store it as the release.

NB:  Since the release is data source specific (not species specific),
if you're running the parsing step of the Xref code for more than one
species, only the build number for the last species will be stored as
the release of UniGene.  So, keep species separate.