Distributing and Visualising Biological Objects at Infobiogen

Eric Viara (4), Stuart Pook (1,3), Frédéric Guyon (1), Christophe Cussat-Blanc (1), Guy Vaysseix (1,2) and Emmanuel Barillot (1,2)

(1) GIS Infobiogen, 7 rue Guy Môquet - BP 8, 94801 Villejuif, France
(2) Généthon, 1 rue de l'Internationale, 91000 Evry France
(3) ENST, C109, Dpt Informatique, 46 rue Barrault, 75013 Paris, France
(4) Sysra Informatique, 7 rue de Bièvre, 92140 Clamart, France

Our aim is to design a complete object-oriented environment for biological data management. In that respect, Sysra Informatique and Infobiogen first developed an object-oriented database management system EyeDB with funding from Sysra Informatique, ANVAR and the EC BIOTECH program. Then we have implemented a complete Java mapping of the EyeDB object model.

We also designed a CORBA layer on top of EyeDB. An ad-hoc compiler can automatically generate an IDL and an ORB server from the database schema. A language allows to define a more elaborate mapping from the database schema to a target IDL. This allows to export different views from the same database. It also permits to target a common IDL from databases with different schemas.

We believe that the biological community should now adopt a standard to define the semantic of its objects. With colleagues at EBI (Patricia Rodriguez-Tomé et al), we have defined a standard description for the genome maps in the form of an IDL and we submit it to the bioinformatic community for adoption and comments. This is done through the Life Sciences Research Task Force of the Object Management Group.

Biological data now constitutes a huge amount of information with which it is difficult to interact. Particularly it is quite impossible to visualise these large spaces of information with the traditional display techniques. We propose to use new concepts of visualisation and navigation well suited for huge volume of data. We have implemented these new concepts in two applications:

a map viewer, (MappetShow) that gives a clear view of very dense maps (several hundreds of markers); it uses the following concepts: (multi-)focus+context techniques, dynamic layout, multi-scaling display.
a generic navigator for large information spaces: Zomit. It consists of a generic Java applet client and a C++ library used to develop a server. It has been used to implement a genome browser: ZoomMap. It is based on the concepts of zooming, semantic zooming, portals and magic lenses.

See http://www.infobiogen.fr/services/ for details.