Hi Guenter,

   I see the boltz failure is only when you paste a sequence.  I see I made a change on October 6, 2025 to check if you specify the "protein" option that it only accepts protein chains (if for instance you specify a whole structure that include both protein and DNA).  And that change broke the code if you paste in a protein sequence.  I've fixed it, so the October 21, 2025 build will work.  Thanks for reporting the problem.

Tom




On Oct 20, 2025, at 12:28 PM, Guenter Fritz <guenter.fritz.phenix.ccp4@gmail.com> wrote:

Hi Tom,

thanks a lot for the prompt answer!

I have checked meanwhile also my macs. Same like on my Rocky8 machine, does not work if I cp paste a sequence there. But surprisingly it works if I use the sequence of a loaded structure, e.g. from  #1   =>  see below C.)

daily build versions and macs:

An old Intel macbook pro
version 1.11.dev202510022155 (2025-10-02)
Works (yes slow). I cp /paste a short sequence.
Command:
/Users/fritzg/boltz22/bin/boltz predict /Users/fritzg/Desktop/boltz_uq-test/uq-test.yaml --use_msa_server --accelerator cpu --no_kernels


An M1 macbook pro:
A.)
version 1.11.dev202509090102 (2025-09-09)
Works. I cp /paste a short sequence.
Command:
/Users/fritzg/boltz22/bin/boltz predict /Users/fritzg/Desktop/boltz_uq-test_20250909/uq-test_20250909.yaml --use_msa_server --accelerator gpu --no_kernels

B.)
version 1.11.dev202510170018 (2025-10-17)
Fails. I cp /paste a short sequence.
Command:
/Users/fritzg/boltz22/bin/boltz predict /Users/fritzg/Desktop/boltz_uq_test_dev20251017_seq/uq_test_dev20251017_seq.yaml --use_msa_server --accelerator gpu --no_kernels

Running boltz prediction failed with exit code 0:
command:
/Users/fritzg/boltz22/bin/boltz predict /Users/fritzg/Desktop/boltz_uq_test_dev20251017_seq/uq_test_dev20251017_seq.yaml --use_msa_server --accelerator gpu --no_kernels
stdout:
Boltz version 2.2.0
MSA server enabled: https://api.colabfold.com
MSA server authentication: no credentials provided
Checking input data.
Processing 1 inputs with 1 threads.
Failed to process /Users/fritzg/Desktop/boltz_uq_test_dev20251017_seq/uq_test_dev20251017_seq.yaml. Skipping. Error: 'NoneType' object is not iterable.

The yaml input file is just:
version: 1
sequences:


But it works if I use the sequence of a loaded structure:
C.)
version 1.11.dev202510170018 (2025-10-17)
Works. I use the sequence of mol #1
Command:
/Users/fritzg/boltz22/bin/boltz predict /Users/fritzg/Desktop/boltz_uq_test__20251017_from_struc/uq_test__20251017_from_struc.yaml --accelerator gpu --no_kernels

The yaml input file is:
version: 1
sequences:
  - protein:
      id: [A]
      sequence: QQRLIFAGKQLEDGRTLSDYNIQKESTLHLVLRLRGGIIEPSLRQLAQKYNCDKMICRKCYARLHPRAVNCRKKKCGHTNNLR
      msa: /Users/fritzg/Downloads/ChimeraX/BoltzMSA/uq-test_20250909/uq-test_20250909_0.csv





Many thanks and best regards,
Guenter


Hi Guenter,

  I have seen that error when you don't specify any sequences to predict. I'll fix the code so it doesn't let you do that and explains the problem. 

  What was your Boltz command shown in the ChimeraX Log?  In the Boltz user interface panel you have to press the Add button to add a sequence to the prediction. 

    Tom


On Oct 20, 2025, at 8:24 AM, Guenter Fritz via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:

Dear ChimeraX developers,

really a great job implementing boltz into chimeraX !!

I have just installed the latest daily build 2025 Oct 20 on a Redhat 8 (Rocky 8 derivative) machine.

Boltz-2 installs with no visible problem via chimeraX (like before). When I try to run predictions, I get an error which seems to be due to an almost empty yaml file created by chimeraX. Only the first line ("version 1.0") is in the yaml file. Boltz (of course) fails.

There should be no problem with permissions. I just installed also the latest release of chimeraX and boltz-1 via chimeraX. Runs as expected.

PS: I had no try yet on my mac (will check later).

Best wishes,

Guenter


_______________________________________________
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/