search

RedesignScience / cvpack

Useful Collective Variables for OpenMM
https://redesignscience.github.io/cvpack/latest
MIT License
11 stars 1 forks source link

Question about path CV #92

Closed gitkol closed 4 months ago

gitkol commented 5 months ago

Hi, I have been using path CV extensively with OpenMM-PLUMED, and got very excited to learn about your fantastic CV library that runs on the GPU. My understanding, though, is that your path CV defines the milestones in CV space. Do you have any suggestion how to use your library when the milestones are defined by intermediate structures between two conformations identified with their atomic coordinates as opposed to s and z, which in this case would just be a trivial sequence of (1,0), (2,0), etc.?

Thank you very much!

Best regards,

Istvan

craabreu commented 5 months ago

Thank you for the question and the feedback.

A new path-CV class is necessary to use the library in the way you mentioned. It has to accept reference structures as milestones and use RMSD to measure the distance between the current state and each milestone. It would be identical to PathInCVSpace in all other features, which makes such implementation relatively simple.

I started PR #93 to implement this new class. It might be available in the next release version.

gitkol commented 5 months ago

That would be fantastic, thank you very much! I'd be happy to test it if that helps.

Istvan

craabreu commented 5 months ago

It helps a lot! Thank you very much.

The latest CVPack version (v0.12.0) includes the new PathInRMSDSpace class. It is already available for installation via conda/mamba or pip.

gitkol commented 5 months ago

Amazing, thank you Charlles! I’ll install and try it with an example from my OpenMM-PLUMED runs.

Best,

István

On Sat, Apr 20, 2024 at 7:47 PM Charlles Abreu @.***> wrote:

It helps a lot! Thank you very much.

The latest CVPack version (v0.12.0) includes the new PathInRMSDSpace https://redesignscience.github.io/cvpack/latest/api/PathInRMSDSpace.html class. It is already available for installation via conda/mamba or pip.

— Reply to this email directly, view it on GitHub https://github.com/RedesignScience/cvpack/issues/92#issuecomment-2067816175, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKDNJ2QNNMMSACERS5I5YCLY6L5CJAVCNFSM6AAAAABGJVYY3GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANRXHAYTMMJXGU . You are receiving this because you authored the thread.Message ID: @.***>

craabreu commented 4 months ago

I'm closing this issue. Please feel free to either reopen it or create a new one.