[Bioperl-l] writing out featurepairs/computation objects, semantic parsing

[Jason Stajich]

Since there is no format for FeaturePairs or Computation objects, I don't
know that anyone has come up with solid proposals.  For computation
objects it is pretty easy to just stuff things in the notes field in the
rich sequence format feature table (genbank/embl).  As for feature pairs,
I'm not sure - for a SimilarityPair it would be

similarity      20..30
                /method=TFASTX
		/score=112.3
		/hit=AF11234.1:1..10


But let's go back a step - we've talked about a couple of major changes to
SeqIO that have not been discussed lately.

 o Factory objects for creating sequences rather than hardcode Bio::Seq or
   Bio::Seq::RichSeq
   (I've just finished doing this, this morning, next email will detail
    the changes)
 o Semantic parsing to build Gene,CDS, Exon,Intron,UTR objects from
   genbank/embl parsing
 o Rewrite SeqIO to be event based and/or include a grammar for parsing
   the location fields.

-jason
-- 
Jason Stajich
Duke University
jason@cgt.mc.duke.edu

---------- Forwarded message ----------
Date: Mon, 22 Apr 2002 09:29:45 -0400
From: "Liu, Mingyi" <mingyi.liu@gpc-biotech.com>
To: 'Jason Stajich' <jason@cgt.mc.duke.edu>
Subject: RE: [Bioperl-guts-l] Notification: incoming/1168 (fwd)

BTW, is there any existing discussion and agreement on how Computation and
FeaturePair features should be written out?  I sure wouldn't want efforts
wasted on wrong ideas that may spring to mind when I work with them.

Best,

Mingyi