Stable and fast federated search for integration of distributed knowledge of life science data

In this study, in order to stably and efficiently use life science data provided by SPARQL endpoints distributed on the Web, we developed a federated search system for queries that trace the relationships between classes. To do so, by using information of classes and class-class relationships obtained by crawling each SPARQL endpoint, a federated class graph is constructed. We developed a federated search system that enables efficient searches across SPARQL endpoints by using the information of the path on the federated class graph.