[gmx-users] Constraint distance question

[Warren Gallin]

Hi,

	I am trying to look at the free energy differences as a function of end-to-end distance of peptides in water.  I am running a set of simulations with the two atoms of interest joined by a type 2 constraint, with the length varying from 0.5 to 3 nm.  However, these simulations all explode, even though I am starting with conformations from an unconstrained simulation that are taken through another round of energy minimization and relaxtion before running the constrained simulations.

	There is an mdrun option, -rcon, that I have not been setting.  I read this as the distance over which atoms interacting through the constraint will be communicated between processors for the P-LINC algorithm.  since some of the end-to-end distances can be relatively large, I was wondering if, in principle, I should be setting -rcon to the constarint distance +1 nm to ensure that the constraint is not randomly disappearing from consideration when the two ends of the peptide are in different domains.

Thanks for any insights.

Warren Gallin