[Biojava-l] Questions

Wed, 17 Oct 2001 08:08:58 +0200

Thomas,

one obvious possibility is to look at NCBI's ASN.1 object model on handling
bibliographic data. This has proven useful and, in addition, is available
in XML format.

Uwe

Thomas Down <td2@sanger.ac.uk>@biojava.org on 10/16/2001 04:44:21 PM

Sent by:  biojava-l-admin@biojava.org

To:   Ron Kuhn <rkuhn@Cellomics.com>
cc:   biojava-l@biojava.org
Subject:  Re: [Biojava-l] Questions

On Tue, Oct 16, 2001 at 10:29:28AM -0400, Ron Kuhn wrote:
> Is there a forum related to the BIOJAVA software??? That is really where
I
> should be posting this email.

Yes, this mailing list is where most discussion happens.  Please
post any questions/bug reports/suggestions/patches here.

> I am trying to use your software to parse Genbank and SwissProt flat
files.
> I made 2 fixes (1 for each) and all seems to work fine. I am successfully
> parsing the files and interpretting sequences. My only problem is that
the
> information about the references to the sequences should be heirarchical
(in
> SwissProt - the RP, RX, RA, RT and RL fields relate to a specific RN) but
> the object used to hold the reference information (Annotation) is NOT
(XML
> is though). If multiple RNs exist and 0 or more RTs exist per RN, then
how
> are we supposed to relate the RTs to the correct RN???

BioJava's currently not very good at handling the metadata associated
with sequence files -- any improvements would be welcome.

When the Annotation objects were originally designed, there
was a general suggestion that they should be used as hierarchies
to model this kind of thing.  So you could have a structure like:

   Top-level annotation
          |
          |      references
          |----------------------> ArrayList
                                      |------->Reference annotation
                                      |           |
                                      |           +----> journal
                                      |           +----> author
                                      |           +----> title
                                      |
                                      +-------> Another reference

Actually, it might be even better to have a dedicated object
model for bibliographical references, but nobody has been particualary
enthusiastic about working on this.

     Thomas.

_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l