Date: Thu, 30 Jan 2003 17:58:32 -0800
From: "David A. Case"
Subject: Re: Quasi-harmonic analysis

On Thu, Jan 30, 2003, Pratul Agarwal wrote:
>
> Can anybody enlighten me on the difference between quasi-harmonic analysis and
> principal component analysis? Or are they one and the same thing?

They only differ in minor ways: in quasiharmonic analysis, the coordinates
are mass weighted, whereas in PCA they are not.

In quasiharmonic, one generally takes the inverse square root of the
eigenvalues of the fluctation matrix, converts to cm**-1 units, and reports
the *lowest* values (corresponding to the greatest motion).

In PCA, one generally reports the eigenvalues themselves, and looks at the
*highest* values (also correponding to the greatest motion).

It is more typical in PCA analysis than in quasiharmonic to work with just a
subset of atoms (such as the Calpha atoms of a protein.

>
> Also, is anybody aware of a published study which used AMBER's quasi-harmonic
> analysis (quasih) for a protein? A reference would be great.
>

Probably the first Amber quasiharmonic study is this one:

%A M.M. Teeter
%A D.A. Case
%T Harmonic and quasiharmonic dynamics of crambin
%J J. Phys. Chem.
%V 94
%P 8091-8097
%D 1990

Here is a more recent one:

%A W. Cornell
%A R. Abseher
%A M. Nilges
%A D.A. Case
%T Continuum solvent molecular dynamics study of flexibility in
interleukin-8
%J J. Mol. Graph. Model.
%V 19
%P 136-145
%D 2001

However, it is certainly true that the most useful papers on quasiharmonic
and PCA analysis come from non-Amber labs.

..hope this helps...dac