VMD-L Mailing List

From: Mayne, Christopher G (cmayne2_at_illinois.edu)
Date: Mon Jun 15 2015 - 10:17:01 CDT

Mohan,

a) The modredundant flag is essential to our methodology, which is described in detail in the ffTK paper (Mayne et al. JCC 2013). We use this flag to take advantage of a feature of Gaussian in which we can redefine the molecular system using internal coordinates that project along the full bonding topology. Distortions applied to compute the Hessian are therefore not just arbitrary distortions in Cartesian space, but map directly onto bonds and angles. In turn this facilitates the isolation of bond/angle distortions to the resulting PES, while minimizing contaminations from other terms, as desired for fitting parameters.

b) I don't know enough about the details of the b3lyp vs mp2 to make much of an informed comment from here. I can say that in some cases, the output parameters can be quite sensitive to the parameterizations options. A good example of this is the threshold angle, which is defined in the Advanced Settings. The default angle is set to 10, which means that if the value of the angle from the MM-optimized geometry at the given iteration is within 10 degrees of the QM-optimized geometry (target data), then there will be no contribution to the objective function for that angle--it is within the acceptable threshold. This default works quite well for saturated aliphatic systems, but not very well for rigid aromatic systems, for which angles deviate by much smaller angles.

Regards,
Christopher Mayne

On Jun 14, 2015, at 11:34 PM, Mohan maruthi sena wrote:

Dear Sir,
              I ran with the route section intact (default) and now I could not get the above error. I could get the bond and angle parameters. I tried with different basis set and is working fine. I have the following queries:

a) We are using geom=(allcheck, modredundant), is using modredundant compulsory, we are not mentioning an any redundant coordinate in the input file. But removing the modredundant option from the input file leads to error. Could you explain the reason for using the option modredundant?

b) I have used b3lyp/6-31g* level qm calculations for hessian and I got very less values for force constants and equilibrium bond lengths? But in the example mentioned in fftk tutorial these values are comparable to general force field parameters used in the md simulations. Could you please suggest some possible reasons for low values of force constants? Please find the attached files.

Thanks & Regards,
Mohan

On Sat, Jun 13, 2015 at 9:51 AM, Mohan maruthi sena <maruthi.sena_at_gmail.com<mailto:maruthi.sena_at_gmail.com>> wrote:
Thank you very much for your valuable time to look into the files and reproducing the error. I will try to use the default QMtool/fftk route section. I will try to run qm calculations with default route section and will let you know.

Thanks & Regards,
K. Mohan Maruthi

On Fri, Jun 12, 2015 at 10:22 PM, Mayne, Christopher G <cmayne2_at_illinois.edu<mailto:cmayne2_at_illinois.edu>> wrote:
Mohan,

Apologies for taking me a couple of days to get to this. I can confirm that the error is reproducible, and that the error is occurring in the QMtool code when parsing and/or manipulating the hessian data. This is almost certainly because you have modified the route section to read:
----------------------------------------------------------------------
 # opt=redundant freq b3lyp/6-31+g(d,p) geom=(connectivity,printinputor
 ient) scf=tight
 ----------------------------------------------------------------------

as opposed to the QMtool (and ffTK) default:
----------------------------------------------------------------------
 # MP2/6-31G* Geom=(AllCheck,ModRedundant) Freq NoSymm IOp(7/33=1) SCF=
 Tight
 ----------------------------------------------------------------------

In the future please include the fact that you have made such modifications in the description of your error. This information is quite pertinent and will likely save others a good deal of time tracking down the source of the error. Modifications to the route section will almost certainly change the structure/details of the output Gaussian log file. QMtool uses specific keywords to parse the desired data from this file, and I suspect that something has changed with these keywords in the output due to the modified route section.

I will try to track down where the keyword mismatch is occurring, but it will have to wait until next week. From the input route section it appears that you want to change the basis set used for the calculation. I'll assume that you have good reason for this, so in the meantime you might try using the QMtool/ffTK default route section only modifying the basis set / level of theory option (i.e., leave the rest alone) and see if the output is compatible with QMtool's parsers. Something like:

# b3lyp/6-31+g(d,p) Geom=(AllCheck,ModRedundant) Freq NoSymm IOp(7/33=1) SCF=
 Tight

Regards,
Christopher Mayne

On Jun 8, 2015, at 9:59 PM, Mohan maruthi sena wrote:

Dear Sir,
             Thanks for a reply. Please find the required files as attachment.

Thanks & Regards,
K. Mohan

On Mon, Jun 8, 2015 at 10:26 PM, Mayne, Christopher G <cmayne2_at_illinois.edu<mailto:cmayne2_at_illinois.edu>> wrote:
Mohan,

The error appears to be occurring deep within the QMTool procs. Can you send me the following files (files used in the Opt. Bonded "Input" section)?

PDB containing optimized geometry
PSF containing optimized charges
LOG file from the Gaussian calculation of the Hessian
PAR files for the in-progress parameter file and any associated parameter files

I'll see if I can reproduce the error and then track down the problem.

Regards,
Christopher Mayne

For future reference, questions regarding ffTK are typically (although not always) more appropriate for VMD-L rather than NAMD-L.

Date: Sun, 7 Jun 2015 05:05:44 +0530
From: Mohan maruthi sena <maruthi.sena_at_gmail.com<mailto:maruthi.sena_at_gmail.com>>
Subject: namd-l: FFTK - error

Hi all,
          I am trying to generate parameters for a molecule using fftk. I
am currently using vmd 1.9.2 version. After performing optimization of
molecule in gaussian, I perform hessian calculations to generate bonded
parameters. When i try to generate bonded parameters, the following error
message is displayed,

. can't use empty string as operand of "/"
can't use empty string as operand of "/"
   while executing
"expr {$fc/double($n*$n)}"
   (procedure "assign_fc_zmat" line 37)
   invoked from within
"assign_fc_zmat $inthessian_kcal"
   (procedure "::QMtool::load_gaussian_log" line 56)
   invoked from within
"::QMtool::load_gaussian_log $hessLog $hessLogID"
   (procedure "::ForceFieldToolKit::BondAngleOpt::optimize" line 81)
   invoked from within
"::ForceFieldToolKit::BondAngleOpt::optimize"
   (procedure "::ForceFieldToolKit::gui::baoptRunOpt" line 40)
   invoked from within
"::ForceFieldToolKit::gui::baoptRunOpt "
   invoked from within
".fftk_gui.hlf.nb.bondangleopt.runOpt invoke "
   invoked from within
".fftk_gui.hlf.nb.bondangleopt.runOpt instate {pressed !disabled} {
.fftk_gui.hlf.nb.bondangleopt.runOpt state !pressed;
.fftk_gui.hlf.nb.bondangleopt

Can you please suggest me a way to overcome this error.

Thanks & Regards,
Mohan Maruthi

<fftk_error.tar.gz>

<enox_bond.par><hess.log>