[gmx-users] Fatal error: Frame contains more atoms than expected

[SANTINI sebastien]

On Wednesday 19 May 2004 14:34, David van der Spoel wrote:
> On Wed, 2004-05-19 at 14:14, SANTINI sebastien wrote:
> > On Tuesday 11 May 2004 13:02, David van der Spoel wrote:
> > > On Tue, 2004-05-11 at 12:22, SANTINI sebastien wrote:
> > > > On Tuesday 11 May 2004 11:56, David van der Spoel wrote:
> > > > > then I don't know. If you have saved the trajectory in the trr file
> > > > > as well, you can try to restart the simulation from 1800 ps.
> > > >
> > > > thanks for your help,
> > > > I have already tried to restart the simulation from 1.790 ns.
> > > > the simulation have crashed around 200 ps after with the same
> > > > problem.
> >
> > It seems the problem came from the peptide (a dimer) that went out of the
> > box.
> >
> > using a larger box with the comm_mode = Linear the crash does not happen
> > anymore (but i'm still waiting for longer simulations).
>
> do you mean you had the comm-mode = angular when the crash happened, but
> with comm-mode = linear it works?
>

(maybe it's important : I use GROMACS 3.1.4)

I used comm-mode = angular in my mdp file but I think this option was not 
taken into account because I have a warning : line unknown in the mdp file.
When I look to the trajectory, the dimer move out of the box.
Using comm-mode = linear the center of mass is replaced in the center of the 
box.
my protein AND my box are asymetric so I have just increased the size of the 
box, then if the protein rotate it can not go out of the box.

Am I wrong ?

-- 
____________________________________________________________________

Sebastien SANTINI, PhD Student
sebastien.santini at igs.cnrs-mrs.fr
Tel:   +33 4 91 16 44 55                Fax:   +33 4 91 16 45 49
Laboratory of Structural & Genomic Information (IGS), UPR2589 - CNRS
Institute of Structural Biology & Microbiology (IBSM)
31 Chemin Joseph Aiguier, 13402 Marseille Cedex 20, FRANCE
____________________________________________________________________