[Biopython] gff3 problem

[Chris Fields]

On Apr 5, 2011, at 8:12 AM, Peter Cock wrote:

> On Tue, Apr 5, 2011 at 2:00 PM, Michal <mictadlo at gmail.com> wrote:
>> Hi Peter,
>> Thank you for your response. By any chance do you know how get the child
>> feature like CDS and exons from the previous gff3 file?
> 
> I would guess (without having had time to really play with Brad's latest
> code yet) they are in the parent feature's subfeatures list.
> 
> This is something that might change - historically the subfeatures have
> been used for the GenBank/EMBL join model, whereby the exons of a
> gene/CDS are held as subfeatures of the same type. There was basically
> just a flat list of all the top level features. We may want to review this for
> the GFF parser and perhaps introduce explicit parent child relationships
> between different types (as specified in GFF3, or inferred for GenBank
> or EMBL files).
> 
> Peter

Though bioperl doesn't explicitly support it, we have classes that do this.  We're also planning on having a class that uses SO ontologies to check relationships and types.  

chris