[gmx-users] Possible problem with g_hbond

[Andrew DeYoung]

Hi,

If you have time, I'm wondering if you can please help me to think through a
problem that I seem to be having with g_hbond.  I am looking for a way to
list the _individual_ hydrogen bonds as a function of time, by indices --
more than just determining the number of hydrogen bonds at every timestep.
So, at every timestep, I would like to somehow list the hydrogen bonds that
exist during that time.  At timestep 1, for example, a hydrogen bond might
exist between oxygen26 and hydrogen353.  But at timestep 2, this same
hydrogen bond may no longer exist because oxygen26 and hydrogen353 have
drifted beyond the cutoff.  At timestep 3, the hydrogen bond may exist again
because the atoms have drifted together.  

Do you know if generating such a list is possible, perhaps with g_hbond?

I was thinking that perhaps such information would be listed in the log
file, which, according to the manual, is output using the -g flag.  So, I
have tried the following command in g_hbond_d (I had run double precision
simulations):

g_hbond_d -f traj.trr -s topol.tpr -n index.ndx -g test.log

(I select two non-overlapping groups in index.ndx when prompted.)  I also
tried the same without "_d":

g_hbond -f traj.trr -s topol.tpr -n index.ndx -g test.log

But in both cases, I do not obtain test.log or any log file; I only obtain
hbnum.xvg from the -num flag, which is non-optional output.

Have you ever experienced the -g flag apparently not working in g_hbond?  Do
you have any thoughts about what I can try differently?  Thanks!

Andrew DeYoung
Carnegie Mellon University