[Biopython-dev] Bio.GFF and Brad's code

[Brad Chapman]

Jose and Peter;

> > Right now I'm storing the result as repr for the SeqRecords, but I would like
> > to write gff files at the end. I've read the discussion regarding Brad's code
> > and I've found it very interesting.
> > I need to write those gff files so couldl use Brad's code or my own, but it
> > would be great if I could contribute to Biopython at the same time.

Awesome. Please do use my code for output and feel free to fork and
make suggestions; I'm happy to integrate changes:

http://github.com/chapmanb/bcbb/tree/master/gff

> > At the time being I don't think a consensus about what a SeqFeature should
> > represent and how. I think Peter made a proposal about adding a parent and
> > children properties, is this a good way to solve the problem?
> > Best regards,
> 
> Brad's code is using the SeqFeature differently to existing bits of
> Biopython, and adding a separate child/parent mechanism for the
> kind of usage required for GFF(3) looks like one way forward allowing
> use to keep full backward compatibility. I'm actually going to see Brad
> in person next month at a workshop, and I'm hoping we can squeeze
> in a little in person debate on this then (assuming we don't settle it
> here on the mailing list first of course).

What do you think we need to modify in the GFF parsing code to bring
this in line? I'd really like to see this get into Biopython, but am
not sure how to clear the blocking issues. If we can put together a
list of specifics, I can try and put together time to tackle that.

Brad