From Bioinformatic Web Portals to Semantically Integrated Data Grid Networks
We propose a semi-automated method for redeploying bioinformatic databases indexed in a Web portal as a decentralized, semantically integrated and service-oriented Data Grid. We generate peer-to-peer schema mappings leveraging on cross-referenced instances and instance-based schema matching algorithms. Analyzing real-world data extracted from an existing portal, we show how a rather trivial combination of lexicographical measures with set distance measures yields surprisingly good results in practice. Finally, we propose data models for redeploying all instances, schemas and schema mappings in the Data Grid, relying on standard Semantic Web technologies.