Figure 1

Architecture of our prototype system comprising of different levels. The figure shows the architecture of the prototype consisting of different levels (0). The resource level, provides and abstraction for elements (E1...En) of public biomedical resources (such as GEO and PubMed). The annotation level, uses a concept recognition tool called mgrep (developed by Univ. of Michigan) to annotate (or tag) resource elements with terms from a dictionary; constructed by including all the concept names and synonyms from a set of ontologies accessible to the ontology level. The annotation tables contain information of the form "element E was annotated with concept T in context C". At the index level, a global index combines all the annotations to index them by ontology concepts and provides information for the form "Concept T annotates elements E1, E2,..., En". See main text for details.