[Biojava-l] ACeDB?

[Matthew Pocock]

Mike,

Good questions. The answer is partly selfish, and partly alturistic. ACeDB
is an object-inspired database system, optimized for storing information
about genomes. So, it handles stoorage of DNA, features within sequences,
who did what to which clone, and how they did it, and the like.

http://www.acedb.org/ for more accurate information

Several of the complete genomes have been released as ACeDB databases which
contain both the sequences and the biological information about the genomes.

Jade is the original java bindings, but we found them slow and are not very
java-like. Also, they require a pipeline of server/clients to be set up -
this requires root access. ACeDB is developed in an office four doors away
from us, so when they said that they were planning a pure socket
server/client implementation, we decided to write a thin Java client. It
turns out to be quite easy, and is very fast - IO bound. It means for us
that we can manipulate ACeDB sequences using all the code in our toolkit
without having to ask ACeDB to dump intermediate files. For example, we can
run an HMM over chromosome 22 without making chromosome22.fa ever and all
the annotations are visible to us.

The ACeDB team are happy, as it is an independantly implemented client, and
catches different bugs to their client. We are happy as it is a different
implementation of Sequence, and lets us do realy cool things.

I hope that others will be happy as they can use it to write analysis
scripts, corba servers, ACeML dumpers and all manner of things that would be
harder otherwise.

Hope that helps. Does anybody else have a view on this?

Matthew

Mike Marsh wrote:

> Most of my work is in developing tools and testing algorithms for
> protein analysis.  I know nothing of processing raw genome data for
> sequencing projects.  I don't code in perl and haven't ever used any
> bioperl code.
>
> That having been said, can someone (I don't need a response from
> everyone) tell me what ACeDB is, and why there is such an emphasis on it
> in the posted code?
>
> thanx,
> mike
>
> -------------------------------------------------------------
> Mike Marsh
> Graduate Student in Structural and Computational Biology
> Baylor College of Medicine.  Houston, TX
>
> FON: 713/798-6034
> Permanent Email:  mikemarsh@bigfoot.com
> -------------------------------------------------------------
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l