The Protein Data Bank in Europe (PDBe) update paper published in Nucleic Acids Research on 14 November 2026 documents expanded literature annotations, molecular dynamics links and AI-compatible access to structural biology data. The authors bring together ongoing developments in the resource. For universities, a practical distinction is the separation between richer evidence and the infrastructure needed to query it.
Literature and dynamics add context to structures
The authors report literature-derived residue annotations covering 120,569 PDB entries from 18,453 publications with licenses allowing distribution of derived data. Their October 2026 manuscript snapshot also lists annotations for 32,427 additional structures withheld because of licensing restrictions. The expansion includes publications beyond a structure’s original paper, connecting later findings to existing entries.
The update also describes cross-links to MDposit simulations and simulation-derived solvent-accessibility and membrane annotations. These complement the experimental structures archived in the Protein Data Bank. PDBe-KB explains that its consortium collates functional, biophysical and biochemical annotations from partner resources.
Three MCP components serve different research tasks
The paper describes Model Context Protocol (MCP) servers for PDBe API, Search and Graph access. The PDBe software repository distinguishes the underlying tasks:
| Component | Research task | Access distinction |
|---|---|---|
| API | Retrieve entries, assemblies, ligands, publications and validation data | Tools expose PDBe REST operations |
| Search | Find structures using filters, facets, grouping and pagination | Queries use PDBe’s Solr search service |
| Graph | Inspect the knowledge graph schema and query relationships | Database queries require a local PDBe-KB Neo4j setup |
The published Python package provides integration with MCP-compatible AI clients. Its documentation states that PDBe does not provide a public running graph database for this server and directs most users toward API and Search. The repository makes the software available under Apache License 2.0.
What the update means for university research support
For research offices and university computing teams evaluating adoption, the distinction makes local graph infrastructure a separate support decision. PDBe-KB’s annotation role also makes the origin of each result consequential: an experimental structure, a mined literature statement and a simulation-derived property represent different evidence.
The paper says the underlying PDBe services remain authoritative and API tool responses add no separate provenance object. The practical implication for research data provenance is to retain identifiers, query details and software information alongside AI-assisted results. For teams applying the FAIR Data Principles, this makes traceable reuse across those evidence layers a priority.
Sources
- Nucleic Acids Research: PDBe in 2027, published 2026-11-14; checked 2026-11-15.
- PDBeurope: PDBe MCP Servers repository, checked 2026-11-15.
- PyPI: pdbe-mcp-server documentation, checked 2026-11-15.
- EMBL-EBI: PDBe-KB background and FAQ, checked 2026-11-15.
