Dear Gustavo,
The problem is that you have a multi-dimensional problem here. There is
coupling between all the parameters that include any two of these atoms and
so as you change one thing you effect something else. E.g you can use
improper torsion parameters to assign an out of plane minimum but angle
terms around 3 of the atoms in the torsion will act to keep the system
planar...
There are also issues in that while the force field equation covers most of
conformational space there are certain conformational changes, e.g. N-H out
of plane deformations on NMA twisting, that can't be accounted for
energetically with a simple un-coupled approach to each parameter.
I have been working on a multiple dimension parameter minimiser for amber
that can do simulataneous parameter fits. I have not tested it sufficiently
yet to ensure it works properly but we could always try it out with your
system. Essentially this will fit multiple parameters to an energy surface
(e.g an abinitio one). If you want me to try this out send me a private
email with your prmtop (with your best parameters so far) and an inpcrd
file. As well as some more info on what you have tried so far, what QM
method you have used to get the energies etc.
Ps. there must be a minimum energy for the data give below. I suspect at the
moment that it is in the wrong place, outside of the 115 to 123 degree
window you give data for below.
All the best
Ross
/\
\/
|\oss Walker
| Department of Molecular Biology TPC15 |
| The Scripps Research Institute |
| Tel:- +1 858 784 8889 | EMail:- ross.rosswalker.co.uk |
|
http://www.rosswalker.co.uk/ | PGP Key available on request |
_____
From: owner-amber.scripps.edu [mailto:owner-amber.scripps.edu] On Behalf Of
Gustavo Pierdominici Sottile
Sent: 19 August 2004 05:31
To: amber.scripps.edu
Subject: AMBER: angle bending
Hi,
I write this mail because I have a problem and can not resolve it.
The problem is the following:
I have parametrized a couple of streching and bending parameters, and have
never have any problem. But there is a bending that seems imposible to
parametrize.
The model consists of a benzene but with a nitrogen(B) replacing an
hydrogen.
Moreover, there is a methyl group bound to that nitrogen.
The bending I want to characterize is that between the carbon of the
ring(A), the nitrogen(B)
and the carbon of th e methyl group(C).
To do this I make an energy plot with gaussian. The points where the energy
was calculated
have the same configurations, except for the angle A-B-C.
The following step was to calculate the non bonded energy to the same
configurations.
When substracting the non bonded energy to the ab initio one, the result is
not harmonic.
I will send you some values so as to add information about the problem:
The ab initio minimum is at 120 °, all the energy are in Kcal
angle ab energy non bonded difference
115 0.67 48.43 -47.76
120 0 46.99 -46.99
123 0.23 46.38 - 46.15
In other calculations I allways found a minimum value, but in this case when
the angle is below 120°, the difference energy decreases instead of
increasing.
Please if someone have had this problem or kwons what is happening, write me
Thanking in advance
Gustavo
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Fri Aug 20 2004 - 02:53:01 PDT
