[gmx-users] trjconv, pbc and breaking of multiple peptides

[ms]

Hi,

I am doing REMD simulations of multiple homopolymeric peptides in a PBC 
box, in vacuum (it's a custom coarse grain). GROMACS 4.0.5. I want to 
analyze features of the system that require to use g_gyrate to find out 
the moments of inertia of the system, for example.

I understand g_gyrate is one of the few tools that is *not* PBC aware. I 
am trying therefore to correct my REMD trajectories using several 
variants of trjconv, with simple command lines like:

trjconv -f xac_12855257520.xtc -o xac_12855257520_pbcatom.xtc -pbc nojump

...but alas, when I see them with ngmx, the peptides tend to break. I 
tried both pbc nojump and pbc atom, with same results (pbc nojump 
perhaps is a bit worse). I can't use pbc cluster because they're not 
always clustered and I am interested in both the clustered and 
non-clustered states.

I'd like to give more information about that but I don't really know 
what is relevant to debug this kind of issue. I looked in the mailing 
list but (despite multiple woes about pbc nojump etc.) I didn't find a 
helpful thread. Any hint?

thanks!
Massimo

-- 
Massimo Sandal, Ph.D.
http://devicerandom.org