On Wed, Apr 13, 2016, Kenneth McGuinness wrote:
>
> So I would like to use gaussian to calculate atomic partial charges (new to
> this completely so any help would be great!)
>
> The paper <http://dx.doi.org/10.1080/07391102.2015.1092475>, works on the
> molecule I am interested in, suggests to use the NBO method for calculating
> partial charges
>
> I would like to/ or maybe not reproduce their results for comparison ( I am
> open to suggestions since I am interested in the molecule binding to
> proteins rather than molecule binding to carbon nanotubes as in the paper)
>
> Looking at the Amber manual: Antechamber (RESP, CM1 ... methods are read
> in)
Partial charges for molecular mechanics generally aim to be somewhat
"over-polarized" to try to represent electrostatic interactions in water,
not in the gas phase. In the "amber" family of force fields, the RESP
procedure is very commonly used (see the R.E.D. website for scripts that
make this pretty straightforward.) But there is nothing "magic" about
the RESP procedure (it relies on systematic errors in low-level, HF
6-31G* calculations)--it just leads to the same mistakes for everyone.
An alternative procedure that we have worked on is described in JPCB
117:2328, 2013. I'm not necessarily suggesting you use this, but the
paper describes many of the tradeoffs involved in trying to find simple
fixed charge molecular mechanics models that are appropriate for condensed
phase simulations.
You are certainly welcome to use whatever charges you wish, or which are
recommended in papers you trust, and you can indeed just copy such charges
into the mol2 file that antechamber would create.
...good luck...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 13 2016 - 18:00:04 PDT
