Bioperl-db
----------

Ewan Birney


Bioperl-db is a first attempt at making a clean, simple, relational
database representation of sequence orientated bioinformatics objects. It
uses the Bioperl (and therefore the associated biocorba and bio* projects
in general) object model as a source for its design. It provides storage
for a number of associated pieces of information

   - Different "databases" (data collections) of sequences
   - Raw sequence information 
   - Sequence feature information 
   - annotation of sequences including
     - references
     - comments
     - database links

As well as providing this relational structure, bioperl-db comes with Perl
bindings that comply to the Bioperl "interface" definitions for the
appropiate objects, and the database code can store Bioperl interface
definitions. This object layer was modelled after the Ensembl/GadFly
"Adaptor" style interface and in particular provides for "lazy"
fetch-on-demand access of the data,
 

The result is a flexible system which can, for example,

   - read a flat file of sequences into the database

   - provide client/server access to sequence data objects, using 
     the SQL client/server connection

   - provide other bioperl interface using functions, eg, 
     running the Annotation Workbench or the Bioperl CORBA server
     from it

   - read from a different bioperl interface into the database,
     eg, a different BioCorba server can be cached locally 
     using this mechanism


I will discuss the design and implementation of the bioperl-db package