[gmx-developers] PosRes development: Include cylindrical constraints?

[Jochen Hub]

David van der Spoel wrote:
> Jochen Hub wrote:
>> p.s.: Would you agree if I add a little comment on that issue to the 
>> respective wiki-page?
>>
> Sure.
>
> You're also welcome to provide a reference implementation.

Well, I need to finish my PhD first :-), but then I should find some 
time to indeed implement a new and more flexible posres routine.

I would suggest to add 3 new mdp option, such as posres_grps, posres_r0 
and posres_direction. In posres_grps one could define groups which are 
subject to some flat-bottom posres potential. The width of the flat 
bottom would be defined in posres_r0. posres_disrection could allow 
options such as xyz, xy, x, y, or z which would refer to:

xyz : sphere-like flat-bottom potential
xy: cylinder-like flat-bottom potential
x, y, or z, layer-like flat-bottom potential in x-,y-, or z-direction

Such an implementation would be quite straight-forward, I guess. To 
allow posres_r0 to change within an FEP run one could allow to give two 
posres_r0 values per posres-group.

Any suggestions? Agreement?

Cheers,
Jochen


>
>> Jochen
>>
>> Jochen Hub wrote:
>>> Dear developers,
>>>
>>> I read about your plans to implement posisiton restraints with a 
>>> flat-bottomed potential:
>>>
>>> http://wiki.gromacs.org/index.php/Position_Restraints_Development
>>>
>>> It may be of interest to implement this potential in way such that 
>>> it allows to employ a cylinder-like potential:
>>>
>>> V(x,y,z) = 0.5*k*( \sqrt(x^2+y^2) - r0)^2 * H( \sqrt(x^2+y^2) - r0)
>>>
>>> where H is the heaviside step funktion and r0 the radius of the 
>>> cylinder. This would be a kind of flat-bottomed potential, which, 
>>> however, only applies in the x-y-plane. Such potentials are useful 
>>> for computing PMFs for permeation through membrane channels since 
>>> they allow to have a well-defined area per channel (or channel 
>>> density in the membrane).
>>>
>>> So, this may be something to keep in mind when extending the posres 
>>> code.
>>>
>>> Cheers from Goettingen,
>>> Jochen
>>>
>>
>>
>
>


-- 
************************************************
Jochen Hub
Max Planck Institute for Biophysical Chemistry
Computational biomolecular dynamics group
Am Fassberg 11
D-37077 Goettingen, Germany
Email: jhub[at]gwdg.de
************************************************