A journey to Semantic Web query federation in the life sciences

Kei-Hoi Cheung 2 H Robert Frost 1 M Scott Marshall 0 Eric Prud'hommeaux 6 Matthias Samwald 4 5 Jun Zhao 3 Adrian Paschke 7 0 Informatics Institute, University of Amsterdam , The Netherlands 1 VectorC, LLC , Hanover, NH 03755 , USA 2 Center for Medical Informatics, Yale University School of Medicine , New Haven, CT 06511 , USA 3 Department of Zoology, University of Oxford , Oxford, OX1 3PS , UK 4 Konrad Lorenz Institute for Evolution and Cognition Research , Altenberg , Austria 5 Digital Enterprise Research Institute, National University of Ireland Galway , IDA Business Park, Lower Dangan, Galway , Ireland 6 World Wide Web Consortium, Massachusetts Institute of Technology , Massachusetts, MA 02139 , USA 7 Freie Universitat Berlin , Germany Background: As interest in adopting the Semantic Web in the biomedical domain continues to grow, Semantic Web technology has been evolving and maturing. A variety of technological approaches including triplestore technologies, SPARQL endpoints, Linked Data, and Vocabulary of Interlinked Datasets have emerged in recent years. In addition to the data warehouse construction, these technological approaches can be used to support dynamic query federation. As a community effort, the BioRDF task force, within the Semantic Web for Health Care and Life Sciences Interest Group, is exploring how these emerging approaches can be utilized to execute distributed queries across different neuroscience data sources. Methods and results: We have created two health care and life science knowledge bases. We have explored a variety of Semantic Web approaches to describe, map, and dynamically query multiple datasets. We have demonstrated several federation approaches that integrate diverse types of information about neurons and receptors that play an important role in basic, clinical, and translational neuroscience research. Particularly, we have created a prototype receptor explorer which uses OWL mappings to provide an integrated list of receptors and executes individual queries against different SPARQL endpoints. We have also employed the AIDA Toolkit, which is directed at groups of knowledge workers who cooperatively search, annotate, interpret, and enrich large collections of heterogeneous documents from diverse locations. We have explored a tool called FeDeRate, which enables a global SPARQL query to be decomposed into subqueries against the remote databases offering either SPARQL or SQL query interfaces. Finally, we have explored how to use the vocabulary of interlinked Datasets (voiD) to create metadata for describing datasets exposed as Linked Data URIs or SPARQL endpoints. - Conclusion: We have demonstrated the use of a set of novel and state-of-the-art Semantic Web technologies in support of a neuroscience query federation scenario. We have identified both the strengths and weaknesses of these technologies. While Semantic Web offers a global data model including the use of Uniform Resource Identifiers (URIs), the proliferation of semanticallyequivalent URIs hinders large scale data integration. Our work helps direct research and tool development, which will be of benefit to this community. Background As the number, size, and complexity of life science databases continue to grow, data integration remains a prominent problem in the life sciences. These disparate databases feature diverse types of data including sequences, genes, proteins, pathways, and drugs produced by different kinds of experiments, including those that involve high-throughput technologies such as DNA microarray, mass spectrometry, and next generation sequencing. The challenges involved in integrating such data include inconsistency in naming, diversity of data models, and heterogeneous data formats. The benefits of integrating these disparate sources of data include discovery of new associations/relationships between the data and validation of existing hypotheses. Numerous life science databases can be accessed publicly via the Web. The data retrieved from different databases are displayed using the HyperText Markup Language (HTML) and rendered by Web browsers (e.g., Internet Explorer and Firefox). Hypertext links are used to connect data items between different Web pages. Data integration using hypertext links, however, is burdensome to the user [1]. HTML works well to expose the results of scripted (canned) queries but does not expose the database structure to data users who would wish to construct their own queries. To automate integration of data in HTML format, we need to rely on methods such as screen scraping to extract the data from the HTML documents and integrate the extracted data by custom scripts. This approach is vulnerable to changes in display and location of Web pages. Such changes, together with changes in database structure, significantly increase the code complexity of data integration. To address this problem, approaches have been developed to facilitate data integration on a larger scale. Some representative approaches include EBI SRS [2], Atlas [3], DiscoveryLink [4], Biokleisli [5], Biozon [6], etc. In general, these approaches fall into two categories: data warehouse and federated database. The data warehouse approach relies on data translation in which data from different databases are re-expressed in a common data model on a central repository. The federated approach features query translation in which data are kept in their local databases and a global query can be translated into a set of local database subqueries whose results are unified and presented to the user. There are pros and cons for each approach. Data warehouses typically wrestle with the concurrency issue (keeping the data up-to-date with respect to a data source). Each time a member database is changed, the data translation code will need to be modified and/or re-executed, depending on the nature of the change. On the other hand, data warehouse query performance is good because queries are run locally. In the federated approach, data concurrency is not an issue, but query speed may be slow, especially when large amounts of data are transferred over the network. The Semantic Web [7] transforms the Web into a global database or knowledge base by providing: i) globally unique names through the Uniform Resource Identifiers (URIs), ii) standard languages including the Resource Description Framework (RDF), RDF Schema (RDFS), and the Web Ontology Language (OWL) for modeling data and creating ontologies, and iii) a standard query language SPARQL [8]. Enabling technologies such as ontology editors (e.g., Protg), OWL reasoners (e.g., Pellet and FaCT++) and triplestores with SPARQL endpoints (e.g., Virtuoso, AllegroGraph and Sesame) help make the Semantic Web vision a reality. While these core and enabling technologies are maturing, there are new technological developments that can help push the Semantic Web to a new level of data interoperability. For example, Linked Data [9] is (...truncated)