[Bioperl-l] SearchIO to GFF (was: Getting 'features'fromSearchIO?)

[Dan Bolser]

2009/5/13 Jason Stajich <jason at bioperl.org>:

> Dan ---
> I am slowly cleaning up and migrating more of the pipelines that we use to
> generate polished GFF for loading into Gbrowse.  You should just set a goal
> of making good GFF3 and many of the visualization and query issues go away.

I see. That's actually really useful to know.


> For genomes I tend to have a separate scaffold file which is just the 1 line
> per chromosome/scaffold/contig listing.

Is this the file that defines the 'entry points' for each sequence
object? i.e. instead of generating this on the fly while processing
the blast results, do it once initially for every sequence object you
have?


> Then a file for each analysis ie. Protein to genome BLAST, EST to genome,
> remapped Pfam domains to genome coordinates, gene predictions, etc. Keep it
> simple at first you know what you've updated when something stops working
> the way you expect.
> -jason

Thanks all for the help. Very much appreciated!

Dan.


P.S. dumb question but... what defines a HIT in blasttable format? is
it strictly the set of HSPs with the same query and subject, or is
phase and relative position also taken into account?