[gmx-users] g_msd

[Rainer Boeckmann]

> could you give some comments about the g_msd program? When option -mol is used
> how does the program calculate the diff.const?
>
> %g_msd -f traj.trr -s topol.tpr -n index.ndx -mol -mw -lateral z
>
> if traj.trr is used as input, i would expect to get MSD of the
> center of mass of the molecule as a function of time.
>

msd as a function of time is given in msd.xvg. Be careful, the index-file has to
contain the molecule numbers, not the atom numbers!


> instead, the output diff_mol.xvg contains diff.constants as a
> function of molecular number..?

yes.


>
> What are -beginfit and -endfit that the program used to fit
> a straight line and compute the diff.const?
>

The program fits a straight line from -beginfit to -endfit to the obtained mean
square displacements (msd.xvg). The lateral diffusion coefficient is then given
by 1/4*slope of this fitted line.


>
> p.s. the system i simulate is a bilayer mixture of two types
> of lipids in a water box.
>
>

you should be aware that you need very long simulation times in order to get a
good estimate of the diffusion coefficient at the nanosecond time scale (we used
100ns simulations of a 128 lipid patch to get reliable diffusion coefficients on
the time scale of up to 6ns). Other problems that might/will appear in the
determination of the diffusion coefficient for individual lipids in a bilayer
concern the possible motion of the monolayers with respect to each other and the
periodic boundary conditions.

good luck,
rainer


--
Dr. Rainer Böckmann
Theoretical Molecular Biophysics Group
Max-Planck-Institute for Biophysical Chemistry
Am Faßberg 11
D-37077 Göttingen

tel:    +49(551)201-1141
mobile: +49(172)9531168
fax:    +49(551)201-1089

email: rboeckm at gwdg.de

http://www.mpibpc.gwdg.de/abteilungen/071/rainer/index.html