
Hi Charity, The ChimeraX AlphaFold prediction is working again thanks to Colabfold developer Milot Mirdita who updated Colabfold to work with numpy 2. It will work in your current ChimeraX -- the fix was done in a Colab notebook that ChimeraX fetches each time you run a prediction, so no need to get a new ChimeraX. Tom
On Mar 17, 2025, at 7:20 PM, Tom Goddard via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Hi Charity,
I've disabled the ChimeraX AlphaFold prediction service. It will now display a message saying it is broken and give links to the bug reports so people can see whether any progress is made to fix it. This prevents trying the "Use fallback runtime version" fix. I figure so few ChimeraX users will know about that fix it is better to provide a clear message that the service is broken and not likely to be fixed for a while (weeks?).
Tom
On Mar 17, 2025, at 4:20 PM, Tom Goddard via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Hi Charity,
I have not found a way to fix the ChimeraX AlphaFold predictions yet. Here is a way to work around the problem caused by Google Colab updating from numpy version 1 to 2. Colab allows you to use a fallback runtime that has numpy 1.
When you run an AlphaFold prediction the Google Colab "AlphaFold Run" panel warns that the code was not written by Google and asks if you want to "Run anyways". Press "Cancel". Next press the "Connect GPU" text on the upper right part of the AlphaFold Run window -- this starts a Colab runtime and is necessary for the next step. Now press the "Commands" text on the upper left part of the AlphaFold Run window which will show a menu, scroll to the bottom of that long menu and select "Use fallback runtime version". It will ask you to confirm you want to use the fallback version. Then you can press the Predict button again in ChimeraX and it will now run correctly. Once your Google Colab session ends you will have to do these steps again to "Use fallback runtime version".
The Colabfold project has a Google Colab notebook that also runs predictions and it is also broken and I have posted a bug report to that project. Our best hope of fixing this in ChimeraX is that the ColabFold project figures out how to run with numpy 2.
https://github.com/sokrypton/ColabFold/issues/693
Tom
On Mar 17, 2025, at 1:21 PM, Tom Goddard via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Hi Charity,
Thanks for reporting the problem. Google Colab updated from the Python numpy library from version 1 to version 2 which has broken Colabfold that ChimeraX is trying to run. Colabfold reinstalls numpy version 1 over the default numpy 2, but Python does not get restarted and so it continues to use numpy 2 and fails with the error you saw. I am looking into how to fix this.
Tom
On Mar 17, 2025, at 12:46 PM, Charity Conlin via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Hello,
I am trying to use ChimeraX to predict the interaction of 2 proteins using alphafold: Beta-Lactoglobulin and Lactoferrin. I have plugged the 2 protein sequences into the alphafold tool to try to predict the interaction, but I get a failure code in google colab. I am wondering how I can get around this? I think it is because the wrong versions of numpy and colab are installed during the program.
Thanks,
Charity
<image.png> _______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu <mailto:chimerax-users@cgl.ucsf.edu> To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu <mailto:chimerax-users-leave@cgl.ucsf.edu> Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
_______________________________________________ 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/
_______________________________________________ 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/
_______________________________________________ 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/