VMD-L Mailing List

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Wed Aug 15 2007 - 20:36:13 CDT

Hi Cesar,
this is a good (and hard) question, but unfortunately there's no clear
answer. It really depends on exactly what you want to get out of your
calculation. To get a qualitative look at what's going on in the
interaction between the groups of interest, it should be sufficient to
just do the calculations between the solute components, perhaps with a
dielectric constant chosen based on the distance between them. If you
want to see what the overall electrostatic potential looks at a given
point during the trajectory, you'd probably want to use the PMEPot
plugin instead. If you really want to go for the true (screened)
electrostatic force between your two components, then I'd arrange things
as you had below, with selection1 being moleculeA and selection2 being
molecule B plus a solvent slice. This is just my gut feeling of roughly
what you'd want to do though; I haven't tried it myself, and certainly
don't have a theoretically rigorous description of what you're looking
at. You may want to go into some literature on solvent effects on this
length scale as you think of what to calculate. This, unfortunately, is
one of the downsides of explicit solvent models (there are, of course,
up sides as well ;-).

Best,
Peter

Cesar Luis Avila wrote:
> Hi Peter,
> thanks for your fast reply. So, it was as i suspected. How should I
> include the solvent in between the molecules on force calculation? It
> is difficult for me to figure it out. Should I include the solvent in
> one of the two selections? for example
> selection1: molecule A
> selection 2: molecule B + solvent slice?
>
> For sure, the number of atoms selected will change from frame to frame
> in this case. I suppose it will introduce an additional problem when
> trying to understand the results.
>
> Thanks again for your help
> Cesar
>
>
> Peter Freddolino escribió:
>> Hi Cesar,
>> this is a good (and important) question. When NAMD calculates pair
>> interactions, it *only* takes into account the atoms which are members
>> of the pair interaction selections, so the forces you obtain from it on
>> the interactions between two sections of solute should be regarded as
>> the forces in vacuum, and interpreted accordingly. You'd have to include
>> some section of solvent in your pair interaction calculations to get its
>> effect included.
>> Best,
>> Peter
>>
>> Cesar Luis Avila wrote:
>>
>>> Dear all,
>>> I want to calculate non bonded forces (VDW and ELECT) between two
>>> groups of molecules during the course of a simulation. For this I plan
>>> to use NAMD energy plugin, with the exact same parameters used for the
>>> simulation (cutoff, pme). This may be a silly question but, when the
>>> electrostatic forces are calculated, does it take into account the
>>> water in between the molecules? In other words, the forces between the
>>> molecules should be regarded as that on vacuum or in a solvated system?
>>> Regards
>>> Cesar
>>>
>>
>>