Bio2RDF Network Of Linked Data
Marc-Alexandre Nolin
1-4
, Peter Ansell
2
, François Belleau
1
, Kingsley Idehen
3
,
Philippe Rigault
1
, Nicole Tourigny
4
, Paul Roe 2, James M Hogan 2, Michel
Dumontier
5
1
Centre de recherche du CHUL, francoisbelleau@yahoo.ca , marc-
alexandre.nolin@genome.ulaval.ca , philippe.rigault@genome.ulaval.ca ,
2
Queensland
University of Technology, p.ansell@qut.edu.au , p.roe@qut.edu.au , j.hogan@qut.edu.au ,
3
OpenLink Software, kidehen@openlinksw.com ,
4
Université Laval,
nicole.tourigny@ift.ulaval.ca ,
5
Carleton University, michel_dumontier@carleton.ca
Abstract. Background: The Bio2RDF project (http://bio2rdf.org) work to
create a network of coherent linked data across the life sciences databases. The
project is open source and the following result is from the input of this
community. Results: Databases have been converted and linked together with
semantic web technologies. The process is to normalize any external URIs in
each databases while we do the conversion from the original format to RDF.
Conclusion: Each converted databases have his own SPARQL point provided
by a Virtuoso Triplestore. Every documents can be retrieve using a REST
interface similar for any databases. The REST URL also happen to be the URI.
Other tools are also available using the REST interface.
Keywords: Linked data, semantic web, URI normalization
1 Introduction
In the State of the nation in data integration for bioinformatics, Goble and Stevens
(2008) advised the bioinformatics community to address the issue of identity and
naming, a necessary condition to facilitate data integration. As part of the Bio2RDF
project to build an integrated bioinformatics warehouse on the semantic web,
resources have been assigned Universal Resource Identifiers (URIs) that are
normalized around the bio2rdf.org namespace (Belleau, F., et.al., 2008). Bio2RDF
has created an RDF warehouse that serves over 70 million triples describing the
human and mouse genomes (Belleau, F., et.al, DILS2008 in press).
The Linked Data
1
initiative aims to make possible the browsing of information on the
semantic web, and follows four basic rules. First, resources should be named with
Universal Resource Identifiers, or URIs. Second, HTTP URIs are ideal as they
provide ownership and resolution. Third, information should be resolvable on the
web, and fourth, it is necessary to connect data on the web. Building a critical mass of
linked data on the web is a crucial goal for the semantic web. DBpedia (Auer, S.,
1
http://www.w3.org/DesignIssues/LinkedData.html