PDBImages: A new, open-source macromolecular structure imaging pipeline

Title image with structure in background, announcing the new PDBImages pipeline

The PDBe team, in collaboration with David Sehnal (NCBR, Masaryk University, Czech Republic), has developed a new pipeline for generating images of macromolecular structures. The new tool, PDBImages, takes advantage of Mol*(), the versatile, open-source software for interactive molecular visualisation platform, to provide more consistent and scalable visualisation for a range of different macromolecular views.

One significant benefit so the PDBImages pipeline is the creation of Mol* state files (.molj format) along with the generated images. These state files contain all the relevant information to recreate the exact visualisation that is displayed in the image within the Mol* viewer. These Mol* state files are generated by the PDBImages pipeline and can be directly uploaded within the Mol* viewer, allowing interactive exploration of the structure in the context of the displayed image.

 

Image of 7y7a structure, the largest in the PDB, generated by PDBImages
Image of the complete structure of PDB entry 7y7a, which is the largest structure in the PDB archive. The new PDBImages software has no problem generating visualisations of such huge complexes.

 

The PDBImages tool offers significant advancements in functionality. It reads atomic XYZ coordinates from PDBx/mmCIF or binaryCIF files and constructs predefined views of the macromolecular structures. For each of these views it then outputs the following files: 

  • Image files (PNG format)
  • Mol* state files - these allow replication of the view within the Mol* viewer
  • Metadata files (JSON format)

 

Image of PW9 ligand in PDB entry 7ov7 generated by PDBImages software
Ligand environment image for a Keggin molecule (CCD ID PW9) in PDB entry 7ov7. This new image generated through the PDBImages software displays the interactions within complex ligands with impressive clarity.

 

In total, the software generates nine distinct types of images, with eight of these created for PDB structures and displayed on various PDBe webpages and a ninth used to generate images for AlphaFold structures. Users can customise the type of image and view and even choose whether to render axis arrows. This high level of flexibility is complemented by the ability to handle input in several formats, including PDBx/mmCIF and binaryCIF, with the tool also able to fetch additional data from PDBe API endpoints.

 

This video shows the loading of a Mol* state file from the PDBImages pipeline. This example highlights entity 2 from PDB entry 6lu8 and allows visualisation of this context directly within Mol*, including the possibility to toggle visualisation of each entity within the viewer.

 

PDBImages leverages a 'headless' version of Mol*, enabling it to render images directly from the command line without a graphical interface. The innovative pipeline is open-source and has a modern codebase that includes comprehensive documentation and tests, making it user-friendly and easily maintainable.

If you are interested in using the PDBImages tool, it has been released as an , with the code available through the (Apache 2.0 license).

 

Schematic of PDBImages process at PDBe
Schematic image of the implementation of the PDBImages pipeline, highlighting the inputs, processes and outputs of the software.

 

In a world increasingly reliant on complex molecular visualisations, this new, open source data pipeline offers a modern, scalable, and user-friendly tool for the scientific community.