[gmx-users] NMA for a group

[Bert de Groot]

Mark Abraham wrote:
> Christian Seifert wrote:
>> On Friday 22 June 2007 10:59, Bert de Groot wrote:
> 
>> The hessian is the second (numerical) derivative of the atoms in the
>> three dimensions. In the future, I want to calculate the hessian of
>> molecules next to proteins, so the "small droplet of explicit water"
>> will not work here.  I do not understand, why I need to calculate the
>> second derivatives of the atoms of the whole system, if I just need it
>> of a small molecule.
> 
> Er, technically the second derivative of the energy of the whole system
> with respect to the coordinates of the atoms, but I'm sure you know that
> :-) You could certainly produce an algorithm to do this for just a
> subset of atoms' coordinates, but so far I guess it hasn't been done
> because people haven't perceived any value in it. So if you're convinced
> it's a useful thing to do, you can either calculate the lot and extract
> the ones you want, or rework the code.
> 

In addition, be aware that you should expect some negative eigenvalues in such a
case as for the subsystem you're most likely not in a minimum anymore. As Mark
and David already noted it's not immediately clear which physical interpretation
would correspond to such an approach.

Bert

______________________________________
Bert de Groot, PhD

Max Planck Institute for Biophysical Chemistry
Computational biomolecular dynamics group
Am Fassberg 11
37077 Goettingen, Germany

tel: +49-551-2012308, fax: +49-551-2012302
http://www.mpibpc.mpg.de/groups/de_groot