VMD-L Mailing List

From: Pawel Kedzierski (pawel.kedzierski_at_pwr.edu.pl)
Date: Thu Nov 05 2015 - 02:04:17 CST

Dear Yue,
My guess is that this could be simply a sum of rounding errors as the
charges are saved with 3 decimal places.
You could manually correct this on one atom or distrubute +0.001 over
five, the button "Adjust charge" is meant for this.
Not a big deal how you do it as atomic charges in CHARMM FF usualy have
the precision to hundreths only.
Regards,
Pawel

W dniu 04.11.2015 o 21:05, Amira Yu pisze:
> Dear Dr. Mayne,
> Thank you so much for you replying, I think I clicked “Calculate
> from PSF” button, otherwise the optimization process would give me
> error message. And I think biphenyl is well established in CGenFF but
> not with the hydorxyl subs. Even from the ParamChem, the charges and
> parameters are still having relative big penalties for
> hydroxylbiphenyl, that is why I am going througth the parameterization
> process. In terms of assigned constraints, I did use the default value
> (-1, 1) which is a fairly large range, but I am not sure whether this
> constraints will lead to a non-zero sum of charges.
> Thank you again.
> Best regards,
> Yue
>
> 2015-11-04 14:51 GMT-05:00 Mayne, Christopher G <cmayne2_at_illinois.edu
> <mailto:cmayne2_at_illinois.edu>>:
>
> Yue,
>
> I don’t know for sure what the problem is; however, my best guess
> is that there is a problem how you assigned the charge groups
> (Charge Constraints box) and/or didn’t click the “Calculate from
> PSF” button. I have seen in very rare cases an optimized charge
> total that doesn’t exactly match computed the “optimized sim” from
> the Charge Constraints section, which is the purpose of the
> “Adjust Charge” box in the Results section.
>
> HOWEVER,
>
> I’m not really sure that you need to be doing parameterization at
> all, as hydroxybiphenyl is a very simple organic molecule that is
> well represented in the CGenFF.
>
> Regards,
> Christopher Mayne
>
> > On Nov 4, 2015, at 9:09 AM, Amira Yu <amirayuyue_at_gmail.com
> <mailto:amirayuyue_at_gmail.com>> wrote:
> >
> > Hi All,
> > I am doing a charge optimization in ffTK for a neutral charged
> aromatic molecule (hydroxylbiphenyl). I assigned the initial
> charges according to the CGenFF parameters. Deleted the aromatic H
> atoms (0.115 according to CGenFF). However, after running the
> charge optimization several times (first run with Stimulated
> annealing and followed with downhill), the charges seemed
> converged. When I checked the total sum of the charges, the total
> sum was not zero, it was -0.005 (I set the tolerence in the
> "advanced settings" to be 0.001). I used the similar method for
> 2-2 biphenol which is a highly symmetric molecule, the sum of the
> charges are very close to zero (3E-16). I am not sure whether the
> non-zero sum of charges represents the charge optimization is
> still not good or it was okay for some kind of tolerence. The
> other reason why I am concerned the sum of charges is that, I have
> to use CHARMM later on to build a solvate system, and this
> non-zero charge will let CHARMM show some level-zero warnings (the
> sum of charges is not an integer).
> > Thank you in advance.
> > Best regards,
> > Yue
> > --
> > Chemical Engineering
> > YU Yue
> >
>
>
>
>
> --
> Chemical Engineering
> YU Yue
> 859-536-6507