AMBER Archive (2009)

Subject: Re: [AMBER] diffusion coeficients

From: case (
Date: Fri Jul 17 2009 - 06:59:41 CDT

On Fri, Jul 17, 2009, felix wrote:

> I would like to know if the langevin thermostat (NTT=3, gamma_ln =3.0)
> is appropriate to estimate hydrodynamic properties such as diffusion
> coefficients. Otherwise we have to use ntt=2

Neither option listed above will work in any simple fashion, since all
thermostats modify Newtonian dynamics. Here is a recent paper that discusses
one way to extract diffusion constants from simulations (using ntt=0):

%A V. Wong
%A D.A. Case
%T Evaluating Rotational Diffusion from Protein MD Simulations
%J J. Phys. Chem. B
%V 112
%P 6013-6024
%D 2008

The above uses explicit solvent simulations. If you wanted to use langevin
dynamics and implicit solvent, you might start from a very old paper:

%A R.W. Pastor
%A M. Karplus
%T Parametrization of the friction constant for stochastic simulations of
%J J. Phys. Chem.
%V 92
%P 2636-41
%D 1988

But you would need a value of gamma_ln very different from 3, and probably
would have to do a fair amount calibration as well. (I have not tried this
method myself....)


AMBER mailing list