=Paper=
{{Paper
|id=Vol-1795/paper39
|storemode=property
|title=Updates from the EMBL-EBI RDF platform
|pdfUrl=https://ceur-ws.org/Vol-1795/paper39.pdf
|volume=Vol-1795
|authors=Thomas Liener,Tony Burdett,Simon Jupp
|dblpUrl=https://dblp.org/rec/conf/swat4ls/LienerBJ16
}}
==Updates from the EMBL-EBI RDF platform==
<pdf width="1500px">https://ceur-ws.org/Vol-1795/paper39.pdf</pdf>
<pre>
         Updates from the EBI RDF platform

               Thomas Liener1 , Tony Burdett1 , and Simon Jupp1

 EMBL-EBI, Wellcome Genome Campus, Hinxton, Cambridgeshire, CB10 1SD, UK
         tliener@ebi.ac.uk, tburdett@ebi.ac.uk, jupp@ebi.ac.uk


    The EBI RDF platform[1] was released in 2013 and provides access to several
EBI databases as RDF Linked Data. The RDF platform provides a SPARQL
endpoints and a Linked Data browser to the data in addition to making data
dumps available over FTP. At release the RDF platform1 contained data from
the Reactome, BioModels, BioSamples, Gene Expression Atlas and ChEMBL
databases in addition to the collaboration with Uniprot to deliver Uniprot RDF.
Since the release, additional EBI databases have begun generating RDF ex-
ports that will be made available via the RDF platform. These include the
Ensembl core data2 and the text-mined data from Europe PubMedCentral3 .
Additional datasets in preparation are the Genome Wide Association Studies
(GWAS) dataset4 , a new version of the Expression Atlas data and the complete
set of ontologies from the Ontology Lookup Service (OLS)5 .

    This poster will present the recent updates to the platform and our plans
for a consolidated infrastructure to deal with the increasing size and demand on
the platform. In a bid to improve the performance of queries that span multiple
datasets and provide a consolidated view over the EBI data, we are moving to
provide a single SPARQL endpoint to access all the EBI RDF data. We have
developed a series of scripts for the automated build and deployment of RDF
datasets based on metadata described using the HCLS dataset description best
practice guidelines6 . Over the coming year we will expose analytics of com-
mon query patterns (e.g. percentage of queries that combine data from multiple
datasets, most heavily accessed concepts and common predicates) that will bring
new insight into the usage of the data. We will use these analytics to deliver an
enhanced user interface with improved search and browsing capabilities along
with smart visualisations that reflect common usage patterns.

References
[1] Jupp et al. The EBI RDF platform: linked open data for the life sciences, Bioin-
  formatics (2014) 30 (9): 1338 - 1339

1
  http://www.ebi.ac.uk/rdf/
2
  https://www.ebi.ac.uk/rdf/services/ensembl/
3
  https://europepmc.org/
4
  https://www.ebi.ac.uk/gwas
5
  https://www.ebi.ac.uk/ols
6
  https://www.w3.org/TR/hcls-dataset

</pre>