[gmx-users] A periodic boundary trajectory with lots of objects: how to get rid of artefacts?

[Mark Abraham]

On 24/11/2010 2:48 AM, ms wrote:
> Dear GROMACS users,
>
> I am trying to visualize a trajectory with hundreds of peptides which 
> come and go out and in the faces of the periodic box. Problem is, when 
> they cross the periodic boundary, the visualization software (I tried 
> both VMD and PyMol) create visual artefacts by visualizing 
> nonsensically long bonds crossing the box -which I understand, since 
> the algorithms seem to have no clear notion of PBC.
>
> Using -pbc nojump or similar, given my experience with identical 
> systems only containing less peptides, is not going to work (peptides 
> are going to be "broken").
>
> Since all I want is some nice snasphot of the trajectory to show and 
> not analysis, I feel that a much easier solution would be to find a 
> setting in the visualization softwares that simply do not visualize 
> bonds longer than a threshold. However I can't find such a setting. 
> Does one exist? Does a "PBC mode" exist? What kind of solution do you 
> propose? I googled quite a bit but I couldn't find much.

If you only want a snapshot, then simply write a single chosen frame to 
a .gro or .pdb using trjconv. Load only that into VMD, and it won't 
generate PBC-crossing bonds.

If you want a movie, then there I seem to recall that there is a VMD 
option to generate bonds for each frame, and not to use those from the 
first frame. Ask their mailing list. Such bonds will not traverse the 
PBC, obviously. There is no general solution for bonds visualized on a 
single set of coordinates, however - over a trajectory, either molecules 
appear to diffuse out of the box, or appear to break.

Mark