[gmx-users] Box Exploding Error

[Mark Abraham]

VENKATESH HARIHARAN wrote:
> I would like to thank everyone for their help.  Several individuals responded
> to my initial post.
> The consensus seemed to be to use distance restraints. 
> My question is, because distance restraints
> creates penalties if the
> distance between two atoms EXCEEDS UPPER BOUNDS, wouldn't that apply more for
> pulling
> two terminal atoms apart?

So the penalties start really high and get smaller. Where's the problem?

> In example, as the peptide is being pulled
> apart, if the distance exceeded a certain
> value, a penalty would be assessed
> to the force.  In my simulation, with the terminal ends coming together,
> how
> UPPER bounds be set?  Would I set my upper bounds very low (i.e. - 0.5
> nm range)?  This way, even the starting
> structure, with the distance between
> the terminal ends starting at ~3.5 nm, would be assessed a penalty. 

Sure, that's the purpose - you're just doing a rough change to suit your 
whim, once you're happy with your termini, take off the distance 
restraints, equilibrate and simulate for your real objective.

> Another
> idea suggested by the GROMACS manual was to incorporate a [bonds]
> type 6 exclusion in the topology, to create a
> harmonic potential between the
> two atoms.  Would this serve to help bring the two ends together during
> simulation?

Yes, but it has a less flexible form for the potential.

Mark