RE: AMBER: AMBER 9 - Force Field Options - Question

From: Yong Duan <>
Date: Thu, 18 Oct 2007 09:21:00 -0700

Hi, Carlos,

I thought Adrian was questioning the validity of parm94 because his main
concern was the difference in dielectric environment in which the charges
were derived and parm94 was derived in "gas-phase" :).

You are right, to validate this, a set of highly accurate absolute binding
free energies would be desirable. If this is what you meant, the answer is
no, we have not done it. However, we have calculated the solvation free
energies of the two charge sets and found they are indeed quite similar.


-----Original Message-----
From: [] On Behalf Of
Carlos Simmerling
Sent: Thursday, October 18, 2007 3:07 AM
Subject: Re: AMBER: AMBER 9 - Force Field Options - Question

Hi Yong,
I think Adrian was asking if you had published a validation of
the combination, not saying anything about something being
gas phase. I'd be interested too, and I admit that I also did not
read the paper that addresses his question- can you give us the ref?
as far as I could see it's not in the ff03 paper. This would be really
helpful for people doing protein/nucleic acid complexes, like I do,
just to see confirmation that the electrostatics are interchangable.

On 10/17/07, Yong Duan <> wrote:
> Adrian:
> You obviously did not read the paper :).
> Parm94 charges are effectively a set of condensed-phase charges, not
> gas-phase charges. If you look at the dipoles, ff03 and parm94 dipoles are
> quite simular. The ff03-approach is preferred because it gives a
> way to represent the solvent polarization effect.
> yong
> -----Original Message-----
> From: [] On Behalf
> Adrian Roitberg
> Sent: Wednesday, October 17, 2007 11:27 AM
> To:
> Subject: Re: AMBER: AMBER 9 - Force Field Options - Question
> Yong Duan wrote:
> > Simple answer to your first Q is yes. We actually do that too. FYI, for
> > large molecules we often derive charges using parm94 approach. For small
> > molecules and amino acids, we prefer DFT/cc-pvtz(pcm=ether) charges.
> Yong,
> quick question, what do you mean by 'we prefer' ? I presume some
> validation of the procedure must be there, right ?
> I do not see how charges derived with such different dielectric constant
> would be compatible, and in fact I can imagine some simple examples
> where it would fail.
> Not that I believe all the other techniques out there either, but just
> wanted to see what your data was to say that this worked. I believe that
> the community would benefit from seeing this work if you have not yet
> published it.
> Thanks !

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Sun Oct 21 2007 - 06:07:31 PDT
Custom Search