OpenBSA - Tools and standards for distributed computing
=======================================================
Juha Muilu(*, Martin Senger and  Alan Robinson

European Bioinformatics Institute
Wellcome Trust Genome Campus
Hinxton, Cambridge CB10 1SD, UK

(*Presenting author


Background
----------

Biomolecular sequence analysis specification (BSA) is one of the first
standardised API's for bioinformatics community developed in
collaboration with commercial and academic organisations.  The
specification was adopted in Dec 1999 by the Object Management Group
(OMG).

The BSA specification comprises two independent modules described
using the CORBA IDL:

The Analysis module defines API for executing and controlling remote
analysis applications as well as XML metadata, which specifies
information needed to run the applications i.e. how to assign input
parameters and data, and how to adapt the results to a format
understood by the client. The metadata is also available to clients
for use in making user interfaces etc.

The analysis module is independent of data format used by the
applications. The results are sent to client as a strings by default,
but the results can be adapted to CORBA objects as well. The
BioObjects module of the BSA specification contains definitions for
certain basic biological entities like DNA/AminoAcid sequence,
alignment and similarity search results, that can be used with the
analysis server.


OpenBSA
-------

In the talk, we will present recent developments in the OpenBSA
project, which provides an open source implementation of the BSA
specification along with some client side Java components for
accessing analysis services, pipelining and viewing the results.

The core of the project is a Java based package called AppLab, which
implements the analysis part of the specification. AppLab takes care
of activating analysis instances and sending progress events, results
and/or intermediate data back to the client. The AppLab package also
includes tools for converting the EMBOSS and GCG application
description metadata to the XML format understood by the BSA
compatible server.  This provides immediate access to dozens of
different sequence analysis applications.

Modular architecture of the AppLab and the OpenBSA client components
allows to utilise existing biocomputing frameworks easily. Integration
of the bioCORBA, bioPerl and bioJava components will be discussed.

BSANE
-----

In the talk we will also describe possible modifications to the
BioObjects module to make it more compatible with the other open
source efforts, especially with the bioCORBA. These modifications will
be done within the "Biomolecular Sequence Analysis Entities" (BSANE)
specification requested by the LSR.


Links
-----

OMG:   http://www.omg.org
LSR:   http://www.omg.org/lsr
BSA    http://www.omg.org/cgi-bin/doc?dtc/00-11-01
BSANE: http://industry.ebi.ac.uk/~muilu/OMG/BSANE/