[DAS] System requirements

jfreeman jfreeman@variagenics.com
Tue, 04 Dec 2001 15:05:41 -0500


Given the following use case:

A small academic lab has discovered, for example, a new way of
determining human alternate splicing, and has a great deal of data that
they want to publish to the world, they decide on trying das.  The local
software person decides that the easiest install given the documentation
and assumed information in the install documentation, given both Dazzle
and Ldas is Ldas.  He uses the ensembl version of Kent's goldenPath to
determine where everything is and puts all of the data on the Ldas
server in ensembl coordinates.  He installs and begins serving data on
Ldas and views it with geodesic or omnigene.  His PI wants to see as
served in its context on the ensembl website, he opens up
(http://www.ensembl.org/), and goes to manage sources, he adds his Ldas
server, and assuming it works as described in the Ldas documentation,
what happens?

A: Ldas and the ensembl contigview are incompatible, PI gets mad,
software person restarts process with Dazzle
B: Das protocol is implementation independent and the data is shown in
contigview; he and his PI are happy.

Which is the likely scenario?  A or B.
Has anyone had this experience?

Thanks for your time,


Thomas Down wrote:
> On Tue, Dec 04, 2001 at 05:15:11PM +0900, Todd Taylor wrote:
> > Hello,
> >
> > We are thinking about setting up a DAS server at our institute here in
> > Japan.
> Cool.
> > Can anyone give us some recommendations as for the installation system
> > requirements? What kind of system (OS, memory, hard disk, cpu) are
> > others running it on?
> Obviously, it's quite strongly dependant on the amount of
> data you're serving, and the way in which you're storing it.
> But looking at current servers, requirements don't seem to
> be excessive.  Using Dazzle, I can serve human chromosomes
> (in Ensembl-format MySQL databases) without any trouble
> on my aging laptop (Pentium II/300, 160Mb RAM).
> In general, so long as you've got enough hard disk for the
> data, I can't see any real trouble.
> > And, can anyone explain the difference between the Dazzle server and
> > LDAS?
> LDAS is a perl-based server, backed by MySQL relational
> databases (with it's own schema, based on GFF).
> Dazzle is a Java servlet based DAS server framework.  Other
> than the language issue, the key difference is that it uses
> little `data source' plugins, which can be backed by files,
> relational databases, or any other means of data storage.
> The plugins present data to the server core as BioJava
> Sequence and Feature objects.
> If you want to hook a DAS server to an existing database,
> this is probably easier with Dazzle (at least so long as you
> have some basic Java experience).  If you're expecting to
> have to convert the data into a new schema anyway, then go
> with your language preferences.
>     Thomas.
> _______________________________________________
> DAS mailing list
> DAS@biodas.org
> http://biodas.org/mailman/listinfo/das