[gmx-users] g_hbond

Mark Abraham Mark.Abraham at anu.edu.au
Sat Apr 18 01:42:28 CEST 2009

Dmitry Lupyan wrote:
> Dear GROMACS Users,
> I'm trying to use g_hbond (v4.0.3) to analyze a simulation which was
> generated with NAMD (trajectory was converted to .trr w/VMD).  The
> g_hbond requires (i) a trajectory file, and (ii) a 'run input file'
> (tpr/tpa/tpb).  How do i generate this tpr/tpa/tpb file? Is there any
> way I can use a plain PDB file for this?

No, you can't get away with a plain coordinate file. To identify the 
atoms, an ordered mapping from the atoms in the .trr file to atom names 
in a coordinate file would suffice, however g_hbond also needs a 
description of the atomic connectivity. I guess that's so it can exclude 
nearest-bonded-neighbour atoms from being observed withing H-bonding range.

There's no easy solution - you have to take a suitable coordinate file, 
apply pdb2gmx, check the atom ordering is still the same, and then 
supply it to g_hbond.


More information about the gromacs.org_gmx-users mailing list