[gmx-users] Terminus Atoms genrestr?

[Teemu Murtola]

Hi,

On Fri, Feb 11, 2011 at 21:05, Justin A. Lemkul <jalemkul at vt.edu> wrote:
> I don't know of any Gromacs tool (other than pdb2gmx) that is smart enough
> to determine terminal residues on its own.  If you have multiple copies of
> the same molecule, then you know how many atoms are in each, so you can
> script a little loop that will do it for you, otherwise I don't think
> there's anything that will do this job on its own.  Maybe g_select, but its
> capabilities and syntax are not yet well-documented.

If your terminal residues have some atoms that are not present in any
other residues (e.g., OXT could be such an atom name), you can try to
achieve what you want with g_select with a selection like (for alpha
carbons of the terminal residues)
  name CA and same residue as name OXT

@Justin: If you have concrete suggestions on how the documentation
could be improved, I would be happy to hear them, and consider them
when I have time to think about the documentation.  I do agree that
it's not perfect, and this has been, in part, a conscious decision.
This is because g_select can really make use of only part of the
capabilities of the selection syntax, and currently there are no other
tool that uses selections at all.  Writing a complete documentation of
the selection syntax would require writing a lot of text that would
have very little use in the present version, and could just serve to
confuse people more.  Partial documentation that would only cover
g_select would probably need to be largely rewritten to cover
everything later, so this has not been very high on the priority list
either.  Still, the only critique I've heard so far (from several
people, it seems) is a vague "the documentation sucks", which doesn't
really help in improving things, or even figuring out where people
face problems.  g_select does provide some on-line help (by typing
"help" in the selection prompt), which should at least briefly
describe the syntax.

BR,
Teemu