Molecular information fusion in ondex

Abstract

Current biological knowledge is buried in hundreds of proprietary and public life-science databases available on the World Wide Web (WWW) and millions of scientific publications. Gaining access to this knowledge can prove difficult as each database may provide different tools to query or show the data and may differ in their structure and user interface or uses a different interpretation of biological knowledge than others. Systems approaches to biological research require that existing biological knowledge (data) is made available to support on the one hand the analysis of experimental results and on the other hand the construction and enrichment of models. Data integration methods are being developed to address these issues by providing a consolidated view of molecular information fused together from multiple databases. However, a key challenge for data integration is the identification of links between closely related entries in different life sciences databases when there is no direct information that provides a reliable cross reference. Here we describe and evaluate three data integration methods to address this challenge in the context of a graph-based data integration framework (the Ondex system). We give a quantitative evaluation of their performance in two different situations: the integration and analysis of different metabolic pathways resources and the mapping of equivalent elements between the Gene Ontology and a nomenclature describing enzyme function.