Ontologies for Description, Discovery and Integration of Genomic Data Sources State of The

Sources: State of The Art and Propositions

M.-D. Devignes, N. Boudjlida and M. Smaïl

Laboratoire Lorrain de Recherche en Informatique et ses Application, CNRS, INRIAUniversité Henri Poincaré, BP 239, 506 VANDOEUVRE LES NANCY Cedex – France

malika@loria.fr

Abstract

Considering the variety and the heterogeneity of genomic data sources we figure an almost ideal situation wherethe following facilities are provided: (i) one can easily locate existing data sources and know about their content,the services they offer, the way to access the data, (ii) one can easily identify and select among data sourcesthose that can contribute to satisfy one's need, (iii) one can easily describe a navigation scenario across selecteddata sources and execute this scenario in order to extract and combine information coming from these sources[1].

Steps towards this situation have been experienced in our previous work on user-oriented systems for collectingand integrating genome information [2, 3]. Generalization of this work may consist in an open infrastructurewhere data sources can be described, compared and accessed. Catalogs of genomic data sources already exist butare difficult to maintain and exploit. Indexing source description with elements of a domain ontology wouldallow to guide navigation across data sources in relation to a given need or view of the domain.

Existing ontologies and/or projects for such purpose will be reviewed. We propose to set up an architecturebased on a registry. Documents stored in the registry would be composed of source metadata such as type ofcontent, relationship with other sources, access to the data, offered services. Metadata would be indexed withone or more domain ontologies. The architecture would enable navigation across the documents according tovarious points of view. A navigation scenario, aimed at collecting the data that answer a specific question, couldbe constituted dynamically upon interactive navigation. Besides the numerous difficulties of the project, theproblem that would be useful to discuss in the workshop is the choice or creation of the ontologie(s) that couldbe used to index source descriptions and guide users in their navigation.References

[1] N. Boudjlida, M.-D. Devignes and M. Smaïl, Services for a genomic open distributed environment, in XML-EnabledWide Area Search in Bioinformatics Workshop, League City, TX, December 2000.

[2] M.-D. Devignes, A. Schaaff and M. Smaïl, Collecte et intégration de données biologiques hétérogènes sur le web –Xmap, application dans le domaine de la cartographie du génome humain. Ing. Syst. Info. 7, 45-61, 2002.

[3] M.-D. Devignes, Y. Norsa, M. Smaïl, P. Collet and L. Domendjoud, From putatives promoter sequence to genomiccontext: biological data collection on the web using a generic application (Xprom), Poster at the First European Conferenceon Computational Biology, 2002.