[gmx-users] Re: [gmx-developers] flexible water model

[Justin A. Lemkul]

Nilesh Dhumal wrote:
> Hello Justin,
> 
> I used 0.1fs timestep and I could get dielectric constant around 56 which
> looks ok. To get the dielectric constant around ~75, should I reduce the
> timestep around 0.05fs.
> 

I have no idea.  Try it.

> I used temperature 298k for my simulation and average tmep. I got is
> 297.83 K. I am doing NPT and used 1.0 bar as reference pressure in my
> .mdp. I am getting average pressure around 2.58 bar.
> 
> I run the equilibration for 1ns.
> 
> Energy                      Average       RMSD     Fluct.      Drift 
> Tot-Drift
> -------------------------------------------------------------------------------
> Temperature                  297.83    10.9687    10.9643 -0.00165807  
> -1.07774
> Pressure (bar)              2.57551    4246.51    4246.51 -0.0360116  
> -23.4075
> 
> Why there is change in pressure? Is there any wrong with simulation?
> 

Not necessarily.  Pressure fluctuates wildly, a fact that is stated on this list 
weekly.  You may want to try equilibrated with the Berendsen barostat (which I 
find relaxes the system nicely), then move to a more correctly defined NPT 
ensemble using Parrinello-Rahman.  For systems that are far from equilibrium, 
P-R allows for much larger fluctuations.  You're also using a very small value 
of tau_p (0.5); you should probably be using something on the order of 2.0 for 
equilibration.

Also realize that you're looking at 2.57551 +/- 4246.51 and wondering if it's 
statistically different from 1.0.  You may find this post useful:

http://lists.gromacs.org/pipermail/gmx-users/2010-October/055403.html

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================