[gmx-users] Msd of molecule center of mass (g_msd -mol with -rmcomm)

[Matthew Hoopes]

Hello,

Does anyone have additional details about how the calculations are done for
the -mol flag of g_msd when using -rmcomm? I am comparing results with and
without the -mol flag. I presume that when -mol is used the file specified
by the -o flag contains the average msd of the centers of mass of the
molecules chosen with the index group and when -mol is not used, -o contains
the average msd of atoms in the selected group.

I have two specific questions:

(1) Why would the msd of the COM of molecules be greater than the msd  of
the atoms at low dt (or at all length scales in some cases)? This does not
happen when -rmcomm is not used but then the diffusion is wrong (too fast).

(2) Why does the -mol msd curve up along a vertical asymtote at tmax when
using -rmcomm?

My system is a mixture of lipids in a bilayer run for 200 ns where COM of
only the system was removed by mdrun. Trjconv was used to extract one
leaflet of all lipids without water into an XTC file (e.g.
upperleaflet-nojump.xtc). On this extracted sub-system, g_msd was run with
-rmcomm and -lateral z while selecting for one lipid type with an index
file. This was tried with and without -mol and comparing the outputs led to
the questions above.

Thanks,
Matt

-------------------------------------------------------
Matthew I. Hoopes, Ph.D. Candidate
Biophysics Graduate Group
Department of Chemical Engineering and Material Science

University of California, Davis
3118 Bainer Hall
1 Shields Ave.
Davis, CA 95616

mihoopes at ucdavis.edu
530-752-6452
-------------------------------------------------------