New PDBe-KB aggregated views of protein structure

21 March 2019

We are delighted to announce the new protein-centric aggregated views, developed as part of our new collaborative resource Protein Data Bank in Europe Knowledge Base (PDBe-KB).

PDBe-KB is a community-driven resource managed by the PDBe team, collating functional annotations and predictions for structure data in the PDB archive, giving researchers a more comprehensive view of publicly available protein structure data.

PDBe entry pages traditionally focus on single PDB entries, with limited information on entries that are related to the same macromolecule. The PDBe-KB team is proud to present a new type of entry page, that focuses on full length proteins instead of single PDB entries. Currently, you can use PDB or UniProt identifiers of proteins to display all the related PDB data, from the list of all the available PDB entries, to functional annotations of ligand binding sites, macromolecular interaction interfaces, and related publications.

Comprehensive view

"Thousands of labs around the world produce data and annotations for protein structures," says Sameer Velankar, Team Leader of PDBe. "Today's challenge is making all the data available in one place and presenting them in an accessible way. Right now, there are many databases that contain protein structure data, but they lack the exposure and accessibility to enable them to be used more widely. PDBe-KB aims to integrate protein data and annotations from existing resources and present everything in a useful way."

Multiple uses

One of the innovative elements of PDBe-KB are the protein-specific aggregated views. These pages highlight the available information related to structures for specific proteins, including structural and functional annotations, domains, ligand-binding sites and more. The protein pages can save researchers a lot of time by essentially answering the question 'what do we currently know about this protein?'.

PDBe-KB also enables researchers to see what molecules bind to their protein of interest, highlighting sites of protein-protein interactions and small molecule binding. This information is particularly useful for drug development and can help researchers prioritise which drugs could interact in a desired way with a specific protein.

PDBe-KB could prove useful for researchers studying small molecules, pathways or specific diseases and conditions. It aims to open up protein structure data to researchers who don't necessarily have structural biology training.

The bigger picture

"The difference between PDBe and PDBe-KB is a bit like the difference between satellite images and Google Maps," says Mihaly Varady, Scientific Programmer at EMBL-EBI. "Satellite images give us a good idea of what is there, but you need a lot of background knowledge to understand what you're looking at. Instead, Google Maps is easy to use and offers additional information from other sources. Just like Google Maps can offer a comprehensive view of a street, PDBe-KB enables scientists to visualise all the available data on a specific protein in an intuitive way."

PDBe-KB aggregated views are currently based on UniProt accession numbers, which are widely used in the life sciences community. In future, the infrastructure developed by the PDBe-KB resource could enable aggregated views for other entities within PDB data, broadening this beyond individual Uniprot accession and, for instance, focussing on small molecules or complexes.

PDBe-KB currently collaborates with 24 resources from seven countries that provide predicted and evidence-based annotations of structural data derived from PDB structures. As a project coordinator, EMBL-EBI is responsible for creating the data infrastructure, access mechanisms and maintaining community agreed standards for the new PDBe-KB resource.