Information Technology Reference
In-Depth Information
comprising more than 300 annotated services and providing a simple web
client for the specification and execution of workflows.
Conceptually, the key feature of SADI is to regard services as “annotators”
that perform particular changes to the input data object and then return
it as output. Technically, input and output of a SADI service are defined
in RDF format in terms of OWL classes, and the behavioral semantics of
the service simply corresponds to the transformation that is performed to
get the output from the input OWL class. As a result of this, the OWL-
DL domain models can be regarded as abstract workflow specifications [348],
and standard OWL reasoners and SPARQL query engines can be applied
to compose adequate services automatically. In particular, it is possible to
describe the expected workflow outputs in terms of SPARQL queries, and
then let the system determine and execute the (sequence of) services that
transform the initial input to the final output.
Semantic Workflow Tool
The Semantic BioinformaticsServiceEnvironmentdescribedin [256] is another
approach that applies standard Semantic Web technologies for the description
and composition of services: It uses OWL-S [214] for semantic descriptions
of services, OWL for the definition of domain-ontologies, and a Description
Logic (DL) reasoner for the automated discovery of services. In addition, the
described environment comprises the Semantic Workflow Tool as a graphical
user interface for (semi-) automatic workflow design and workflow execution.
With the interface it is possible to construct service pipelines dynamically via
semantic discovery of matching services, that is, by searching for a services that
provide or use the input or output data of other services in the workflow.
The described environment focuses on Web Services as primary computa-
tional resource, and has been applied in the bioinformatics domain on pop-
ular publicly available Web Services. In order to provide a domain-specific
framework, the originally domain-independent OWL-S ontology has been ex-
tended with ontological concepts from the bioinformatics domain, and then
also used for the service annotations. In order to support interoperability and
reusability, the Semantic Workflow Tool provides a functionality for exporting
workflows in the Taverna workflow format.
8.1.2 Workflow Systems for Bioinformatics
Similarly, if not even more, interesting for comparison with Bio-jETI, how-
ever, are general-purpose workflow systems that have been designed for the
bioinformatics domain [128]. In particular, the bioinformatics community has
adopted the ideas of the Semantic Web earlier than other domains, and ac-
cordingly also systems for semantically guided workflow composition have
been around for several years now. Without making a claim to be complete,
the following gives an overview of the most popular approaches to workflow
systems that have originated in the bioinformatics community,
Search WWH ::
Custom Search