Towards Semantic Harmonization of Biobanks 'beyond the obvious'
In this pilot project we apply existing advanced Semantic Web technologies and innovative biosemantics Web Services for 'harmonizing' 7 Dutch BioBanks, and an additional 6 from Groningen for reference. Via semantic search services and mining technologies, biobank researchers will be able to compile information about BioBanks relevant to their research questions. For instance to find peers with relevant expertise or data, or groups of BioBanks best suited for setting up a new collaboration. Our aim is that BioBank researchers can find answers to questions such as 'which BioBank or researcher has the right type of data suitable for my experiment?'. This pilot program will make extensive use of existing automated technologies that obviate the need for the formulation of top-down attempts to enforce standards of data formats and collection procedures that can often frustrate attempts at data harmonization. The semantic capabilities and web services we propose will provide intuitive and real-time access to data, using data descriptions and other textual information in the BioBank while at the same time maintaining owner-specified information security. Importantly, a semantic web approach will also enable queries going beyond retrieval of explicit data, and will open the door to automated, statistical approaches to ‘inference’, whereby terms in the descriptions of BioBanks and terms mined from vast literature corpora can be associated using concept profiles derived from existing biomedical literature and databases via ontologies. In turn, this will open the door to the discovery of important and otherwise unknowable relations between BioBank holdings, researchers, diseases, epidemiological trends and hosts of other relevant information: searching 'beyond the obvious' The key objectives of this pilot program is (1) the identification of infrastructural requirements for scaling semantic BioBank harmonization to a nationwide and international scope and (2) working closely with BioBank users to develop, test, and refine Web Service applications to meet the needs of practicing biomedical researchers. Our approach will complement applications that require a top-down standardization procedure.