19th International CODATA Conference
Category: Interoperability

Combining Applications and Databases Integration Approaches in a Common Distributed Genomic Platform

Pierre-Emmanuel GROS (gros@limsi.fr), Joan HÉRISSON (herisson@limsi.fr), Nicolas FEREY (ferey@limsi.fr), and Rachid GHERBI (gherbi@limsi.fr)
LIMSI-CNRS, Université Paris-Sud, France
http://www.limsi.fr

Many databases and software tools are currently available for accessing genomic data. In general, each tool uses a different type of data in exchange protocols, and usually they offer specific services. Our system approach, named GenoMEDIA, addresses the problem of integrating these heterogeneous databases and software utilities within a generic distributed platform. GenoMEDIA provides two specific middleware, named Lydia and Antje, for this integration. Lydia offers facilities for working simultaneously with a variety of Services and Databases; Antje is designed to allow users to manage multiple heterogeneous remote databases in a uniform vision. The aim of this paper is to present GenoMEDIA and how heterogeneous databases and remote services are integrated, in particular how Antje was designed, implemented and tested with various genomic databases. Lydia is used to call a service through a RDBMS and Antje allows users to manage multiple remote databases through a single interface. The GenoMEDIA architecture allows a user to send an augmented unified query through standard SQL that involves remote heterogeneous databases and services. Unlike GeneWeaver, GenoMEDIA uses this SQL approach for managing existing DBMS technologies (PHP, ODBC,..). Like in Oracle's distributed database architecture, Antje’s goal is to facilitate access to databases on multiple computers by making them appear as if they were a single file available through a user-friendly interface. Antje uses the notion of remote view instead of the notion of database link. Users do not have to use the database link into the SQL statements.In this paper, we will present shortly the Lydiaconcept, our service integration middleware. This will allow the reader to better understand its connections with Antje on the GenoMEDIA platform. The major part of the paper will discuss how Antje was designed, implemented and tested in efforts to unify remote biological databases. Finally , we will present our integration graphic user interface for Antje; AntjeViewer. The main goal of AntjeViewer is to allow users to easily integrate databases without any lines coding of SQL.