AMBER Archive (2009)
Subject: Re: [AMBER] help needed for HID
From: FyD (fyd_at_q4md-forcefieldtools.org)
Date: Mon Jul 13 2009 - 15:27:13 CDT
> I was just testing the generation of partial charges of HID residue
> using RESP method. I capped HID with ACE on N-terminal and NME on
> C-terminal and optimized the molecule using
> #p HF/6-31G* gfinp gfprint geom=connectivity
This is not a geometry optimization here...
> then generated the ESP by using:
> #p HF/6-31G* SCF=tight gfinp gfprint geom=connectivity pop=MK IOp(6/33=2)
> After that I followed RESP method (two steps) usnig resp program
> (Amber10). I found that there is a difference between my calculated
> charges and that inside xleap.
> atom amber charge my calculated charge
> N -0.4157 -0.320085
> CA 0.0188 0.026449
> C 0.5973 0.603543
> O -0.5679 -0.588240
> CB -0.0462 -0.242978 big difference here ???
> CG -0.0266 -0.044801
> CD2 0.1292 0.112621
> NE2 -0.5727 -0.568863
> CE1 0.2057 0.181391
> ND1 -0.3811 -0.246646
> H 0.2719 0.227475
> HA 0.0881 0.093945
> HB2 0.0402 0.110400
> HB3 0.0402 0.110400
> HD1 0.3649 0.306931
> HD2 0.1147 0.109502
> HE1 0.1392 0.128956
> Sorry, does it make any sence? I know the charge can be changed by
> changing the molecular orientation, but a little bit, not like this.
This problem has been reported several times in the Amber mailing
list. Charge values based on MEP computation are affected by many
factors: accuracy of the optimized geometry, molecular orientation of
the optimized geometry, charge fitting approach. Thus, charge values
in AMBER force fields cannot be reproducible.
What is the maximum incertainty for an atomic charge value ?
This is impossible to say, it depends on the choices you are going to
make in charge derivation.
This is why we develop the R.E.D. program to rigorously define all the
parameters that affect charge values. This means that if you use
R.E.D. you will not be able to reproduce charges in Amber force
fields, but people will be able to reproduce the charge values you are
going to derive using R.E.D.
AMBER mailing list