PDB by keyword and filter by method or resolution to find relevant structures.

get_structure_info

Retrieve detailed metadata for a specific PDB ID, including authors, resolution, and experimental method.

download_structure

Download PDB coordinate files for a given structure ID to your local workspace.

search_by_uniprot

Find PDB structures linked to a specific UniProt accession.

get_structure_quality

Provide quality indicators such as resolution, R-factors, and Ramachandran validation.

replace_metal

Swap metals in HETATM records, enabling quick metal substitutions (for example Co to Zn).

mutate_residue

Mutate amino acids in the structure, specifying residue position and target residue.

remove_hetatm

Remove specific ligands or heteroatoms from the structure.

remove_chain

Remove a chain from a multi-chain structure to focus on a subset.

list_hetatm

List all ligands and heteroatoms present in the structure.

get_modified_structure

Retrieve a PDB file that reflects the latest modifications applied to the structure.