VMD-L Mailing List

From: Mayne, Christopher G (cmayne2_at_illinois.edu)
Date: Wed Jan 07 2015 - 09:52:14 CST

Elena,

The screencasts were made before we added a proper Bond & Angle optimization module in which data was extracted directly from the Hessian. Adding the optimization layer tends to give much better results. Part of why we used PRLD in the screencasts is because it is part of CGenFF, so you can directly compare your optimized parameters against CGenFF values (don't expect them to exactly match, for a variety of reasons). Taking a quick look at a few of these, the parameters shown in the screencast (directly extracted from the Hessian) are too low. I did a quick optimization using the latest ffTK and obtained values that fit more with those in CGenFF.

Regards,
Christopher Mayne

On Jan 6, 2015, at 7:32 AM, Elena Lilkova wrote:

> Dear VMD community,
>
> I want to parameterize a novel molecule with FFTK. So, I decided to repeat
> the example from the screencasts. I am currently at the Bond and Angle
> Parameters Optimization step. I noticed that the angle force constants
> that I obtain are significantly different from the ones displayed in the
> screencast. They are at least an order of magnitude, sometimes even two
> orders, larger than what you show. I even tried to start with values,
> similar to the those from the screencast and then optimize, but once again
> I ended up with force constants in the range 40. - 350.
> I am new to novel molecules parameterization, but these force constants
> seem quite large to me. I followed strictly the procedure from the
> screencasts and I don't know why I obtain such different results. Any
> ideas/explanations/comments?
> I am using VMD 1.9.2a42. The QM simulations were done with Gaussian 09.
>
>
> Elena Lilkova
>
> elilkova_at_phys.uni-sofia.bg
>
> PhD student,
> Atomic Physics Department,
> Faculty of Physics,
> University of Sofia "St. Kliment Ohridski"
>
> 5 James Bourchier Blvd.
> 1164 Sofia,
> Bulgaria
>
>
>
>
>
>
>
>