Hi Alan,
In principal there is no reason why one can't mix these two force fields.
The equations are suitably similar that it should work, however how well the
actual charge models interact is probably open to lots of debate.
One key requirement is that the scaling of 1-4 interactions is different in
GAFF to in CHARMM. For GAFF 1-4 EEL is divided by 1.2 and 1-4 VDW by 2.0.
CHARMM does not employ any scaling but instead uses different parameters for
1-4 VDW interactions. In principal these two force fields can be mixed
'correctly' in AMBER 11. One would probably have to hack the prmtop a bit to
get it to work properly but there is now support for the 1-4 scaling to be
different for each dihedral type. I have not tried this mixing myself though
and one should consider it an advanced topic.
So I think the real question is not so much if one can correctly mix these
two force fields but whether the interaction between something described
with CHARMM and something described with GAFF is correct. I.e. H bonding
between a GAFF ligand and a CHARMM protein for example.
This is really just something that would have to be carefully tested but I
see no reason why it would not work okay.
All the best
Ross
> -----Original Message-----
> From: Alan [mailto:alanwilter.gmail.com]
> Sent: Wednesday, September 08, 2010 10:36 AM
> To: AMBER Mailing List
> Subject: [AMBER] mixing GAFF with CHARMM
>
> Hi there,
>
> As the developer of ACPYPE (a tool for adapting antechamber topologies
> generated to other MD engines) I usually see people asking me about
> mixing
> GAFF with CHARMM.
>
> I myself do some sort of mixing buy using GAFF with CNS (Engh & Huber
> parameters) but I believe I know what I am doing. However, I never used
> Charmm so...
>
> Well, what I am looking is for advices and maybe references concerning
> this
> topic. How tabu is this and what would be the implications and
> potential
> problems. For example, AFAIK oplsaa and amber/gaff have a lot in common
> (ok
> parameters -- bonded and nonbonded -- may differ, but sometimes just
> slightly, however topology -- in the sense how atoms are connected --
> is
> very similar if not the same!).
>
> Many thanks in advance,
>
> Alan
>
> --
> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
> Department of Biochemistry, University of Cambridge.
> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
> >>http://www.bio.cam.ac.uk/~awd28<<
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Sep 08 2010 - 11:30:03 PDT
