Related Work - User-Level Workflow Design: A Bioinformatics Perspective - page 162

the Departments of Biology and Mathematics and Computer Science at
Emory University. Recently it has predominantly been used for genomic re-
search (cf., e.g., [156, 46, 44]).
The Galaxy web interface provides a rich (and extensible) collection of
computational tools for various general and bioinformatics tasks and prede-
fined access methods to common bioinformatics databases. Upon workflow
execution, Galaxy automatically creates and stores provenance information
and metadata about the executed tools and workflows. As an open web plat-
form, it furthermore functions also as community space for sharing workflows,
data, and results, and provides the possibility to equip tools and workflows
with tags and annotations.
jORCA
jORCA [215, 147], developed at the University of Malaga (Spain), addresses
interoperability issues of bioinformatics web services that arise from the use of
incompatible data and communication formats. Therefore it combines the in-
tegration of heterogeneous services into one uniform framework with service
discovery and automatic workflow composition functionality. jORCA inte-
grates web services from a number public repositories, such as the BioMoby
[341], DDBJ [165] and EBI [250, 167, 118] web services. As usability is a
major goal of jORCA, its user interface is conceived as desktop client that
allows users to “compose, edit, store, export, and enact workflows” [147].
The automatic workflow composition functionality is based on the Mag-
allanes system [265], which uses a breadth-first pruning algorithm (cf. [85])
for finding a shortest path from source (i.e., input data type) to target (i.e.,
output data type). Thus, jORCA can create input-output pipelines automat-
ically, but in contrast to the temporal-logic synthesis methods applied by
PROPHETS, it is neither able to handle additional constraints, nor to use a
shared memory for more complex data transfer.
Kepler
Kepler [28] has been designed as a system for speeding scientific workflows in
the scope of a cross-institution project at UC Davis, UC Santa Barbara, and
UC San Diego. It has been applied to different domains, such as astronomy,
ecology, geology and biology (cf., e.g., [53, 126]).
Extending the Ptolemy II system [91] for heterogeneous, concurrent mod-
eling and design, one of Kepler's most notable characteristics is its support
for different models of computation for the workflow models. For instance,
synchronous data-flow enactment can be chosen as the adequate model of
computation for simple data transformation workflows, whereas continuous-
time enactment is suitable for time-dependent workflows where the model is
described in terms of differential equations. Workflow development in Kepler