Avoiding ratelimits by searching against ColabFold DBs locally

[IlyesAbdelhamid]

Hello,

I've been running a lot of simulations with ColabFold and I noticed that a RATELIMIT had been encountered for some of them. To avoid this issue, I then decided to build ColabFold databases (setup_databases.sh) and search against them locally. Is it possible to pass the path of ColabFold databases to colabfold_batch command and stop requesting the ColabFold server's API in order to process MMSeq2 search? Or should I change my workflow and run first colabfold_search then colabfold_batch ?

Thank you in advance for your reply!

Sincerely, Ilyes

[cclough]

ilyes - I came up against the same issue with Ratelimits. Taking the databases local seems like a good idea. Did you find a workaround in the end?

[dbolser]

Check here fro a solution: https://github.com/sokrypton/ColabFold/issues/563