[gmx-users] g_msd and make_ndx

[toma0052]

Hello,
     I am looking to measure the diffusion coefficient of a subset of water
molecules moving through a pore in a lipid bilayer.  In looking at g_msd and
the -mol option, the manual states that my index file needs references to
molecule numbers rather than atom numbers.  I was wondering how I go about
this.  The help for make_ndx is a bit unclear to me.  It seems the selections
that I make all end up giving a reference to the atom number.  Am I missing
something here?  How do I get an index file with a reference to the molecule
number?

Thanks,
Mike Tomasini