Big linked cancer data

The amount of bio-medical data available on the Web grows exponentially with time. The resulting large volume of data makes manual exploration very tedious. Moreover, the velocity at which this data changes and the variety of formats in which bio-medical data is published makes it difficult to access them in an integrated form. Finally, the lack of an integrated vocabulary makes querying this data more difficult. In this paper, we advocate the use of Linked Data to integrate, query and visualize bio-medical data. The resulting Big Linked Data allows discovering knowledge distributed across manifold sources, making it viable for the serendipitous discovery of novel knowledge. We present the concept of Big Linked Data by showing how the constant stream of new bio-medical publications can be integrated with the Linked Cancer Genome Atlas dataset (TCGA) within a virtual integration scenario. We ensure the scalability of our approach through the novel TopFed federated query engine, which we evaluate by comparing the query execution time of our system with that of FedX on Linked TCGA. Then, we show how we can harness the value hidden in the underlying integrated data by making it easier to explore through a user-friendly interface. We evaluate the usability of the interface by using the standard system usability questionnaire as well as a customized questionnaire designed for the users of our system. Our overall result of 77 suggests that our interface is easy to use and can thus lead to novel insights.