Jason Stewart, Open Informatics
Date: Wednesday, February 05
Time: 1:30pm - 2:15pm
Location: California Ballroom A & B
Stewart examines the MAGE project, and its relationship to ongoing efforts to standardize data communication in the field of gene expression informatics. The goal of the MAGE project is to develop an open source infrastructure for microarray gene expression data handling, and it currently includes three primary pieces:
- The MicroArray Gene Expression Object Model (MAGE-OM)
- The MicroArray Gene Expression Markup Language (MAGE-ML)
- The MAGE Software ToolKit (MAGEstk)
MAGE-OM is defined using the Unified Modelling Language (UML), and the
MAGEstk UML parser is used to generate the remaining pieces from the
UML model. MAGE-ML is an XML-based markup language based on a Document
Type Definition (DTD) created by the MAGEstk UML parser. Finally, the
UML parser creates a complete software API for Java, Perl, C++,
While MAGE-OM is specific for microarray gene expression data, the
MAGEstk framework has much broader use. The UML parser can read any
UML object model and create an XML DTD and a software API for either
the four supported languages.
Because of the growing use of Web services within bioinformatics (such as BioMOBY, MyGrid, OmniGene, and the I3C), MAGEstk can greatly enhance the ability of researchers to share data via the WWW.
What’s particularly interesting about Stewart’s presentation? “MAGE itself is a bit controversial, because it is an attempt to standardize scientific data, and scientists always buck at the thought of someone reining them in. Given that the major journals have recently agreed to require that all scientists must meet the MIAME requirements for data submission, people all over the world will have to begin using MAGE to submit their expression data.
“MAGE only handles one small corner of bioinformatics -- microarray gene expression data -- though, so MAGEstk framework is definitely the coolest piece, because it is general and can handle any kind of scientific data. It's timely and interesting because it provides an (almost) push-button mechanism to enable scientists to hook up their data to a Web service architecture and communicate their data over the net.
“People who need to export expression data for publication will need to learn about MAGE. People who have other kinds of data, but don't yet have a bioinformatics infrastructure to handle it, will want to hear about MAGEstk.”