plumed2
plumed2 copied to clipboard
plumed
Hydrogen bonds including angles
We should add the possibility to include a switching function on the angle to define a hydrogen bond. I do not know if there is a conventional way to do it.
We might also use the approach discussed here: http://arxiv.org/abs/1410.5447
Hey, is the HBOND_MATRIX the best way to set number of Hbonds as colvar in plumed? Do you know where I can find an example of input using this functions?
Maybe a continuous version of this criterion? http://mdtraj.org/1.6.2/api/generated/mdtraj.wernet_nilsson.html
This has the advantage of requiring only one threshold with respect to other angle/distance combinations.
but actually isn't HBOND_MATRIX doing what we want here @gtribello ? I think we can close this issue and maybe document this better and point to it from COORDINATION
Wow, I've never seen HBOND_MATRIX... That's great @gtribello !
Is it efficient? I would say it required three loops. Another perhaps missing feature it computing HB between A and B (as in COORDINATION with two groups).
Also: is it possible to hardcode sensible defaults for the switching functions?
Hello @GiovanniBussi . I don't know how efficient it is. The link cells are used at least. I did it as an experiment to see if I could get it to work. In general, I think the better thing to do for this sort of thing is to use the Michelino PAM thing. I also coded this but it is not in the master yet.
On Jun 19, 2017, at 15:30, Gareth Tribello [email protected] wrote:
Hello @GiovanniBussi . I don't know how efficient it is. The link cells are used at least. I did it as an experiment to see if I could get it to work. In general, I think the better thing to do for this sort of thing is to use the Michelino PAM thing.
Is this a criterion to define h-bonds? I wouldn’t mind having (also ? ) a more standard definition as the one Toni suggested.
Max
I also coded this but it is not in the master yet.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
Here is the paper: https://infoscience.epfl.ch/record/202770/files/text.pdf?version=1
This is nice as you use data learn what a hydrogen bond looks like in your system and then you use the criterion the machine learning algorithm comes up with. I hesitate to put in some defaults as I don't know enough about typical hydrogen bond geometries to be able to say what they should be.
Thanks! I will give it a look, but I will include also a more standard definition as Toni suggested
Max
On Jun 19, 2017, at 15:41, Gareth Tribello [email protected] wrote:
Here is the paper: https://infoscience.epfl.ch/record/202770/files/text.pdf?version=1
This is nice as you use data learn what a hydrogen bond looks like in your system and then you use the criterion the machine learning algorithm comes up with. I hesitate to put in some defaults as I don't know enough about typical hydrogen bond geometries to be able to say what they should be.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.
Maybe a continuous version of this criterion? http://mdtraj.org/1.6.2/api/generated/mdtraj.wernet_nilsson.html
This has the advantage of requiring only one threshold with respect to other angle/distance combinations.
based on
(1, 2) Wernet, Ph., L.G.M. Pettersson, and A. Nilsson, et al. “The Structure of the First Coordination Shell in Liquid Water.” (2004) Science 304, 995-999.
On Jun 19, 2017, at 15:43, Gareth Tribello [email protected] wrote:
Where or what was Toni's suggestion though?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.