[BioRuby] bio.pdb + biographics

[Adam Kraut]

Even though I work with PDB's often I still write one-off parsers for each
particular project since a generic parser tends to break more often than
not.  There's a good post about this at:

http://boscoh.com/protein/parsing-pdb-files-sometimes-you-really-should-reinvent-the-wheel

The challenges to having a robust PDB parser seem to be caused by
inconsistency in the legacy format.  Since the PDB remediation (
http://remediation.wwpdb.org/) maybe things could be better?  The few times
I've used Bio::PDB it was too slow for me, and this will only get worse if
you try to include every possible use someone could have for macromolecular
structures.  It would be nice if the Ruby language had something new to
offer as I recall similar frustrations when using the PDB parts of Bioperl.
The best PDB manipulation scripts I've used are part of the MMTSB toolset:

http://blue11.bch.msu.edu/mmtsb/Main_Page

But it's still Perl and I'd love to see or help Ruby become more useful in
structural bioinformatics.  Bio::Graphics seems to be focused mostly on DNA
sequences.  I've been meaning to try adding secondary structure glyphs
(Helix glyph, Sheet (arrow) glyph, etc) onto protein sequences similar to
what you might see in the results of a secondary structure prediction.  The
topology graphics like those generated in PDBsum are very nice but just
having the glyphs along a sequence would be convenient.  If I really want to
see a structure in detail it's best to use a something that supports 3D
graphics such as VMD or PyMOL in my opinion.

-Adam


On Thu, Feb 21, 2008 at 2:09 PM, K. Shameer <shameer at ncbs.res.in> wrote:

> Dear Alex,
>
> >> Infact I am specifically interested in
> >> something that has not taktions to start wen up by any of the
> Bio-toolkit
> >> libraries :).
> >> Can we (Alex, Jan, all other bioruby pros and Shameer) work on to
> >> develop a module that integrated both Bio.Pdb and Bio-Graphics module
> to
> >> generate odular topology diagrams ? Please let me know your comments,
> we >> can discuss and work on this :) !!!
> > The parser reads the SHEET / HELIX / TURN records from the PDB file.
> > I've never really played with them, but they look easily readable.
>
> It is always better to get the structure information from DSSP/SSTRUC sort
> of accessory files than the orginal PDB files. The secondory structure
> information will be in detail in these accessory files. If it works well
> with PDB, we can parse the  SHEET / HELIX / TURN as well.
>
> >
> > Going from there to an overall topology diagram (are you thinking of
> > something like this:
> > http://www.ebi.ac.uk/thornton-srv/databases/pdbsum/1tca/domA01.gif)
> >   is more complicated.
> Yep, some thing similar and elaboraded. If it is available as a  module
> the interactivity will be high. In PDBSum only domain based topology
> diagrams are available. They are getting domain information from CATH. In
> our module there should be many possibilities : user can define the
> canvas, color the sec str, print residue no, residue details etc. The
> module should also be able to plot different chains of a given protein as
> well... (i think am all excited, am getting unlimited ideas ... :) )
>
> > I guess the first step would be analysing the
> > structure to see how the secondary structure elements interact and
> > then applying some kind of layout algorithm (putting sheets side by
> > side, etc...) Once you have the layout, using Bio-graphics, Cairo or
> > something else to do the drawing is pretty straight forward (I assume!).
>
> There you go - You are right. I think we can easily trace out the
> structure layout from DSSP/SSTRUC file. I will talk to my Boss and get
> back to you with details.
> >
> > I know Roman Laskowski, who runs PDBSum. So I could ask him how his
> > topology generation script works - if that's the kind of thing you are
> > thinking of.
>
> That will be great !! I asked Roman Laskowski about the availability of
> that program, thats not available for the public at the moment.
>
> Thanks,
> Shameer
>
> _______________________________________________
> BioRuby mailing list
> BioRuby at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioruby<https://lists.open-bio.org/mailman/listinfo/bioruby>
>