=Paper=
{{Paper
|id=Vol-1114/Demo_Iersel
|storemode=property
|title=Linked Data Services for Scientists Exemplified Using Drosophila Melanogaster Datasets
|pdfUrl=https://ceur-ws.org/Vol-1114/Demo_Iersel.pdf
|volume=Vol-1114
|dblpUrl=https://dblp.org/rec/conf/swat4ls/IerselAR13
}}
==Linked Data Services for Scientists Exemplified Using Drosophila Melanogaster Datasets==
<pdf width="1500px">https://ceur-ws.org/Vol-1114/Demo_Iersel.pdf</pdf>
<pre>
    Linked Data services for Scientists exemplified using
            Drosophila melanogaster datasets

                        M.P. van Iersel, N. Anwar, L. Reynolds

                          General Bioinformatics, Reading, UK




1      Introduction


   Semantic technologies allow scientists to rapidly integrate data from experimental
datasets and online databases. On the computational side, the massive scale of these
datasets poses some challenges. We find that there are additional challenges on the
human side. Problems we encounter in practice include translating human questions
to SPARQL queries, adjusting data models and ontologies to match the problem space
of scientists, visualising data, and providing user-friendly access.

   To address these issues, General Bioinformatics applies the following principles.
Open source tools such as PathVisio and Cytoscape are adopted to provide
visualisation options. Custom inferences layers are added to semantic data to address
research questions. For example, we apply inferences on top of BioPAX pathways to
enable the extraction of a bipartite biochemical reaction network with a single
SPARQL query. Finally, we provide training materials for scientists for visualisation
tools, ontologies, and SPARQL.

   Our linked data system is supports life science researchers to answer research
questions. For example, the system can gather supporting evidence from a variety of
databases to identify interesting target proteins. Our public demo focuses on
Drosophila, but the same principles apply to any model organism.


2      Results

   We have integrated several Drosophila melanogaster datasets and built a public
showcase. This showcase integrates diverse data types such as genes, proteins,
FlyAtlas expression data and BioPAX pathways. This data is stored in a triple store
running in a cloud instance. We provide several options for visualizing the data. First,
a custom Cytoscape plugin can extract triple data and visualize it in network form.
Secondly, a PathVisio plugin can be used to create visualisations on top of pathway
diagrams derived from WikiPathways (See Fig. 1).
   By following the demo, you should get a clear idea how linked data can be used to
integrate diverse Drosophila datasets, how this can be visualized in pathway and
network context, and how this can help to answer research questions.




   Fig. 1: a schematic overview of the integrated Drosophila data. A collection of
  databases (top) have been transformed into RDF and integrated using standard
ontologies (middle). Two different visualisation tools are provided to create different
                                 views oft he data.


3     Availability

   The integrated data, SPARQL tutorial and videos, RDF downloads, cytoscape
session       and    live     SPARQL       endpoint are all   available    at
http://fly.cloud.generalbioinformatics.com/

</pre>