2024-07-17 : 37051 EMDB map entries, 21507 PDB coordinate entries RCSB PDB | PDBe

Twelve golden rules for deposition of EM data in EMDB and PDB

Every now and then depositors of EM data approach us with questions (or gripes) about the deposition process, in particular when both a map and a model are to be deposited. We appreciate that the process at present is not ideal, since there are two separate deposition systems to deal with (EMdep for EM maps and AutoDep or Adit for atomic coordinate models). The wwPDB and Unified Data Resource partners are currently developing a new joint Deposition and Annotation system that will be used by all deposition sites and with which you will be able to deposit maps and models at the same time. We expect this new system to go live next year. To make deposition a little bit easier in the meantime, we have formulated a few “golden rules.”

Here is what you need to do to deposit EM volume data (incl. tomographic reconstructions):

  • 1.   Deposit the final reconstruction(s) in EMDB.

    We do not archive raw data or intermediate steps of your structure determination and it is only the final reconstructions and some information describing the experiment that need to be uploaded or entered via the EMdep deposition system (UK site: pdbe.org/emdep, US site: emdep.rutgers.edu).

  • 2.   Your map must be in MRC, CCP4 or Spider format; it may optionally be compressed.

    To upload a compressed map, use zip, gzip or bzip2 compression schemes. All uploaded maps will be converted by us and archived and distributed as CCP4 files. Files uploaded in MRC format need to follow a strict interpretation of the file-format description. Note that many programs generate “.mrc” files that are modifications or extensions of the MRC format. In the best-case scenario, the deposition system will ignore the additional information, but some modifications may lead to errors during the reading of the map-header parameters, and in the worst case they will crash the deposition system.

  • 3.   Please check the converted map.

    Once your map has been converted into CCP4 format, EMdep will now allow you to download the converted map. Please do this and check the map visually, check the density range, and check that the map overlaps with any fitted models you are planning to deposit in the PDB.

  • 4.   You can deposit only one map per session.

    The deposition session assigns one accession code (EMD-nnnn) to one map. Therefore, even if your experiment resulted in multiple maps (e.g., heterogeneous data separated into multiple reconstructions), each one has to be deposited in a separate session. To make deposition easier, use the option “Based on previous submission” in EMdep - this will populate many of the fields in the deposition form based on the previous submission.

  • 5.   The voxel spacing used in the map must be specified.

    The voxel spacing (Ångstroms per pixel in X, Y, and Z) is crucial for determining the physical dimensions of the map. The numbers entered in EMdep will be used to rescale the map!

  • 6.   The contour level of the map must be specified, except for tomographic reconstructions.

    The contour level is important for the interpretation of the map. It defines the surface of the structure, and calculation of volume and molecular mass are dependent on it.

  • Once you have uploaded a map, you have the option to deposit a “fitted model” to the PDB. Such a model can be the result of a rigid-body docking of an existing PDB entry, or it may be the result of docking plus refinement (e.g., flexible fitting), or it may have been determined de novo using the EM map alone. Please note that the two EMDB deposition sites use the same software for EM depositions (EMdep), but different software for PDB depositions - the European site (PDBe at EBI) uses AutoDep and the US site (RCSB PDB) uses Adit. Here is what you need to do:

  • 7.   Always deposit the EM map first (in EMDB) and then any fitted models (in the PDB).

    Always start by depositing the EM map as described above. Towards the end of your deposition, you will get the option to deposit your fitted models to the PDB. This option will remain active even after you have submitted your EM map.

  • 8.   Deposit your fitted models in the PDB.

    All fitted models may be deposited, even if they are the result of rigid-body transformation of an existing PDB model. If the map doesn’t support atomistic interpretations because of lack of resolution, consider omitting water molecules, small ligands, ions, and side chains, or providing a C-alpha trace (for proteins) or a phosphorus trace (for polynucleotides).

  • 9.   Model coordinates must be in the same reference frame as the EMDB map.

    Your model must fit the map that will be archived in EMDB, i.e., the result of the conversion to CCP4 format by the deposition process. See golden rule #3.

  • 10.   For symmetric models the atomic coordinates should only include the asymmetric unit.

    When you deposit your model you can provide the transformation matrices that are required to generate the complete assembly.

  • 11.   The correct model sequence must be specified.

    This can be confusing! In cases where the sequence of the fitted model differs from that of the EM map, it is the former that needs to be specified during the deposition of the fitted model. However, the genus/species description entered in both the EMDB and PDB depositions should be that of the sample on which the EM data was collected.

  • 12.   Your model must make chemical and physical sense.

    If the atomic coordinates represent a composite model (made of two or more separate models) make sure they do not clash or overlap with each other. Flexible fitting may introduce incorrect chirality into the model. All such errors or problems should be corrected prior to deposition. There are standard software packages available to carry out these kinds of geometric checks (e.g., MolProbity and WhatCheck).

You can find more information about EM deposition under “Additional Links” on the deposition page. For specific questions, please feel free to contact us at: help@emdataresource.org

Prior Support from
National Institutes of Health
National Institute of General Medical Sciences

© 2007-2024 EMDataResource