[Bioperl-l] Species validation

[Ewan Birney]

At the moment as people have noticed we are blowing up on impossible to
decipher species. Of course, some of these really are impossible to
decipher, like the mouse/rat hybrid thing (though it slightly annoys me -
surely this thing *came* from either mouse or rat!)


What I suggest is that we test whether the classification strings make
sense - if they do, we make a standard species object. If not, we make a
specific "unknown species" object, classification "unknown", species
"unknown" but store the original strings somewhere.



Any other thoughts - throwing exceptions is not good in the middle of EMBL
parsing and just letting anything go in there is also not good...




What does genbank do for these entries?