Hi Tom Many thanks for getting back. Really appreciate it. Understood. I have just re-started, colabfold indeed works again. Great! ChimeraX is absolutely amazing. Thank-you!! All the best Arthur Arthur Kaser FMedSci Department of Medicine University of Cambridge M +44 7412 569153 This email (together with any files transmitted with it) is intended only for the use of the individual(s) to whom it is addressed. It may contain information which is confidential and/or legally privileged. If you have received this email in error, please notify the sender by return email (or telephone) and delete the original message. The sender has taken reasonable precautions to check for viruses but the recipient opens this message at his or her own risk. From: Tom Goddard <goddard@sonic.net> Date: Monday, 22 December 2025 at 18:31 To: Arthur Kaser <ak729@cam.ac.uk> Cc: ChimeraX Users Help <chimerax-users@cgl.ucsf.edu> Subject: Re: [chimerax-users] ChimeraX Boltz prediction running eternally or failing submission to MSA server Hi Arthur, As mentioned in the previous discussion of this we have no control over the api.colabfold.com server https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/... Below is the previous message about this. I just tried Boltz and the Colabfold server is responding now. Unfortunately there is no substitute for that server. You cannot easily set it up yourself as it is running on a machine with 2 Tbytes of memory and I think 128 cores in order to be able to deliver MSAs in about 30-60 seconds each. A few months ago I experimented with running Colabfold MSA calculations on a Linux machine with just 64 GB of memory and an Nvidia 4090 GPU https://www.rbvi.ucsf.edu/chimerax/data/local-msa-sep2025/pp_screening.html It worked, and running 500 MSA calculations in one batch can be done reasonably fast (~30 seconds each). But running individual MSAs like the Colabfold server using GPU mmseqs2 took about 5 minutes for each MSA, and requires large databases (~ 1 TB) on an NVMe drive and GPU mmseqs2 to get that speed. It is not easy to set up. Tom Previous mailing list message about Colabfold MSA server not responding follows: t sounds to me like your Boltz prediction not running is because the Colabfold MSA server is not responding. Here is an issue on the Boltz github site from 12 hours ago reporting this problem https://github.com/jwohlwend/boltz/issues/629 When you run a Boltz prediction from ChimeraX it shows in the Boltz panel the progress it is making. For instance it will say "Prediction runing 14 seconds: sequence search submitted" "Prediction running 16 seconds: sequence search running" "Prediction running 22 seconds: structure inference" ... So if it is not finishing you should easily be able to see what step it is stuck on. My guess is you don't get past "sequence search submitted". But I just ran two Boltz predictions and the sequence searches completed in under a minute. So maybe the Colabfold server is back to working. We do not run that server." On Dec 21, 2025, at 6:10 AM, Arthur Kaser via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote: Same here. Boltz2 appears to stall at api.colabfold.com<http://api.colabfold.com/>, regardless of size/ length of the input sequence. Nil return at this step. _______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/