From: Minh Ho (
Date: Thu Oct 08 2020 - 16:39:46 CDT

Hi all,

I have a couple questions regarding the process of bond/angle optimization.

1. I used the input (opt and NumFreq at MP2 level) generated by FFTK in the
"calc. bonded" tab and ran the calculation in ORCA. Upon completion of
ORCA, I attached the output and the .hess file into the "opt. bonded" tab
of FFTK. I followed the instructions from the screen cast and the tutorial
documentation to do optimization. I checked the bond lengths and angles,
and they matched very well with the values in par_all36_cgenff.prm.
However, the force constant (k) for both bonds and angles were not similar.
For example, in par_all36_cgenff.prm, the k value of CG2R61 CG2R61:
305.00, but my k value of those two atoms is around 160. Is this difference
caused by incorrect target data from QM calculations?

2. Should I adjust the lower bound and upper bound of K in bond and angle
in advance settings under the "opt. bonded" tab of FFTK to cover most of k
values in par_all36_cgenff.prm prior to running bond/angle optimization to
obtain reasonable k values?

4. I noticed that if I added par_all36_cgenff.prm into "additional
associated parameter files", I got an error saying FFTK couldn't find vdW
for some atom types. Unless I attached every charmm par files or I didn't
attach any additional par files, I didn't receive any error. Why was that
the case?

5. When the bond/angle optimization was in process, namd2 was executed.
However, only 1 processor was used. Is there any way to change the default
number of processors (i.e. from 1 to 3) used by namd2 everytime the
bond/angle optimization is proceeded?

I am new to the whole small molecule parameterization, and I am learning
these concepts. Your answers will be greatly appreciated.


*Minh Ho*
BS Biochemistry, Class of 2021
Department of Biochemistry and Molecular Biology
University of Massachusetts Amherst