Progres server

Progres lets you quickly search protein structures against pre-embedded structural databases. This can be useful for remote homology detection, protein classification and inferring function from structure.

Due to the way it is trained, Progres works with protein domains rather than whole structures. However, queries can be automatically split into domains before searching.

If you want to search with many structures or embed databases to search against, please use the Progres software. If you use the Progres web server or software, please cite the paper:

Greener JG and Jamali K, Fast protein structure searching using structure graph embeddings, bioRxiv - link

Submit:

Job name:		Give the job a name (optional).
Target database:		Choose a database to search against. afted is the TED domains from the AlphaFold database. scope95/scope40/cath40/ecod70 are domains from classifications of the PDB. pdb100 is the whole PDB split into domains with Chainsaw. af21org is domains from the AlphaFold set of 21 model organisms. See here for more info.
Split domains:		Whether to split the query structure into domains with Chainsaw and search with each domain separately. Recommended for structures above 200-300 residues. Note that whether or not this option is selected, only the first chain in the file is considered.
Minimum similarity:		The Progres score above which to return hits. The default of 0.8 indicates the same fold. Must be 0 -> 1.
Max number of hits:		The maximum number of hits per domain to return. Must be 1 -> 1000.
File format:		By default the format of the uploaded file is guessed from the file extension, but it can be set explicitly here. Supported formats are PDB, mmCIF and MMTF.
File:

Or view an example results page.

If you find an issue with the web server, please open an issue or get in touch.