[MMTK] pb with nma overlap

[hinsen at cnrs-orleans.fr]

On 06.07.2006, at 14:36, Gaelle Debret wrote:

> I use NMA overlap to find which modes better describe transition  
> between two conformations of a protein. Results I get are strange  
> and different from those obtained with webnm@ :

...

> Here is the script I use :
> (NB: structures have already been fitted with profit)
>
> from MMTK import *
> from MMTK.Proteins import Protein
> from MMTK.ForceFields import DeformationForceField

If I remember correctly, WebNMA uses CalphaForceField instead of  
DeformationForceField. For C-alpha models, CalphaForceField gives  
better results than DeformationForceField, so I suggest you do the same.

> for i in range(len(modes)):
>     m = modes[i]
>     o = m.dotProduct(vec) / \
> 	sqrt(m.dotProduct(m)) / \
> 	sqrt(vec.dotProduct(vec))
>     overlap.append(abs(o))

It's also worth checking if WebNMA overlaps are the same as yours -  
they might be the squares, as it is the squares that add up to one.

Otherwise your script looks fine!

Konrad.
--
---------------------------------------------------------------------
Konrad Hinsen
Centre de Biophysique Moléculaire, CNRS Orléans
Synchrotron Soleil - Division Expériences
Saint Aubin - BP 48
91192 Gif sur Yvette Cedex, France
Tel. +33-1 69 35 97 15
E-Mail: hinsen at cnrs-orleans.fr
---------------------------------------------------------------------