[Biopython-dev] User-defined annotations in Stockholm alignment file

[João Rodrigues]

Got in touch with Sean Eddy and apparently the issue is that hmmer reads
the info from the original database file, which is in FASTA format, and
then considers it as description. There's no attempt to parse any of that
info because of the lack of semantics in FASTA headers.

I'd be in favor of adding a sub parser for this info,  although I'm not
sure how popular it would be. Making it separate from the main Stockholm
parser makes sense for me, this is a special case. What do you think?

A ter, 5/04/2016, 01:08, Peter Cock <p.j.a.cock at googlemail.com> escreveu:

> On Tue, Apr 5, 2016 at 8:58 AM, João Rodrigues
> <j.p.g.l.m.rodrigues at gmail.com> wrote:
> >> > My original question was why does AlignIO ignore "custom" annotations
> it
> >> > doesn't know, while writing (StockholmIO, line 254)?
> >>
> >>
> https://github.com/biopython/biopython/blob/master/Bio/AlignIO/StockholmIO.py#L254
> >>
> >> Because as far as I know only a short list of accepted feature types for
> >> the GS lines exist (from PFAM/RFAM). The associated comment about
> >> this could have been prefixed with TODO - do you have a strong use
> >> case for custom annotations?
> >
> > Not really, just a case where I want to add my own annotations to each
> > sequence in an alignment. We could support "custom" annotations directly
> > under the keys found in the Stockholm file, instead of trying to map them
> > somewhere. That would keep the PFAM/RFAM keys "mappable" but extend the
> > format if people want to add extra things. It's a shame such an
> > annotation-friendly format can only take "official" annotations..
>
> Its more I was strongly guided by PFAM as the only major user of the
> annotations in the Stockholm format when I wrote the Biopython code.
>
> Peter
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.open-bio.org/pipermail/biopython-dev/attachments/20160405/cf50b5ca/attachment.html>