Database Reference
In-Depth Information
13.6.1 Workflow Reuse
Workflow reuse is effective at multiple levels: the scientist reuses a workflow
with different parameters and data, and may modify the workflow as part
of the routine of their daily scientific work; workflows can be shared with
other scientists conducting similar work, so they provide a means of codifying,
sharing, and thus spreading the workflow designers' practice; and workflows,
workflow fragments, and workflow patterns can be reused to support science
outside their initial application.
The latter point illustrates the tremendous potential for new scientific ad-
vances. An example of this is a workflow used to help identify genes involved
in tolerance to trypanosomiasis in east African cattle.
77
The workflow was
initially successful because it enabled data to be processed systematically
without a need for manual triage. This same workflow was then reused over
a new dataset to identify the biological pathways implicated in the ability
for mice to expel the
Trichuris muris
parasite (a parasite model of the hu-
man parasite
Trichuris trichuria
). This reuse was made easier by the explicit,
high-level nature of the workflow that describes the analytical protocol.
Workflows bring challenges too. Realistic workflows require skill to produce,
and therefore they can be dicult and expensive to develop. Consequently,
workflow developers need development assistance, and prefer not to start from
scratch. This is another incentive for reusing workflows. Unfortunately it is
easy for the reuse of a workflow to be confined to the project in which it
was conceived. In the trypanosomiasis example, the barrier to reuse was how
the knowledge about the workflow could be spread to the scientists with the
potential need. In this case it was word of mouth within one institution; this
barrier needs to be overcome.
Workflow management systems already provide basic sharing mechanisms,
through repository stores for workflows developed as part of projects or com-
munities. For example, the Kepler actor repository is an LDAP-based direc-
tory for the remote storage, query, and retrieval of actors (processes) and other
workflow components. Similarly, the Southern California Earthquake Center
(SCEC) uses component and workflow libraries annotated with ontologies.
78
These follow the tradition of cataloging scripting libraries and codes. InforS-
ense's online Customer Hub and workflow library
*
allow users to share best
practices and leverage community knowledge potentially across projects.
13.6.2 Social Sharing
The myExperiment project
†
is taking a more social approach, recogniz-
ing the use of workflows by a community of scientists.
79
This acknowledges
*
http://www.inforsense.com/pdfs/InforSense WorkflowLibrary DataSheet.pdf. Accessed on July
20, 2009.
†
http://www.myexperiment.org/. Accessed on July 20, 2009.
