Support for Cooperative Experiments in e-Science: From Scientific Work flows to Knowledge Sharing - Identification of Ligand Binding Site and Protein-Protein Interaction Area - page 137

Disseminate
Results

Design
experiment

Perform
Experiment

Analyze
Results

Emphasis of current SWMS user support

Scope of e-Science

Fig. 7.2 e-Science experiment phases with indications of current and desired support. The gray-
scale represents the focus of the existing workflow management systems: most of the support
currently offered is associated with the execution and result analysis phases ( black bar ).
Dissemination and design phases are less frequently supported ( gray bars )

following a workflow approach, the latter encode the logic of the experimentation
processes and become an important resource to promote knowledge transfer among
scientists. A workflow represents a reliably repeatable sequence of tasks composing
an e-Science application. It describes the pattern of activity enabled by a systematic
organization of resources (Taylor et al. 2007 ). One aspect of e-Science in particular,
the sharing of resources, places greater demands than usual on the experiment
validation methodology. The reason for this is that scientists conducting these
complex experiments often do not have the required expertise to solve all the prob-
lems facing them. It is common that they use third-party components and therefore
need extra assurance to make sure that they are using these components in the proper
way and these third party components are behaving as expected. There are different
views on how the e-Science lifecycle can be defined. A commonly accepted
definition is that the development of a scientific experiment has different activities
or tasks performed at different times (Jacobs and Humphrey 2004 ; Humphrey and
Hamilton 2004 ). These activities belong to different phases of a typical e-Science
application lifecycle, which can be divided into four phases: Design , Execution ,
Analysis , and Dissemination (Fig. 7.2 ).

Design : an iterative process which requires discovering the resources that can be
used, for instance through semantic search. Interoperation between the discovered
resources needs to be established and a proper methodology associated with these
resources needs to be considered. This problem becomes more challenging in the
context of scientific experimentation where the scientists are continuously defining
hypotheses, collecting data, running experiments, revising hypotheses, and publish-
ing results. Multidisciplinary and geographically distributed teams of scientists
need to be able to locate, construct, execute and maintain such workflows (De Roure
et al. 2007 ). To design a successful workflow, modeling and composition tools are