set the label for Uniprot to be the accesion and NOT the ID line names. This is to keep the web looking the same.

Unigene added but commented out for now as the next release does not want them but hopefully the one after will, so want to keep the data in the file

UniGene parser first attempt. Not sure it is worth downloading the .data file for the titles as it is a large file but is far easier than trying to guess from the .seq.uniq file.

Use warning() from Bio::EnsEBML::Utils::Exception instead of warn() from Bio::EnsEMBL::Root.
Fix warning message in pop_stack()

Rationalised database connection; lots of calls to $self->xref()->dbi() and $self->dbi() were causing many database connections to be opened and never closed.

Now only uses one connection per database.

Updated to reflect the fact that a CCDS may be linked to more than one transcript. Hence need to add xrefs only once, i.e. check for existence before adding.

Added add_direct_xref() - will be needed for future parsers that create direct_xrefs (e.g. CCDSParser).

Made add_xref return xref_id of last-inserted xref so that it can be passed to add_direct_xref.