Explore Computed Structure Models Alongside PDB Data

08/31 

Search for all CSMs at RCSB.org | Access User Guide Documentation

In addition to nearly 200,000 experimentally-determined PDB structures, RCSB.org now offers access to ~1 million Computed Structure Models (CSMs) from AlphaFoldDB and RoseTTAFold (from Model Archive). Both software tools build on decades of methodological research on structure prediction and rely on open access to the immense number of sequences in genomic sequence databases and to the PDB archive. This release includes the pre-packaged collection of 999,255 AlphaFold models released on 01-Jun-2022 (based on model organism proteomes; global health proteomes; Swiss-Prot sequences; and MANE (Matched Annotation from NCBI and EMBL-EBI)) and 1,106 core eukaryotic protein complexes produced by RoseTTAFold and AlphaFold2 from the ModelArchive.

Additional CSMs were added in February 2023.

Only experimentally-determined PDB structures are included in search results by default. Move the "Include CSM" from gray- to cyan-colored slider to activate.

Exploring PDB Structures and CSMs at RCSB.org

Experimental structures and CSMs are clearly identified throughout the website: a dark-blue flask icon is used for PDB structures and a cyan computer icon for CSMs.

Left: flask icon indicates an experimental PDB structure; Right: computer icon indicates a CSM

A confidence score called “predicted local distance difference test” (pLDDT) is computed for each amino acid residue to estimate how well the method has converged (i.e., how well the predicted structure agrees with multiple sequence alignment data and PDB structure information). This pLDDT score is used to color 2D and 3D views of CSMs and to sort search results.

RCSB.org users can query, organize, visualize, analyze, and compare experimental structures and CSMs side-by-side:

• Search: User queries can be applied to all PDB structures and CSMs; PDB structures only; and can exclude either PDB structures or CSMs from the search results. Turn the slider button to cyan to include CSMs when searching the tab bar Basic Search or Advanced Search options; CSMs are not included in the search results by default.
• View and Organize Results: By default, search results are ordered using a query-based relevancy score. Results can be resorted using different criteria (e.g., listing experimental PDB structures first, per-residue confidence score (pLDDT)). The Refinements panel can be used to exclude PDB structures or CSMs from the results list.
• Explore Similar Proteins: "Group" summary pages and search results simplify exploration of PDB structures with similar sequence identity/UniProt ID or were deposited as part of the same study.
• Explore Individual Structures: Structure Summary Pages offer details of experimental PDB structures (e.g., 4HHB) and CSMs (e.g., AF_AFP44795F1).
• Assess quality: Analogous to the validation slider for experimental structures, all CSMs report global and local confidence levels as pLDDT scores.
• Visualize in 3D: View experimental PDB structures and CSMs in Mol* from Structure Summary Pages (e.g., AF_AFP44795F1). Use the standalone Mol* 3D Viewer to upload single or multiple data files, align structures, and run Structure Motif Search.
• Download: From Structure Summary Pages, download the ModelCIF data file hosted by the corresponding external archive (AlphaFoldDB or ModelArchive).

RCSB PDB will continue to develop resources to support exploration of experimental PDB structures and CSMs; feedback sent using the Contact Us button above or by email is greatly appreciated.

Register Now for Virtual Crash Course: Exploring Computed Structure Models from Artificial Intelligence/Machine Learning at RCSB.org

Learn how to search, visualize, and analyze CSMs alongside experimentally-determined PDB structures using all of the same RCSB PDB features that you know and love on Thursday September 22, 2022 (1:00 - 3:00 pm EDT).

This event has ended, but Videos from Crash Course: Exploring Computed Structure Models at RCSB.org are available.

Related Resources: Learn more about CSMs vs. Experimental PDB Structures

• RCSB.org Feature Documentation for CSMs
• PDB-101: Computed Structure Models

References

• AlphaFold: Highly accurate protein structure prediction with AlphaFold (2021) Nature 596: 583-589 doi:10.1038/s41586-021-03819-2
  AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models (2022) Nucleic Acids Res 50: D439-D444. doi: 10.1093/nar/gkab1061
• RoseTTAfold: Accurate prediction of protein structures and interactions using a three-track neural network (2021) Science 373: 871-876 doi: 10.1126/science.abj8754;
  Computed structures of core eukaryotic protein complexes (2021) Science 374: eabm4805 doi: 10.1126/science.abm4805
• pLDDT scoring: Highly accurate protein structure prediction for the human proteome (2021) Nature 596: 590–596 doi: 10.1038/s41586-021-03828-1
• Predicting Proteome-Scale Protein Structure with Artificial Intelligence (2021) N Engl J Med 385: 2191-2194 doi: 10.1056/NEJMcibr2113027
• Open-access data: A cornerstone for artificial intelligence approaches to protein structure prediction (2021) Structure 29: 515-520 doi: 10.1016/j.str.2021.04.010