[Gmod-help] Re: [Gmod-gbrowse] multi-segmented feature looks fine at low power, but connecting lines between segments disappear when zoomed in

[Scott Cain]

Hi Burcu,

I finally got around to looking at the sample data and config file you
sent me a few days ago.  There were a few problems I had to fix:

1. GFF3 and the Bio::DB::GFF adaptor doesn't always get along, and I
think your GFF3 is one example of that happening.  I switched to the
Bio::DB::SeqFeature::Store adaptor.

2. With the switch to SeqFeature::Store, you don't need aggregators
any more, so I switched the [EntrezGene] track to use gene:GenBank
features and the gene glyph.

3. I changed all occurrences of 'iD' to 'ID' (I'm off to BioPerl next
to see what caused this so I can make it stop).

4. I added a reference seqeunce line; you had a line like this:

Chr10  GenBank region  1   1994762 .    .    .   ID=NW_047331

I changed it to this:

Chr10  GenBank region  1   1994762 .    .    .   ID=Chr10;Name=Chr10

(of course, I suspect that rat chromosome 10 is bigger than that; the
alternative would be to change column 1 to NW_047331 throughout the
file.)

Switching to SeqFeature::Store will require a few changes, but not too
many; basically, the only things that would be affected are the tracks
that currently use aggregators.  Please let me know if you need any
help with the transition.

Scott


On Mon, Sep 15, 2008 at 4:29 PM, Don Gilbert
<gilbertd at cricket.bio.indiana.edu> wrote:
>
>
> Burcu,
>
> It may be you have to work thru a few changes.  The 'iD' problem likely was
> part of it, your aggregator also needs to be updated with corrections for ID/Parent
> tags.
>
>>> EntrezGene{CDS,exon/mRNA}
>
> This one should work when CDS,exon have Parent=mRNA.ID and mRNA has ID=
> This is equivalent to the processed_transcript aggregator
> Bio/DB/GFF/Aggregator/processed_transcript.pm
>
> Aggregators are good when using Bio/DB/GFF databases; the Bio/DB/SeqFeature/Store
> databases do not use aggregators.
>
> PS, one of the tools, likely bp_genbank2gff3, created those funky 'iD' tags,
> for reasons of its own.
>
> - Don Gilbert
>



-- 
------------------------------------------------------------------------
Scott Cain, Ph. D. cain.cshl at gmail.com
GMOD Coordinator (http://gmod.org/) 216-392-3087
Cold Spring Harbor Laboratory