[Biopython-dev] SeqIO

[Brad Chapman]

Hi Thomas;
Hope you're doing well!

[I ask how many features we want to keep between conversions]
> All of them. I think each GenBank feature has an exact equivalence in EMBL
> and SwissProt (GenPept). So that leaves us just with the definition of the
> corresponding feature names.

[relatedly, I ask in a confusing manner about a "specialized
converter"]
> I don't know if I understood this question...

What I mean is that I'm not sure how I would plug in "lossless"
EBML->GenBank conversion into the current scheme. I can write a
generic writer that will convert a basic SeqRecord to a simple
GenBank (no features). But the way ReadSeq.Convert works now is that
I only get a record, and don't know the starting format. In order to
have "smart" EMBL->GenBank I need to know that the format is EMBL,
so I can look for the conversions.

It seems like a simple thing to do would be to add an optional
second argument to write, so that I could do something like:

def write(self, record, starting_format):
    if starting_format == "embl":
        _do_embl_to_genbank()
    elif starting_format = "swissprot":
        _do_swissprot_to_genbank()
    else:
        _do_generic_to_genbank()

Does this make sense and seem like a good idea? Or am I still making
no darn sense?

[I ask about duplicated SeqRecord stuff]
> I copied everything so that I cóuld play around without breaking e.g. your
> code. Now I think the changes are actually backward compatible - so we
> could move it back.

Yeah, you are more than welcome to add additions that are
back-compatible to the SeqRecord stuff. This will help eliminate
duplication (and thus lots of confusion for me :-).

> P.S. is anybody going to the Atlanta meeting in November ?

Well, I am not going (I was kind of scared away by the
"Bioinformatics after Human Genome" sub-title), but I am only an
hour away from Atlanta so I'll definately "be in the area" :-). Many
of the talks look good though, so I may try to sneak in to listen to
a couple of them.

It's-okay-for-graduate-students-to-be-cheap-ly yr's,
Brad
-- 
PGP public key available from http://pgp.mit.edu/