[gmx-users] Problem: Protein move out of box

[Mark Abraham]

Pan Wu wrote:
> Hi Gromacs users,
>     I am new to Gromacs, and now setting up one very simple system to 
> test Gromacs+Martini. (one protein with solvent surrounded)
>     When I run the MD and saw the trajectory, the protein seems go 
> around randomly, and across the box size. 

Whether that's reasonable depends how much time is elapsing between the 
frames. You may also wish to use trjconv to center each frame on some 
suitable group - e.g. your protein. See trjconv -h and play around with 
the options. mdrun doesn't bother making the output look "nice" since it 
doesn't know what you think "nice" is. :-) It will only write molecules 
that are "broken" across PBC only under some awkward algorithm 
combination, however.

> Because Periodic Boundary 
> Condition, it will come back from the other side. But I don't know 
> *whether the across box size movement will actually cause some problem 
> about protein structure or dynamics*?
> 
>     In my mdout.mdp file, I can see:
>         comm-mode = Linear
>         nstcomm = 1
>         comm-grps  =                      (default, should be system)

Should be fine.

Mark