[gmx-users] DNA hydration water dynamics

soumadwip ghosh soumadwipghosh at gmail.com
Fri Apr 17 06:25:19 CEST 2015

Hi all,
         I am trying to find out the diffusion constant and the rotational
correlation time of water molecules confined in different DNA segments and
for that

1. I am using a cut -off of 5 Angstrom in g_dist with dist option. I am
specifying the COM of DNA bases/backbones as index group 1 and calculating
the no. of water molecules (group 2) within that cut-off for 2 ps of my
trajectory. The command (g_dist along with -dist option) prints the indices
of the SOL molecules within that distances for 2 ps. Now it gives indices
at t=0, 1 and 2 ps. So the water molecules which remain in the vicinity of
the desired segment throughout this 2 ps are selected and made custom index
groups with. Obviously, these indices will be different for different parts
of the DNA strand and so will be the no. of SOL molecules.

2. Now with these index groups of different confined waters I am running
g_msd and g_rotacf and as expected the value of rotational relaxation times
for confined waters are coming out to be higher than bulk and the diffusion
coefficients are also in good agreement with the literature.

Now my question is am I doing anything wrong in this analysis?

If I take a 50 ps trajectory with the same water molecules (i.e less than
100 ps which in turn is the typical time by which a biologically confined
water molecule exchanges with the bulk water) is there any ambiguity in the
analysis part?

The MSD plots obtained are linearly correlated with 2 ps simulation time
but for confined water molecules there are some noises and it is not a
typical  MSD vs time straight line passing through origin as observed for
the bulk water for 2 ps. What could possibly be the reason?

Any kind of help will be hugely appreciated.

Soumadwip Ghosh
Research Fellow
IIT Mumbai,

More information about the gromacs.org_gmx-users mailing list