Creating NoSQL Biological Databases with Ontologies for Query Relaxation

The complexity of building biological databases is well-known and ontologies play an extremely important role in biological databases. However, much of the emphasis on the role of ontologies in biological databases has been on the construction of databases. In this paper, we explore a somewhat overlooked aspect regarding ontologies in biological databases, namely, how ontologies can be used to assist better database retrieval. In particular, we show how ontologies can be used to revise user submitted queries for query relaxation. In addition, since our research is conducted at today's “big data” era, our investigation is centered on NoSQL databases which serve as a kind of “representatives” of big data. This paper contains two major parts: First we describe our methodology of building two NoSQL application databases (MongoDB and AllegroGraph) using GO ontology, and then discuss how to achieve query relaxation through GO ontology. We report our experiments and show sample queries and results. Our research on query relaxation on NoSQL databases is complementary to existing work in big data and in biological databases and deserves further exploration.