[gmx-users] -center -fit dodecahedron : dimer

[Mark Abraham]

nahren manuel wrote:
> Dear Gromacs Users,
> 
>  
> 
> I am performing a MD simulation of a dimer in a dodecahedron box. The 
> simulation stopped after 8 ns (power cut) and i had to restart to 
> complete it fully to 12 ns.
> 
>  
> 
> I then concatenated the two trajectories using trjcat
> 
>  
> 
> trjconv -f promd.trr -s proem.tpr -pbc nojump -o nojump.xtc
> 
>  
> 
> trjconv -f nojump.xtc -s proem.tpr -pbc mol -ur compact -center 
> -boxcenter tric -o center.xtc
> 
>  
> 
> trjconv -f center.xtc -s proem.tpr -fit rot+trans -o fit.xtc
> 
>  
> 
>  
> 
> 1. The above procedures does not center the molecule in the box.

The last operation fits to a structure, permitting rotations and 
translations. If the target is not centered in its box...

> 2. The box seems to shift from one corner to the center. Especially for 
> the duration 8-12 ns (my restart run)
> 
> I feel I am missing something here. Kindly advice.

It's a periodic box - a mathematical construction. The simulation 
doesn't care where the atoms go in relation to it. If you care, you can 
apply trjconv afterwards.

Mark