[gmx-users] R: gmx-users Digest, Vol 86, Issue 183

[Mark Abraham]

On 28/06/2011 8:09 PM, Anna Marabotti wrote:
> Dear Tsjerk, dear all,
> thank you very much for suggestion. One of the main reasons why I never used
> the rhombic dodecahedron box for my simulations is the fact that it is
> really hard to visualize the results with programs such as VMD or Pymol, and
> therefore I cannot check easily during the setting of my system if the work
> proceeds properly. Since I would like to redo my simulations following your
> suggestion, could you please give me some hints on how to proceed in order
> to visualize easily the system?

In my experience, "trjconv -pbc mol -ur compact" is the way to achieve this.

>   For example, now I created a dodecahedric
> box filled with water and ions. I tried to visualize it with VMD and I saw
> it as a rectangular box with the protein at one edge. I tried to use the
> command trjconv:
>
> trjconv -f prot_dod_box_neu.gro -s prot_dod_box_neu.tpr -o
> prot_dod_box_neu2.gro -pbc mol -ur compact
>
> and I obtained an error message stating:
> Fatal error:
> Index[84328] 84329 is larger than the number of atoms in the trajectory file
> (84328)

The contents of the -f and -s files have to match meaningfully. Yours don't.

I'm surprised this error message is as cryptic as it is. I'll make it 
less so.

> I don't find this error neither in the gmx-users list archive (indeed it's
> becoming harder and harder to search into this archive, given that each
> message is visualized 10 times, as I already noticed sometimes ago), nor in
> the Error section. The first .gro file is the one coming from genion, and
> the .tpr file is the one I used to neutralize the system, which is the .tpr
> file I should give to trjconv?

For "-pbc mol", trjconv needs to be told the definition of the molecules 
in the -f file. So you need to supply a .tpr that corresponds to it. 
Generating ions is a tricky case, because the output files do not match 
the input .tpr. Hence its scarcity in the archives. However, since 
you'll shortly be generating a .tpr file for the next stage, it's not a 
real problem for visualization.

>   You see, if every time I try to use a box
> different from the standard cubic one I has to deal with such maybe trivial
> (for you) problems, but difficult and long to solve, I'm quite discouraged
> to use it!

The only problem with the above is a small gap in understanding. I hope 
I've filled that now.

>   This is also a kind request for Gromacs Web site managers: could
> you please add a section for this (recurrent, I suspect) problem of box
> visualization in the "How-to" section? it would be very useful for users, I
> think, and probably also for you (less trivial questions posted to the
> gmx-users list...)

The web page is a wiki - anyone with useful contributions is welcome to 
make them, but registering an account for editing does require 
registration. People who've experienced and solved a problem are often 
well placed to document management strategies :-) Others will come along 
and flesh out details, etc.

Mark