[gmx-users] center of mass motion?

[MOLTENI at unisi.it]

Dear gmx users,
I am running MD on a peptide in water. Looking at a movie of (a part of)
my trajectory, I see that the molecule is often near the box boundary,
and therefore sometimes jumps in from the opposite side of the box (I
use PBC). I still have to have a look at the complete
trajectory.....but: is it normal for the peptide to be near the box
edge, when at the beginning I put it in the middle? Can it be due to
center of mass motion? Does it make sense? 
I checked my md.mdp file (which I modified from the ones in gromacs
examples) and indeed the keyword "comm_mode" doesn't appear in my file,
but "nstcomm" is = 1 .... does it mean that I'm removing center of mass
motion at every step... or never? (i.e., which is the default value for
"comm_mode"?) 
Moreover, the manual says that "nstcomm" should be used only for vacuum
simulations? Why? 

Thanks in advance

Elena