Re: bioperl based database infrastucture for directed graphs

Hi,

2008/1/9, Chris Mungall <cjm <at> fruitfly.org>:
> [cc-d to gmod-schema]
>
> Chado does have some views and pg functions for interval-based
> retrieval. AFAIK there are no accelerators for deep feature graphs,
> as most chado users have relatively shallow gene-model/SO feature
> graphs. It may not be so hard to extend cvterm code for doing this,
> depending on the characteristics of your graphs (the closure of
> feature neighbourhood graphs may be particularly large)

Great! I'm studing Chado and I will have a look at the interval optimizations.
Did any of you compared BioSQL and Chado for huge feature and feature
graph storage/retrieval efficiency? As Sendu pointed to limitations in
Bio::DB::SeqFeature's schema, I'm thinking which of these plataforms
(or maybe another one?) would be best suited for these tasks... for
the moment, I will either extend Sendu's hack of Lincon's modules or
adapt the binning algorithm of Bio::DB::SeqFeature::DBI::mysql to
Chado, if it turns out to be more efficient than the pg functions.

Best,
Robson

PS: I could not find the most recent version of gmod by following the
Download link to gmod(Chado) from GMOD's site to the Sourceforge
download page. Did I miss the right link on the download site or is
this unexpected? Is the version available at IUBio's mirror (0.003-10)
the most recent one?