[gmx-users] Replacing the PRODRG charges

Marc F. Lensink lensink at bigre.ulb.ac.be
Thu Jun 4 22:28:43 CEST 2009


On Thu, Jun 04, 2009 at 03:46:59PM -0400, Justin A. Lemkul wrote:
>
>
> Lucio Ricardo Montero Valenzuela wrote:
>> The target forcefield will be G43a1. But I would move to more recent
>> versions of GROMOS 53a5 or 53a6.
>
> A more empirical charge assignment is often a reasonable starting point for 
> assigning GROMOS charges.  As I understand it, quantum calculations were 
> only done to get electron densities around each atom in various functional 
> groups; the subsequent parameterization was done empirically.  You must 
> then verify that the parameters match some condensed-phase criteria (the 
> force field references will tell you what types).
>
> I have gotten reasonable results based on assigning charges from comparable 
> functional groups within the building blocks in, i.e. ffG53a6.rtp.

I can second that.  charge transfer from comparable functional works
adequately for the gromos force field.  for opls/aa I've achieved
satisfactory results with resp charges.  I expect those will work well
for amber too.  compatibility between resp charges and the gromos
force field is not straightforward.

in any case, _any_ (new) parameterization needs to be verified and
validated!

cheers,
marc



-- 
Marc F. Lensink (Ph.D.)
Structure and Function of Biological Membranes            SFMB
Centre for Structural Biology and Bioinformatics          CSBB
Université Libre de Bruxelles (ULB)     marc.lensink at ulb.ac.be
Boulevard du Triomphe - CP 263, B-1050 Brussels, Belgium
tel: +32 2 650 5411  secr: +32 2 650 2013  fax: +32 2 650 5425



More information about the gromacs.org_gmx-users mailing list