[gmx-users] position_restraint anfd full Dynamics and berendsen Nose hoover Thermostat

[Mark Abraham]

On 7/04/2010 9:08 PM, shahid nayeem wrote:
> Dear users
> Please let me know some basic question. I am sorry if I am asking a
> silly question.
> a) While simulating a protein thermal unfolding at high temperature say
> 500K should I run position restraint dynamics at 500k or at 298K. I am
> doing 100ps position restraint dynamics at 500k. Am I right in doing so.

The usual purpose of position restrained MD is to allow the system to 
achieve close to the ensemble desired for the start of the simulation 
without perturbing the initial structure.

> b) While using Berendsen thermostat, how to select a value of tau_t and
> how does it affects simulation output. The same question for NoseHoover
> thermostat.

Start by reading the GROMACS manual sections.

> c) After completing the 100ns simulation when I view the .xtc file in
> VMD I find that  alpha helix of the protein is unfolded in very first
> frame collected at 10ps interval and reduced by trjconv  at 50ps
> interval. Is it possible. If not then where I am wrong.
> I did Steepest descent minimization, then 100ps position restraint
> Dynamics at 500K and then 100ns full dynamics at 500k.

10ps would be very fast for a lengthy helix to unfold, however you've 
deliberately taken your force field out of the zone where it was 
parameterized, so we can't really infer anything about its behaviour. 
You could also have a broken model physics in your .mdp file, but we 
can't know about that.

The real lesson here is not to do a lengthy simulation without looking 
at the early and ongoing results.

Mark