pdb-database

PDB Database - Protein Structure Database Query Tool

Skill Overview

Access the RCSB PDB global protein structure database to search and download 3D structures of proteins/nucleic acids. Supports text, sequence, and structural similarity searches, suitable for structural biology and drug discovery research.

Applicable Scenarios

1. Structural Biology and Protein Engineering Research

Search for experimental structures of target proteins, obtain homologs for modeling, analyze structural differences of mutants, and study sequence-structure relationships. Supports multi-dimensional filtering by species, resolution, experimental method, etc., to quickly locate required structure data.

2. Drug Discovery and Molecular Docking

Obtain 3D structures of target proteins for virtual screening and molecular docking, analyze ligand binding site features, compare protein–ligand complex structures, and identify similar drug targets. Supports batch processing of multiple target structures to improve drug development efficiency.

3. Computational Chemistry and Bioinformatics Analyses

Download PDB/mmCIF coordinate files for molecular dynamics simulations, perform structural alignment and evolutionary analyses, and integrate structural data into computational pipelines. Provides a Python API for automating large-scale structure processing.

Core Features

1. Multi-dimensional Structure Search

Supports text search, attribute filtering, sequence similarity, and structural similarity searches. Allows precise filtering by organism, resolution, experimental method, and other attributes, and supports combined queries for building complex search conditions. Access to over 200,000 experimentally determined structures and computational models.

2. Structure File Downloads

Offers downloads in PDB, mmCIF, and BinaryCIF formats, supporting single structures or biological assembly files. Coordinate files can be downloaded directly via API or URL and integrated into automated workflows.

3. Metadata Retrieval and Batch Processing

Retrieve metadata such as structure titles, experimental methods, resolution, and organism source, with flexible GraphQL queries. Supports batch processing of multiple PDB IDs to extract structure summaries and quality metrics, facilitating large-scale data analysis.

Frequently Asked Questions

What is the PDB database?

RCSB PDB is the official portal of the worldwide Protein Data Bank, which archives 3D structure data of biological macromolecules (proteins, nucleic acids, etc.). It currently includes over 200,000 structures determined by experiments such as X-ray crystallography, cryo-EM, and NMR, as well as computational models from sources like AlphaFold.

How do I download protein structure files?

You can download directly from the RCSB PDB website or use the Python API (rcsb-api package) for automated retrieval. Download links follow the format https://files.rcsb.org/download/{PDB_ID}.pdb (PDB format) or .cif (mmCIF format). mmCIF is recommended as the modern standard format and supports larger structures.

What is the difference between RCSB PDB and AlphaFold?

RCSB PDB contains experimentally determined structures (observed data), while AlphaFold provides AI-predicted structure models. Experimental structures are highly accurate but have limited coverage; predicted models have broader coverage but may contain local errors. For critical applications like drug discovery, prioritize experimental structures; predicted models can be used for preliminary analysis and template searching.

Which should I choose, PDB or mmCIF format?

mmCIF (also called PDBx) is recommended. mmCIF is the modern standard format, supporting larger structures and more complete metadata without column width limits. PDB is a legacy format suitable only for smaller structures and has strict fixed-width field limitations. Most modern structural biology software supports mmCIF.

Are there limitations to this skill?

Network access is required to use the RCSB PDB API, and some advanced features may be subject to request rate limits. The skill itself provides data access interfaces only; structure analysis and visualization require complementary tools such as BioPython and PyMOL.