EMDB Map Data and Data Models

Introduction

The Electron Microscopy Databank (EMDB) archive stores map data along with metadata describing the sample, sample preparation, image acquisition and processing pipeline. The EMDB XSD data model defines the experimental metadata used to describe entries in the EMDB archive. It also provides a structured framework that outlines relationships between metadata elements and specifies which metadata items are required based on the type of experiment associated with an entry and EMDB policies. The following sections detail the components of the EMDB data model and how they contribute to the description of EMDB entries.

Map Data Format

The EMDB accepts and distributes 3DEM maps/volumes, strictly Coulomb potential maps, derived from several types of 3DEM experimental approaches and/or reconstruction methods. The EMDB distributes maps as MRC files with the extension *.map. These MRC files also follow the CCP4 definition (CCP4 map format), where voxel positions are limited to a grid that includes the Cartesian coordinate origin (0,0,0). The MRC format and header is maintained by CCP-EM (https://www.ccpem.ac.uk/mrc-format/) who the EMDB work with to maintain interoperability of map data stored in the EMDB.

EMDB Data Model: Overview

The EMDB archive holds 3DEM volumes of macromolecular complexes and subcellular structures. Alongside these volumes, the archive provides metadata files containing information on the experimental sample, data collection, processing details, etc.

EMDB Data Model: Data in | Data out

With regard to providing metadata to the EMDB and retrieving metadata from the EMDB, it can be helpful to distinguish between these processes. Curated data is provided to the EMDB by OneDep processes managed and developed by the wwPDB, broadly referred to as ‘data in’. EMDB processes and provides these data as a structured archive to the global community and our wwPDB partners, broadly referred to as ‘data out’. Metadata for each EMDB entry is collected by the OneDep deposition system, either through manual input by the depositor or by automatic extraction from an uploaded mmCIF file. Thus, depositors or developers wishing to provide metadata to the wwPDB OneDep deposition system should do so using the PDBx/mmCIF data model standard.

For accessing EMDB entry metadata, these data are available in two formats: XML and PDBx/mmCIF. The EMDB distributes metadata in both XML and PDBx/mmCIF formats to ensure the continued operations of partnered resources and community processes dependent on either source of information. The EMDB archive has provided XML metadata files since its inception. The archive also serves metadata in PDBx/mmCIF files where these are generated by the wwPDB OneDep deposition system through which EMDB receives deposition data. 

EMDB Data Model: Principles

The continuous advancements in cryoEM and cryoET highlight the dynamic nature of the field and its rapid pace of change. The EMDB and wwPDB continually assesses how the metadata data model supports the representation of 3DEM experiments while ensuring its long-term sustainability and compatibility with the already existing EMDB and PDB archives. To guide these developments, the EMDB engages with the EM community and wwPDB to make changes to its data model.

Currently, the EMDB data model comprises over several hundred distinct metadata elements used to describe an EMDB entry. Any addition of new metadata items or sets of items are versioned, with update details made publicly available. The following subsections outline the structure and update mechanism for both the XML and PDBx/mmCIF data models.

EMDB | wwPDB Data Model: PDBx/mmCIF

The Protein Data Bank Exchange macromolecular Crystallographic Information Framework (PDBx/mmCIF) is the metadata standard of the wwPDB able to store model coordinates and entry metadata. This format is defined by the rules outlined in the wwPDB PDB Exchange Data Dictionary (PDBx/mmCIF Data Model), ensuring consistency in data representation. PDBx/mmCIF provides the standard supporting OneDep depositions and annotations, as well as the archiving of wwPDB data, including EMDB. The associated dictionary file, named mmcif_pdbx.dic, currently used by wwPDB archives has the latest version being 5.xxx.

Metadata within PDBx/mmCIF-formatted files is structured into namespaces, each containing various categories. The primary namespace for metadata related to deposition and annotation in the EMDB archive is the "em namespace." Categories within this namespace begin with the prefix "em" (e.g., em_map, which stores metadata for EM maps). Metadata items are organized into category items of different data types, all formatted using the ASCII character set.

In a PDBx/mmCIF file, metadata follow a specific naming convention:

  • Each item name begins with an underscore (_).
  • It consists of a category name, followed by a period, and then an attribute name (e.g., _em_map.resolution).

This structured approach ensures a standardised representation of metadata, enabling seamless data exchange and archival across the wwPDB ecosystem. It also aligns information sources for an EMDB map-model entry, where map data is stored in the EMDB archive and the corresponding model is housed in the PDB. PDBx/mmCIF files have an extension “.cif”.

For a released entry EMD-xxxxx, its PDBx/mmCIF metadata file is stored in a subdirectory named EMD-xxxxx within the EMDB repository's released entries directory, following this path format:

<EMDB released entries>/EMD-xxxxx/metadata/emd-xxxxx.cif

EMDB Data Model: XML

Download schema
Browse schema documentation
Download Python code to facilitate reading and writing XML header files

The structure and validation rules for EMDB XML metadata files are defined in a version 1.0 XML Schema Definition (XSD) file. XSD schema versions are assigned based on changes to the data model, following the policy outlined here. To date, there have been three major updates to the EMDB XML data model. 

The most recent major update, version 3, introduced in 2018, replaced the long-standing stable version v1.9.6. All v3.x versions are backward compatible, allowing metadata files from earlier v3.x versions to be processed by newer ones. Version 2.x was used internally during OneDep development. All available EMDB XML data models can be found in the EMDB schema section of the EMDB repository. The model XSD file name is emdb.xsd and the latest version is always located here.

An interactive visual representation of the latest emdb.xsd file is available in the Schema Documentation section of the EMDB repository. This visualisation illustrates how the schema structures EMDB entry metadata into seven distinct groupings (XML elements):

  1. admin (holds housekeeping information about an entry such as a depositor provided title, relevant dates, authors and compiled revision history)
  2. crossreferences (lists publications and other entries linked to the entry)
  3. sample (provides list of entities contained in the entry)
  4. structure_determination_list (collates information of all experimental setups used in preparing, processing and determination of the deposited entry files)
  5. map (metadata pertaining to the entry’s primary map)
  6. interpretation (groups all elements used in interpreting the entry, such as an initial model, half maps, etc)
  7. validation (place for referencing files and methodologies used in validating the entry)

For a released entry EMD-xxxxx, its XML metadata file is stored in a subdirectory named EMD-xxxxx within the EMDB repository's released entries directory, following this path format:

<EMDB released entries>/EMD-xxxxx/header/emd-xxxxx.xml