[gmx-users] Very large fluctuations in dg/dl

[Berk Hess]

>From: Patrick Fuchs <Patrick.Fuchs at ebgm.jussieu.fr>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>Subject: Re: [gmx-users] Very large fluctuations in dg/dl
>Date: Tue, 8 May 2007 16:36:35 +0200 (CEST)
>
>Hi Berk,
>I just wanted a confirmation because it's a critical aspect. Beyond the 
>debate between PME and RF, I always read it was important to be consistent
>with the (early) force field parameterization when deriving new parameters.
>Even if it does not make any difference for solvation free energy, I 
>imagine this can give some differences on other properties (e.g. heat of 
>vaporization...), or if you want your new parameters to be consistent with
>the remaining of the force field.
>Ciao,
>
>Patrick

Indeed it is extremely important to be consistent in force-field
parametrization.
But the molecules used in the parametrization of the GROMOS
force fields are so small, that on the relevant length-scales
there is no measurable difference between reaction-field and PME,
except for the very bad energy conservation of reaction-field,
which leads to a temperature increase of 5 K, but luckily the solvation
free energies do change significantly over 5 K.
The heat of vaporization, or any other property, will also not be
affected by the choice between reaction-field of PME.

Note that a reaction-field cut-off of roughly 1.4 nm is required though.
Using a shorter cut-off results in artifacts in the water dipole-dipole
correlation.

Berk.

_________________________________________________________________
Talk with your online friends with Messenger 
http://www.join.msn.com/messenger/overview