Computed Structure Models from AI/ML and PDB Experimental Structures
Download high quality image | Credit: Maria Voigt/RCSB PDB
Researchers have long used the wealth of experimentally-determined structures available from the open-access PDB archive to understand basic principles of protein architecture. More recently, PDB structures have been used as training data for Artificial Intelligence/Machine Learning software tools (e.g., AlphaFold, RosettaFold, OpenFold) that are able to predict three-dimensional structures of proteins and design proteins with novel shapes and biochemical functions.
RCSB.org provides access to these Computed Structure Models from AI/ML (as symbolized on the computer screen in this image) alongside experimental PDB structures (symbolized in flasks).
Experimental structures and CSMs are clearly identified throughout RCSB.org in a similar manner: a dark-blue flask icon is used for PDB structures and a cyan computer icon for CSMs. Simultaneous delivery of PDB data and CSMs provides access to 3D structural information from across the human proteome, model organisms, and selected pathogens.
Related Resources
Browse PDB-101 resources related to Protein Structure Prediction, Design, and Computed Structure Models
