The amount of Arabidopsis data being generated, especially by high
throughput methods, continues to grow at an alarming rate. New web sites
are appearing to host new data sets and analyses are proliferating.
However, most existing and new web sites give little or no consideration
to effective integration of data from multiple sources. We, as a
community, should take more seriously the challenges of data integration
so as to more effectively discover and exploit such data sets.
Web services provide a means whereby data residing at many different
locations can be seamlessly integrated to provide the user with richer
data sets. Unlike web pages that are idiosyncratic in their layout and
content, and must be visited by a researcher one at a time, web services
(in this case BioMOBY) provide data in well structured and agreed
formats, document their availability through a central registry and can
be combined to provide richer views of the data. Under a demonstration
project encouraged by the Multinational Arabidopsis Steering Committee
(MASC) that oversees and attempts to integrate functional genomics
efforts world-wide, our goal was to facilitate the deployment of web
services from 16 data providers (eight each in Germany and in the USA)
that vary both in the kinds of data that they host, and in the level of
expertise at the sites. Training workshops for developers have been held
both at the Max Planck Institute for Plant Breeding Research in Cologne
and The Institute for Genomic Research in Rockville MD that have both
provided the attendees with the knowledge needed to deploy web services
and have resulted in new services.
For information on the project, we invite you to visit our web site
(http://bioinfo.mpiz-koeln.mpg.de/araws), where you will find extensive
"how to" documentation developed in conjunction with our workshops that
is freely available to the community.
We also invite you to check out two "aggregator" web clients that can
simultaneously invoke services from multiple locations in search of a
particular data type (e.g. literature or images). These are "baby steps"
towards the concept of "one-stop shopping" for specific data sets.
The Arabidopsis Image Gallery (tAIGa;
http://mips.gsf.de/proj/planet/araws/tAIGaSearch.html) takes an AGI
locus identifier as input and queries seven different locations for
images corresponding to this locus.
<http://mips.gsf.de/proj/planet/araws/litRepSearch.html> ) an AGI locus
identifier as input and returns PubMed identifiers queried from a number
of different sources including TAIR and Aramemnon.
We'd appreciate your comments on these pilot services and encourage you
to consider becoming actively involved in the distribution of
Arabidopsis data via web services.
Thanks for your attention
The Institute for Genomic Research
9712 Medical Center Drive
Rockville, MD 20850
Office Phone: 301-795-7523
To page me at TIGR: 301-795-7000
Home Phone: 301-990-0878
Cell Phone: 301-204-6300