VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Tue Jul 12 2005 - 17:37:23 CDT

Hi,
  I believe the problem is that VMD's current PDB reader plugin implementation
isn't paying attention to the "altLoc" specifier in the PDB "ATOM" records.
In the short term, I can see what sort of tweak I could add to the PDB
parser to make it do something more sensible with structures like these.
Ultimately, I'll probably have to modify the plugin interface so that the
PDB reader can more directly communicate this information to the core of
VMD, so that one could make atom selections that are directly based off
of the altLoc field. If people have suggestions for how you'd like to
be able to select and/or view these alternate locations within VMD, I'd
love to hear comments. Perhaps the simplest thing might be to have a new
atom selection keyword for 'altloc'. That'd give access to this information,
though the user would have to be cognizant of the meaning.

Getting bonds right will be another level of difficulty. At the present
time VMD doesn't really consider the possibility of one molecule having
more than one structure, except as a sequence of temporal snapshots, such
as in a trajectory. So, at the moment, VMD wouldn't know what to do with
these alternate locations except to either ignore subsequent alternate
atom locations for the purposes of bonding. With just a little extra
logic, we could eliminate all altLoc values other than " " or "A" from
consideration during the automatic bond determination stage of loading
a lone PDB file. The "B" and subsequent altLoc groups could have their
bonds determined amongst themselves, but that would still leave them
disconnected from the rest of the structure, when displayed with an
appropriate atom selection. Perhaps another way of dealing with this
would be to provide a command for recomputing bonds on-the-fly, much
as is done in the "autobonds" representation, but with a parameter allowing
the user to select the "altLoc" group of interest. None of these are
ideal solutions however, and leave much to be desired.

Another possibility, which is perhaps far simpler would be to load
PDB files containing non-empty altLoc identifiers as multiple structures.
This could be conceivably be done entirely within the PDB reader plugin
without changing VMD itself very much.

How would you guys _like_ to be able to view structures with altLoc
information in them? Would you prefer to see these presented as
completely independent molecules within VMD, or would you prefer to
see them by virtue of something along the lines of "altLoc" atom selections?

The "altLoc" field is mentioned briefly in the PDB specs here:
  http://www.rcsb.org/pdb/docs/format/pdbguide2.2/part_62.html

  John Stone
  vmd_at_ks.uiuc.edu

On Tue, Jul 12, 2005 at 09:36:11AM +0200, Peter Schellenberg wrote:
> Hi
>
> there are certain pdb files which contain variations of a structure, like different
> isomerization states of a chromophore and correspondingly more than one
> structure for neighboring amino acids, for example the structures of photoactive
> yellow protein: 1s1y.pdb
>
> The different structures are distinguished in the pdb file by an A resp.B prior to
> the amino acid (and chromophore) name, for example:
> ATOM 1021 OG1ATHR A 101 ....
> ATOM 1022 OG1BTHR A 101 ....
> ATOM 1023 CG2ATHR A 101 ....
> ATOM 1024 CG2BTHR A 101 etc.
>
> Unfortunately VMD draws bonds not only within A and B but also between all
> atoms within the bond radius that belong to the different representations. Is there
> an easy workaround ? Something like select ATHR and BTHR does not work!
> The only way was to select all atoms belonging to one structure, which is quite
> some work and looks odd.
>
> Many thanks in advance Peter
>
> Dr. Peter Schellenberg
> Institute for Physical High Technology Jena
> Dept. of Biotechnical Microsystems
> Albert -Einstein -Str. 9
> D-07745 Jena 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
Email: johns_at_ks.uiuc.edu                 Phone: 217-244-3349
  WWW: http://www.ks.uiuc.edu/~johns/      Fax: 217-244-6078