VMD-L Mailing List

From: Hannes Loeffler (Hannes.Loeffler_at_uibk.ac.at)
Date: Thu May 30 2002 - 01:29:13 CDT

Quoting "Storr, Mark" <Mark.Storr_at_nrc.ca>:

> Hi,
>
> I would like to calculate the mean square displacement for a selection
> throughout a simulation and hence derive the diffusion coefficient for the
> molecule, as I would like to compare say the diffusion coefficient of water
> molecules close to and far from a protein.

The diffusion coefficient may be obtained from the long time behaviour ("long"
means at least several hundred picoseconds, see recent discusson on CCL) of the
MSD. You simply fit a linear function to your MSD at long time and divide the
slope by 6 (and take care of units). This will give you an estimate of the
diffusion coefficient. Please note, that you should calculate the MSD from
several time origins to improve statistics. Take care of PBCs, i.e. unfold the
trajectory before you do any computations.

Why you want to calculate the RMSD in connection with the diffusion coefficient
is rather unclear to me. As the name implies the RMSD is simply sqrt(MSD). As
you may view the MSD as a variance, you may view the RMSD as a standard
deviation. Therefore the RMSD is often used to compare two similar structures.

Hope this helps,
Hannes. 

-- 
Hannes Loeffler | tel: +43-(0)512-507-5166 | fax: +43-(0)512-507-2714
http://www-c724.uibk.ac.at/staff/loeffler/ | mailto:Hannes.Loeffler_at_uibk.ac.at
Institute for General, Inorganic and Theoretical Chemistry
University of Innsbruck, Austria