AMBER Archive (2006)

Subject: Re: AMBER: charge models in antechamber

From: David Mobley (dmobley_at_gmail.com)
Date: Thu Oct 19 2006 - 12:48:14 CDT


Eric,

I did a comparison of hydration free energies for a fairly large
series of small molecules with a number of different charge models
obtained from Antechamber (some with RESP fitting of Gaussian output).
I did not, however, try Gasteiger and Mulliken charges (I think pretty
much everyone these days agrees that Gasteiger charges are bad, and I
don't think Mulliken charges are typically thought to be suitable for
molecular simulation). If you're interested in a preprint of this work
you can get it on my web page.
(http://www.dillgroup.ucsf.edu/~dmobley; look for the publication on
comparison of charge models).

David

On 10/18/06, Eric Hu <list.eric_at_gmail.com> wrote:
> Hi,
>
> I have compared all the available charge models in antechamber on my
> system. Three of the models cm1, gasteiger and mulliken failed to
> obtain an expected intermolecular hydrogen bond while esp, bcc and cm2
> generate similar geometries and association energies. By looking at
> the charges on heteroatoms such as N and O, I found that the charges
> are most polarized in cm1 between N and O among all charge models and
> it is the opposite case in the gasteiger and mulliken models. I wonder
> if there is any more detailed research so far on this subject. Thanks.
>
> -Eric
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu