[gmx-users] Inconsistency in g_dist calculation

[Berk Hess]

>From: sunita at chem.iitb.ac.in
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] Inconsistency in g_dist calculation
>Date: Tue, 16 Nov 2004 11:46:17 +0530 (IST)
>
>Dear gmx-users,
>
>I have calculated the distance between C-alpha atoms of the first and last
>residue of the peptide. The graph should decrease from high value to low
>value since it is the end to end distance of the extended peptide.But what
>I found is  that, during the initial 57 ps the distance is 3 nm and but
>just after this it rises to 3.6 nm.
>
>Then I calculated the distace directly from the VMD by selecting the
>C-alpha atoms and then scale its x-axis to ps and y-axis to nm.When I
>overlaid both the .xvg files, I found that except initial 82 ps, it
>matched perfectly.
>
>Why I am getting such inconsistency while calculating the g_dist (3.1.4).

This is probably due to the periodic boundary conditions.
During the first 82 ps the distance between the two residues is shorter from
one periodic image to another than inside one image.
Unfortunately g_dist does not have an option to turn off the pbc 
calculation.

You could use trjconv -box ... to make the box larger for the g_dist 
analysis.

Berk.

_________________________________________________________________
Play online games with your friends with MSN Messenger 
http://messenger.msn.nl/