Dear ChimeraX team, I have a question regarding data privacy when using the Boltz-1 structure prediction tool within ChimeraX. Is the input or output data temporarily or permanently stored, logged, or shared in any way by UCSF, ColabFold, or any third-party services? Is there an option to ensure that all data remains strictly local and is not stored externally? Thank you, Dario
Maybe experts can give more detail, but as I understand it, using ChimeraX "boltz" installs boltz-1 locally on your own machine: <https://rbvi.ucsf.edu/chimerax/docs/user/tools/boltz.html> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/boltz.html> <https://www.rbvi.ucsf.edu/chimerax/data/boltz-apr2025/boltz_help.html> However, running boltz-1 will use Colabfold server for generating the deep multiple sequence alignment, and we do not have control over how/whether data is stored by that server. Other than that, I don't see anything in the help about data going in/out from your own computer. I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Resource for Biocomputing, Visualization, and Informatics Department of Pharmaceutical Chemistry University of California, San Francisco
On Jul 1, 2025, at 8:38 PM, Dario Passos via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear ChimeraX team,
I have a question regarding data privacy when using the Boltz-1 structure prediction tool within ChimeraX. Is the input or output data temporarily or permanently stored, logged, or shared in any way by UCSF, ColabFold, or any third-party services? Is there an option to ensure that all data remains strictly local and is not stored externally?
Thank you, Dario
Hi Dario, Pretty much all of the current tools use MMSeqs2 <https://github.com/soedinglab/mmseqs2> to do the multiple sequence alignment, and by default reach out to their server to do the job. As I understand it, it caches all alignment results for efficiency (so it can instantly return the cached result for any query identical to one it's seen before). While you can set up your own local server, the requirements are pretty steep - ideally hundreds of GB RAM and a few TB of storage. Best, Tristan On Wed, Jul 2, 2025 at 5:54 PM Elaine Meng via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Maybe experts can give more detail, but as I understand it, using ChimeraX "boltz" installs boltz-1 locally on your own machine:
<https://rbvi.ucsf.edu/chimerax/docs/user/tools/boltz.html> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/boltz.html> <https://www.rbvi.ucsf.edu/chimerax/data/boltz-apr2025/boltz_help.html>
However, running boltz-1 will use Colabfold server for generating the deep multiple sequence alignment, and we do not have control over how/whether data is stored by that server.
Other than that, I don't see anything in the help about data going in/out from your own computer.
I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Resource for Biocomputing, Visualization, and Informatics Department of Pharmaceutical Chemistry University of California, San Francisco
On Jul 1, 2025, at 8:38 PM, Dario Passos via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Dear ChimeraX team,
I have a question regarding data privacy when using the Boltz-1 structure prediction tool within ChimeraX. Is the input or output data temporarily or permanently stored, logged, or shared in any way by UCSF, ColabFold, or any third-party services? Is there an option to ensure that all data remains strictly local and is not stored externally?
Thank you, Dario
_______________________________________________ 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/
-- Altos Labs UK Limited | England | Company reg 13484917 Registered address: 3rd Floor 1 Ashley Road, Altrincham, Cheshire, United Kingdom, WA14 2DT
Hi Tristan and Elaine, Thank you for the information. Ideally, there would be an option to prevent input data from being stored remotely after a job completes, similar to how setting "useMsaCache" to false avoids caching. Kind regards, Dario On Wed, Jul 2, 2025 at 10:10 AM Tristan Croll <tcroll@altoslabs.com> wrote:
Hi Dario,
Pretty much all of the current tools use MMSeqs2 <https://github.com/soedinglab/mmseqs2> to do the multiple sequence alignment, and by default reach out to their server to do the job. As I understand it, it caches all alignment results for efficiency (so it can instantly return the cached result for any query identical to one it's seen before). While you can set up your own local server, the requirements are pretty steep - ideally hundreds of GB RAM and a few TB of storage.
Best, Tristan
On Wed, Jul 2, 2025 at 5:54 PM Elaine Meng via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Maybe experts can give more detail, but as I understand it, using ChimeraX "boltz" installs boltz-1 locally on your own machine:
<https://rbvi.ucsf.edu/chimerax/docs/user/tools/boltz.html> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/boltz.html> <https://www.rbvi.ucsf.edu/chimerax/data/boltz-apr2025/boltz_help.html>
However, running boltz-1 will use Colabfold server for generating the deep multiple sequence alignment, and we do not have control over how/whether data is stored by that server.
Other than that, I don't see anything in the help about data going in/out from your own computer.
I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Resource for Biocomputing, Visualization, and Informatics Department of Pharmaceutical Chemistry University of California, San Francisco
On Jul 1, 2025, at 8:38 PM, Dario Passos via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Dear ChimeraX team,
I have a question regarding data privacy when using the Boltz-1 structure prediction tool within ChimeraX. Is the input or output data temporarily or permanently stored, logged, or shared in any way by UCSF, ColabFold, or any third-party services? Is there an option to ensure that all data remains strictly local and is not stored externally?
Thank you, Dario
_______________________________________________ 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/
Altos Labs UK Limited | England | Company reg 13484917 Registered address: 3rd Floor 1 Ashley Road, Altrincham, Cheshire, United Kingdom, WA14 2DT
Hi Dario, The multiple sequence alignment (MSA) server used by Boltz is maintained by the Martin Steinegger lab at Seoul National University (https://steineggerlab.com/en/) and was developed for ColabFold (an optimized AlphaFold 2). They determine the privacy policy for the sequences sent to the server. We are considering setting up an MSA server at UC San Francisco that would be used by ChimeraX to make MSA calculation not dependent on the single Steinegger lab server. If we did that we could set our own privacy policy. Currently to run Boltz predictions without sending any data out you would have to setup your own MSA server which is a somewhat complex process described here https://github.com/sokrypton/ColabFold/tree/main/MsaServer Tom
On Jul 2, 2025, at 9:58 PM, Dario Passos via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Hi Tristan and Elaine,
Thank you for the information.
Ideally, there would be an option to prevent input data from being stored remotely after a job completes, similar to how setting "useMsaCache" to false avoids caching.
Kind regards, Dario
On Wed, Jul 2, 2025 at 10:10 AM Tristan Croll <tcroll@altoslabs.com <mailto:tcroll@altoslabs.com>> wrote:
Hi Dario,
Pretty much all of the current tools use MMSeqs2 <https://github.com/soedinglab/mmseqs2> to do the multiple sequence alignment, and by default reach out to their server to do the job. As I understand it, it caches all alignment results for efficiency (so it can instantly return the cached result for any query identical to one it's seen before). While you can set up your own local server, the requirements are pretty steep - ideally hundreds of GB RAM and a few TB of storage.
Best, Tristan
On Wed, Jul 2, 2025 at 5:54 PM Elaine Meng via ChimeraX-users <chimerax-users@cgl.ucsf.edu <mailto:chimerax-users@cgl.ucsf.edu>> wrote:
Maybe experts can give more detail, but as I understand it, using ChimeraX "boltz" installs boltz-1 locally on your own machine:
<https://rbvi.ucsf.edu/chimerax/docs/user/tools/boltz.html> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/boltz.html> <https://www.rbvi.ucsf.edu/chimerax/data/boltz-apr2025/boltz_help.html>
However, running boltz-1 will use Colabfold server for generating the deep multiple sequence alignment, and we do not have control over how/whether data is stored by that server.
Other than that, I don't see anything in the help about data going in/out from your own computer.
I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Resource for Biocomputing, Visualization, and Informatics Department of Pharmaceutical Chemistry University of California, San Francisco
On Jul 1, 2025, at 8:38 PM, Dario Passos via ChimeraX-users <chimerax-users@cgl.ucsf.edu <mailto:chimerax-users@cgl.ucsf.edu>> wrote:
Dear ChimeraX team,
I have a question regarding data privacy when using the Boltz-1 structure prediction tool within ChimeraX. Is the input or output data temporarily or permanently stored, logged, or shared in any way by UCSF, ColabFold, or any third-party services? Is there an option to ensure that all data remains strictly local and is not stored externally?
Thank you, Dario
_______________________________________________ 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/
Altos Labs UK Limited | England | Company reg 13484917 Registered address: 3rd Floor 1 Ashley Road, Altrincham, Cheshire, United Kingdom, WA14 2DT
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/
Hi Tom, Having the option to run confidential jobs through a UCSF-hosted server would be a great addition. It might also be worth considering a dual-option setup, allowing users to choose between the UCSF server for sensitive sequences and the existing SNU server for non sensitive cases. This could help balance privacy needs with computational resource demands. Thank you also for the link to set up a local ColabFold API server. Kind regards, Dario On Thu, Jul 3, 2025 at 1:39 PM Tom Goddard <goddard@sonic.net> wrote:
Hi Dario,
The multiple sequence alignment (MSA) server used by Boltz is maintained by the Martin Steinegger lab at Seoul National University ( https://steineggerlab.com/en/) and was developed for ColabFold (an optimized AlphaFold 2). They determine the privacy policy for the sequences sent to the server. We are considering setting up an MSA server at UC San Francisco that would be used by ChimeraX to make MSA calculation not dependent on the single Steinegger lab server. If we did that we could set our own privacy policy.
Currently to run Boltz predictions without sending any data out you would have to setup your own MSA server which is a somewhat complex process described here
https://github.com/sokrypton/ColabFold/tree/main/MsaServer
Tom
On Jul 2, 2025, at 9:58 PM, Dario Passos via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Hi Tristan and Elaine,
Thank you for the information.
Ideally, there would be an option to prevent input data from being stored remotely after a job completes, similar to how setting "useMsaCache" to false avoids caching.
Kind regards, Dario
On Wed, Jul 2, 2025 at 10:10 AM Tristan Croll <tcroll@altoslabs.com> wrote:
Hi Dario,
Pretty much all of the current tools use MMSeqs2 <https://github.com/soedinglab/mmseqs2> to do the multiple sequence alignment, and by default reach out to their server to do the job. As I understand it, it caches all alignment results for efficiency (so it can instantly return the cached result for any query identical to one it's seen before). While you can set up your own local server, the requirements are pretty steep - ideally hundreds of GB RAM and a few TB of storage.
Best, Tristan
On Wed, Jul 2, 2025 at 5:54 PM Elaine Meng via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Maybe experts can give more detail, but as I understand it, using ChimeraX "boltz" installs boltz-1 locally on your own machine:
<https://rbvi.ucsf.edu/chimerax/docs/user/tools/boltz.html> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/boltz.html> <https://www.rbvi.ucsf.edu/chimerax/data/boltz-apr2025/boltz_help.html>
However, running boltz-1 will use Colabfold server for generating the deep multiple sequence alignment, and we do not have control over how/whether data is stored by that server.
Other than that, I don't see anything in the help about data going in/out from your own computer.
I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Resource for Biocomputing, Visualization, and Informatics Department of Pharmaceutical Chemistry University of California, San Francisco
On Jul 1, 2025, at 8:38 PM, Dario Passos via ChimeraX-users < chimerax-users@cgl.ucsf.edu> wrote:
Dear ChimeraX team,
I have a question regarding data privacy when using the Boltz-1 structure prediction tool within ChimeraX. Is the input or output data temporarily or permanently stored, logged, or shared in any way by UCSF, ColabFold, or any third-party services? Is there an option to ensure that all data remains strictly local and is not stored externally?
Thank you, Dario
_______________________________________________ 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/
Altos Labs UK Limited | England | Company reg 13484917 Registered address: 3rd Floor 1 Ashley Road, Altrincham, Cheshire, United Kingdom, WA14 2DT
_______________________________________________ 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/
participants (4)
-
Dario Passos -
Elaine Meng -
Tom Goddard -
Tristan Croll