Hi Tom,
Having a non-integer system charge is non-physical and so there is rightly
no fix for simulating it. You cannot add 0.3 of an ion in the same way that
you cannot have 0.3 of an electron. Generally building a system with
non-integer charge means there is a typo in the entry of the charges, or the
molecule you are using was broken up and never designed to be simulated in
isolation. Typically in AMBER each residue should sum to an integer charge,
the exception is terminal DNA residues which are always designed to be
simulated with their corresponding dual chain pair. This is why the linker
etc in the tutorial does not have an integer charge on its own.
How exactly did you get a charge of 2.3 from the tutorial? What do you mean
by corrected the mistakes?
In the case of the tutorial you should end up with a system summing to an
integer charge. If you do not it would be helpful to know exactly what you
did.
All the best
Ross
From: Tom Williams [mailto:dnaafm.gmail.com]
Sent: Sunday, February 07, 2010 12:01 AM
To: amber.ambermd.org
Cc: ross.rosswalker.co.uk
Subject: Question of Tutorials-A: the total charge is not integer
Hello Amber Community,
How to deal with the case when the total charge is not integer?
If the molecule to simulate is with the charge of non-integer,
how to add ions and make system neutral for MD simulation?
For example, in Tutorial A1, after the charges of Dye and Linker
were obtained from Gaussian and RESP fit, the total
charge is not integer when they (Dye and Linker) are linked.
The original one in this tutorial is:
Total perturbed charge: -2.307860
Total unperturbed charge: -2.307860
After I corrected the mistakes, what i get is:
Total perturbed charge: -2.3074847
Total unperturbed charge: -2.3074847
Thanks for your suggestions!
Tom
---------- Forwarded message ----------
From: Ross Walker <ross.rosswalker.co.uk>
Date: Wed, Jan 27, 2010 at 3:27 PM
Subject: RE: [AMBER] Question of Tutorials-A1 about charges compuation from
Guassian for Amber
To: AMBER Mailing List <amber.ambermd.org>
Hi Tom,
You are correct. The charges in the lib file should be those generated from
the RESP procedure. Gaussian is used to calculate the electrostatic
potential which is used to fit the charges using RESP. It looks like the
wrong lib file was uploaded. I never checked this before.
I will update the file to contain the correct charges.
Thanks for spotting this.
All the best
Ross
> -----Original Message-----
> From: amber-bounces.ambermd.org [mailto:amber-bounces.ambermd.org] On
> Behalf Of Tom Williams
> Sent: Wednesday, January 27, 2010 1:46 PM
> To: amber.ambermd.org
> Subject: [AMBER] Question of Tutorials-A1 about charges compuation from
> Guassian for Amber
>
> I have a question of Amber Tutorials-A1. Anyone can help with this
> question,
> thank you!
> http://ambermd.org/tutorials/advanced/tutorial1/
>
> In this tutorial, it seems to me that amber uses the partial charges
> computed from Gaussian.
> But i don't know why these values of partial charges computed from
> Gaussian
> become
> totally different after importing to Amber.
>
> Pls compare these two file:
> 1) the result from Guassian computation: Figure 1.4 at
> http://ambermd.org/tutorials/advanced/tutorial1/section1.htm
> 2) after manual assigning the type and the partial chage to each atom
> and
> generate the file of fam5.lib and fam5_leap.pdb in Section 3.3.1)
> Saving
> our new library file
> http://ambermd.org/tutorials/advanced/tutorial1/section3.htm
>
> you also can use xleap to see the charges
> $AMBERHOME/exe/xleap &
> loadoff fam5.lib
> loadamberparams fam5.frcmod
> FAM = copy FAM5
> mol = loadpdb fam5_leap.pdb
> edit mol
> you can see the charge of atoms
> They are totally different from the ones shown in Figure 1.4 which are
> calculated from Gaussian.
> Can anyone help exlain this? thanks!
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Feb 07 2010 - 02:30:03 PST
