VMD-L Mailing List

From: Josh Vermaas (vermaas2_at_illinois.edu)
Date: Mon Sep 22 2014 - 10:58:56 CDT

Hi Sebastian,

I think the truthful answer to #1 is that noone sat down and thought
"can we do this better?" and worked it out why the current
implementation is non-optimal (I mean, in principle you only need one
rotation to minimize volume. Figuring out what this rotation is is the
hard part). There is also some danger of making your box too small and
introducing periodic cell artifacts if the system rotates so that it
spans this new, smaller box.

For #2, the most obvious case to me would be systems where some part of
the system was aligned intentionally with an axis. In membrane
simulations, for example, the z-axis is normally aligned perpendicular
to the membrane so that membrane depths are simple to find. For a
soluble protein, it may be more convenient to calculate rotational
diffusion if the three principle axes were aligned with x, y, and z
(even if this makes your box bigger) if you want to precisely determine
how your protein differs from the diffusion of a sphere or ellipsoid.

-Josh Vermaas
On 09/22/2014 09:59 AM, Sebastian Stolzenberg wrote:
> Dear All,
>
> 1.)
> In the Solvate Plugin, I came across the
> ::Solvate::rotate_save_water
> procedure, which rotates a given system along the z and x axes to find
> the orientation with minimum surrounding cuboid volume V0.
> I was wondering why this rotation is not performed along all three
> cartesian axes to allow the algorithm to find even smaller values of V0.
>
> As an example, I am attaching a tcl script I wrote to execute this task
> for the "2D8R" PDB structure:
> As a result, I obtain a 15% reduced volume compared to the one obtained
> from ::Solvate::rotate_save_water.
> When I further solvate this system with a 20A water buffer layer (not
> shown), the total number of atoms is reduced by 29% compared to the
> solvated system produced by the solvate plugin.
>
> 2.)
> At http://www.ks.uiuc.edu/Research/vmd/plugins/solvate/
> I read:
> "Because the solvation box must be oriented along the cardinal axes,
> this can result in significant system size reductions, but should not be
> used if the initial orientation was chosen to facilitate analysis"
> I don't quite understand the last half sentence, what kind of situations
> does it specifically refer to?
>
> Thank you very much, Best,
> Sebastian
>