Research activity in the life science area is becoming increasingly data intensive. Huge amounts of highly heterogeneous data, including high throughput experiment results, publication collections, and clinical records are generated at a fast pace by researchers all over the world. The capability of correlating heterogeneous information stored in separated data repositories is a compelling, yet currently unsatisfied, need for bioinformatics scientists. Recently developed systems address this issue by building knowledge repositories for specific bioinformatics sub-domains such as protein-protein interaction or array expression analysis. In this paper we present an overview of heterogeneous biological data sources and discuss the many difficulties faced by biological data querying and analysis.