pdb-database

What it does

Instructs an agent to access the RCSB PDB public endpoints to perform programmatic structural biology queries. It covers searching by text, attributes, sequence similarity, and 3D structure similarity, retrieving metadata via GraphQL, and downloading coordinate files. It also provides examples for batch requests, pagination, and combining queries with group/logical_operator structures.

How it works

• Use the Search API at https://search.rcsb.org/rcsbsearch/v2/query with POST and a JSON payload to find PDB IDs by text, attribute filters, sequence, or 3D similarity.
• Retrieve structured metadata via the Data API at https://data.rcsb.org/graphql with POST and a GraphQL query (entry, polymer_entity, assembly, chem_comp, etc.).
• Download coordinate files from https://files.rcsb.org/download/{id}.{format} using GET, returning mmCIF, PDB, or FASTA formats.
• For AlphaFold predictions, reference alphafold-database-access; for protein sequences only, use uniprot-protein-database.
• Includes code samples for text_search, attribute_search, sequence/module usage, structure similarity search, and batch retrieval.
• Demonstrates how to compose complex queries with group + logical_operator, including AND/OR combinations and date-range filtering.
• Shows batch pagination patterns with start/rows and a small delay between requests to respect rate limits.

When to use it

• When searching for structures by keyword or property, filtering by organism, method, or resolution.
• When retrieving metadata like resolution, method, organism, and deposition date for sets of structures.
• When downloading coordinate files for simulation, docking, or visualization.
• When building datasets for machine learning or drug discovery pipelines that require structured metadata and coordinate data.
• When needing sequence, polymer_entity, or organism information via GraphQL for multiple entries in one request.

What it can touch

• The REST endpoint for search: https://search.rcsb.org/rcsbsearch/v2/query
• The GraphQL endpoint for data: https://data.rcsb.org/graphql
• The Files endpoint for downloads: https://files.rcsb.org/download/{id}.{format}
• It uses the Python package requests (and optionally biopython for parsing files) as prerequisites.
• No API key is required; rate limits are managed via polite delays and optional exponential backoff on HTTP 429.

Caveats

• License is BSD-3-Clause for the skill description; no explicit runtime license noted in the provided content.
• Requires internet access to public endpoints; the skill demonstrates basic usage patterns but does not guarantee specific response formats beyond examples.
• Some features reference other databases or tools (alphafold-database-access, uniprot-protein-database) rather than the primary endpoints.
• Examples rely on Python code blocks; execution assumes a Python environment with requests installed.