Protein Data Bank MCP Server

Micro-service for accessing the Protein Data Bank (PDB) API through the MCP protocol. This service provides structured access to protein structures, chemical components, and related annotations from the PDB database.

Features

• Access to PDB core API endpoints
• Structured data retrieval for proteins, assemblies, and chemical components
• Support for annotation data from DrugBank, UniProt, and PubMed
• Built on FastMCP framework with async/await support
• Local PDB file storage and parsing capabilities

Setup

1. Install dependencies:

uv sync

2. Configure environment variables in .env:

PDB_MCP_HOST=localhost
PDB_MCP_PORT=8080

Available Tools

Core Data

• structural_assembly_description: Get assembly structures
• chemical_component: Get chemical component details
• polymer_entity: Get polymer entity information
• structure: Get structure details

Annotations

• drugbank_annotations: DrugBank data for compounds
• uniprot_annotations: UniProt protein annotations
• pubmed_annotations: PubMed literature references

Repository Info

• current_entry_ids: List current PDB entries
• structure_status: Check entry status
• unreleased_structures: Get unreleased structure info

Docker Usage

Run with Docker Compose:

docker compose up pdb-server

The server will be available at http://localhost:8080.