Complex Query Formulation Over Diverse Information Sources in TAMBIS

Biologists increasingly need to ask complex questions over the large number of data and analysis tools that are available on the Internet. To do this, the individual resources need to be made to work together. The knowledge needed to accomplish this, for example about the locations of the sources and their capabilities, places barriers between biologists and the questions they would like to ask. The TAMBIS project (Transparent Access to Multiple Bioinformatics Information Sources) has sought to remove some of these barriers, thereby making the process of asking questions against multiple sources more straightforward. Central to the TAMBIS system is an ontology of bioinformatics and biological terms. Users express retrieval requests in terms of the concepts and relationships described in the ontology, rather than by making direct reference to individual sources. This allows TAMBIS to be used to formulate rich, declarative queries over multiple sources. The ontology is constructed in a manner that ensures only biologically meaningful queries can be posed. User’s queries are constructed using an interactive ontology browsing and query construction tool, and are rewritten by a query planner for evaluation using a wrapper layer. This paper provides an overview of the TAMBIS approach to source integration, focusing on the way the ontology is used to support query formulation and refinement.