[BioPython] martel & mindy

[Jeffrey Chang]

On Thu, Jul 18, 2002 at 10:49:54AM +1000, Cath Lawrence wrote:
> (From the UniGene thread)
> On Thursday, July 18, 2002, at 05:45 AM, Jeffrey Chang wrote:
> > On Wed, Jul 17, 2002 at 07:34:25AM -0700, Sagar Damle wrote:
> >> thoughts anyone?  I'm not really a programmer, just a scripter, so I
> >> may be way off-base here.
> > Yes, I think you are right.  For more than trivial use of LocusLink,
> > you really should download the flatfiles and index them by hand.
> > Fortunately, Biopython does provide these capabilities with
> > Martel/Mindy.  If you build parsers for this with Martel, Mindy can
> > automatically take a parser and flat file and build indexes for fast
> > access to the file.
> 
> Oooh, can I do that with PDB? Can anyone point me to documentation?

There's no documentation yet, although Andrew Dalke wrote some pretty
explicit example code in the mailing list sometime between 6-12 months
ago, I think.

Also, I should've been clearer the first time, but this is code in the
CVS and hasn't been released yet.  Mindy is still in development, but
stable enough that people have used it to index GenBank successfully.


> I have the CVS version of BioPython with PDB parser but don't know much
> about its internals yet. It is also pretty slow with big records, and
> building an index on just a few fields would be really useful.

Unfortunately, the current PDB parser in BioPython doesn't use Martel,
so it doesn't get these fancy new capabilities!  :) However, if you're
only interested in a few fields (and not the whole format), it should
be pretty easy to build a parser to handle what you need.

Jeff