Can i use foldseek online server run multiple monomer structure for multiple m8file (m8.) ?

[yang-nei]

there are hundreds of monomer i shall run, but i don't wanna copy-paste pdb. one by one for similarity protein structure, Is there is a better ways that can resolve my problem?

[milot-mirdita]

You can submit the individual monomers as separate chains or as separate models in a PDB file. This should work in theory, however the server has a 1-hour time-limit so it might crash before it processes all queries.

You can install foldseek locally and run the searches on your own machine if you need larger scale searches.

[yang-nei]

Thanks for your reply! may i ask another questions that once i wanna install foldseek on my local computer , is there something limitted on hardwares or softwaresm. Or something u recommand equipments

[martin-steinegger]

To use this optimally, you need a machine with memory (RAM) equal to the amount of residues * 8 bytes. However, you can still run it with less memory, but the search may become slower. For example, for the AFDB50 you would need approx: 8 byte * 54*10^6 (Seqs) * 350 (avg. protein length) = 151GB,

[yang-nei]

ok! how about RAM ,GPU ,CPU ? Is there any preference

[martin-steinegger]

For RAM please see my previous answer. All CPU should work for foldseek. I recommend having one that supports AVX2. No GPU is needed.

[yang-nei]

after struggling, I decide to download foldseek however once I decide to use uniprot all database, there must be 3TB SSD for my computers ,is it ? (alphafold2 copacities) And as above you toldme ,I can use any kinds of RAM and CPU to run foldseek and just need to wait for more time, isn't it?