[gmx-users] Normal Mode Analysis on Calphas only ??

[Berk Hess]

>From: Marc Baaden <baaden at smplinux.de>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] Normal Mode Analysis on Calphas only ??
>Date: Tue, 17 May 2005 17:05:22 +0200
>
>
>Hi,
>
>we started to do some normal mode calculations on a protein complex
>and several mutants. To simplify things (and because the different
>complexes have different numbers of atoms which makes comparison more
>difficult) we wanted to compare eg Calphas only or per residue only.
>
>This seems to be more difficult than expected, and I wonder at which
>point this could be done. Current guesses are:
>
>  - "edit"/modify the hessian matrix .mtx file to only include data
>    on Calpha only .. and from then on "standard" procedures
>
>  - "average"/sum up the eigenvectors.trr for each residue (not sure
>    this is correct)
>
>the last resort kind of solution seems to me
>
>  - extract an ensemble from the NM calculation, do PCA on the Calphas
>    of the ensemble .. but this really seems like serious overkill
>
>I hope I have just missed the simple, obvious and straightforward solution,
>and that some kind person will point it out to me.

There is only one proper way of doing this.
You need to diagonalize the whole Hessian and then reduce it to
a C-alpha only matrix. I forgot the exact formula for this, Herman
Berendsen derived it some years ago, today I don't have time
to search for it.
You can not just take the C-alpha elements, as all non C-alpha
atoms need to be minimized.
The formula is quite simple.

Berk.

_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar - get it now! 
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/