AMBER Archive (2009)
Subject: Re: [AMBER] gaussian 98 input/output for antechamber/resp
From: Karl Kirschner (kkirsch_at_scai.fraunhofer.de)
Date: Thu Aug 13 2009 - 02:26:12 CDT
From my inspection, your Gaussian keywords looks proper for generating an
electrostatic potential. Just for clarification, you are calculating the
cation of CN, which means that four electrons have been removed from the
molecule if one looks at the Lewis structure. I am not sure there are
standard Amber atom types that would properly capture the physics of this
molecule (i.e. Lennard-Jones parameters, bond stretching).
If Gaussian is not printing out the electrostatics after the optimization
(but I seem to recall that it printed them out before and after) doing an
optimization calculation followed by a second calculation to obtain the
electrostatics would solve this issue. You should calculate an esp on an
optimized structure. (You might also want to verify that Gaussian is
calculating the electronic state that you are intending to model - this can
effect the resulting esp.)
The **** is the output, which indicates the number of the ESP Fit Center,
will not effect the esp calculation to my knowledge.
I suggest that you look into using the R.E.D.III program or its website
server, which allows you to address other issues simply - such as the
molecular orientation effects on computing the esp, different types of charge
derivations (e.g. CHELPG), different types of resp weighting, and so on.
On Wednesday 12 August 2009 20:41, wong105_at_llnl.gov wrote:
> Dear Amber users;
> I'm using antechamber (Amber 8)/gaussian 98 to calculate charges
> for a ligand. I start with a mol2 file and obtain from antechamber a
> gaussian input file. The gaussian input file looks like this:
> #HF/6-31G* SCF=tight Test Pop=MK iop(6/33=2) iop(6/42=6) opt
> remark line goes here
> 1 1
> N 1.2410 -4.7650 42.5820
> C 0.1340 -4.4500 43.3000
> I made the assumption that I could take this input file to obtain the
> electrostatic potential for RESP. I also noticed the "opt" keyword,
> which, I assumed served the purpose of relaxing the molecule to obtain
> a reasonable response.
> Two observations:
> 1) I noticed that the electrostatic potential data is only printed
> for the first point in the optimization, rather than the final
> geometry. That would seem to defeat the point of the optimization.
> 2) After 10,000 points, one gets **** for the point number:
> ESP Fit Center 9998 is at -1.493521 -3.825592 -3.033347
> ESP Fit Center 9999 is at -1.062486 -3.825592 -3.033347
> ESP Fit Center **** is at -0.657445 -3.678169 -3.033347
> ESP Fit Center **** is at -0.327253 -3.401105 -3.033347
> ESP Fit Center **** is at -0.111735 -3.027817 -3.033347
> Does the resp program know how to deal with this? It does not yield
> show any errors when I run it to calculate charges, but I wonder if it
> ignores these points. Perhaps others have addressed these points in
> the past, but a quick google search did not yield an obvious answer.
> Any ideas?
AMBER mailing list