Hi Kevin, Paul and others,
On Wed, 2 Jun 2010, Kevin Cowtan wrote:
My understanding from Gerard K is that the next version of the PDB format
will also use unjustified atom names.
Unfortunately, following mmdb, I also backed the wrong horse and assumed
space-padded atom names were here to stay. A certain amount of re-engineering
is going to be required.
My current plan (which will be implemented in clipper, but I guess mmdb needs
something similar) is to have two modes in the coordinate code: a legacy mode
which auto pads atom names on read (which is easy if the element name is
present and requires guesswork otherwise). By selecting this mode programs
will work without other modification (a 1-line change).
wwPDB mmCIF atom names should be related to the chemical component list, as
I am sure you know. The CHEM_COMP_ATOM category contains an item
_chem_comp_atom.pdbx_align, which can be either 0 or 1, and indicates
explicitly how the atom name should be aligned in padded PDB-style output.
This item appears only in the atom list, not in CHEM_COMP_BOND, and so the
alignment information is absolutely not part of the atom name. I hope that
the PDB will continue to maintain this data item for backwards-compatibility
purposes, even after they switch to unjustified atom names.
If you are going to be doing some re-engineering anyway, then you might like
to consider using alignment information if this is available (e.g. the user
has a local copy of the PDB component list), in addition to the element name
and heuristics. Ideally the alignment information should be included in
refinement dictionaries too, but I don't know how likely that is to happen.
Regards,
Peter.
--
Peter Keller Tel.: +44 (0)1223 353033
Global Phasing Ltd., Fax.: +44 (0)1223 366889
Sheraton House,
Castle Park,
Cambridge CB3 0AX
United Kingdom
