[ensembl-dev] Ensembl release 72 is out!

[Alessandro Vullo]

Hi Hiram,

On 26/06/13 23:05, Hiram Clawson wrote:
> Is it known if the prediction pipeline has had something fixed to
> improve these gene definitions ?

Yes, I can confirm this.
Previously, the pipeline was dumping start_codon features with a test to 
check for the presence of start/end codons using a single unique 
vertebrate codon table.

The GTF pipeline is now using a species-specific codon table, and that 
is my guess as to why we're finding these differences. I'm not sure how 
much the content of the data has had an effect in regard to these.

Hope this helps,

Alessandro