Artificial database that delegates to specific databases, with a
smart (well, smartish) guessing routine for what the ids. No doubt
the smart routine can be made smarter.
The hope is that you can make this database and just throw ids at it -
for most easy cases it will sort you out. Personally, I would be
making sure I knew where each id came from and putting it into its own
database first - but this is a quick and dirty solution.
By default this connects to web orientated databases, with all the
reliability and network bandwidth costs this implies. However you can
subsistute your own local databases - they could be Bio::Index
databases (DBM file and flat file) or bioperl-db based (MySQL based)
or biocorba-based (whatever you like behind the corba interface).
Internally the tags for the databases are
genbank - ncbi dna database
embl - ebis dna database (these two share accession number space)
swiss - swissprot + sptrembl (EBIs protein database)
We should extend this for RefSeq and other sequence databases which
are out there... ;)
Inspired by Lincoln Stein, written by Ewan Birney.