Bitbol-Lab / ProtMamba-ssm

ProtMamba: a homology-aware but alignment-free protein state space model
https://www.biorxiv.org/content/10.1101/2024.05.24.595730v1
Apache License 2.0
44 stars 7 forks source link

ColabFold MSA Reproduction for ProteinGym Benchmark #8

Closed smdrnks closed 2 weeks ago

smdrnks commented 1 month ago

Hi!

I am trying to reproduce the results for the ProteinGym benchmark. For that I need to create the target MSAs using ColabFold. Could you say a bit more about which settings were used for the ColabFold MSA generation? Your help would be greatly appreciated. Thank you!

smdrnks commented 1 month ago

Alternatively, would it be possible to provide the alignment files directly somehow?

CyrilMa commented 1 month ago

Hi @smdrnks,

Thank you for reaching out.

You can get all alignments from here

Alternatively, I added here the code that we used to make API calls to ColabFold (https://github.com/sokrypton/ColabFold/tree/main) adapted from their code

Keep in mind that it's their API so please cite them and adapt to their conditions :)

Is it okay to use the MMseqs2 MSA server (cf.run_mmseqs2) on a local computer?

You can access the server from a local computer if you queries are serial from a single IP. Please do not use multiple computers to query the server.

Cyril

smdrnks commented 1 month ago

Hi Cyril.

Amazing! thanks a lot for providing the MSAs so fast. that's really helpful.