Phylogenetic Analysis Workflows - User-Level Workflow Design: A Bioinformatics Perspective - page 74

are the use of the defined terms for semantic annotations of web services in
order to facilitate service discovery and integration, and the more detailed
description of the involved data types in order to improve the data exchange
between services.
In fact, a number of service providers (for instance the BioCatalogue [112]
and BibiServ [127]) have started to annotate their services in terms of EDAM,
and the latest release (6.4.0 of July 2011) of the European Molecular Biology
Open Software Suite (EMBOSS) [264] contains EDAM annotations for the
more than 400 tools of the suite and their parameters.
3.3.2 Domain Model
As any PROPHETS domain model, the domain model for this scenario con-
sists of type and service taxonomies, semantically annotated services, and
domain-specific workflow constraints. The domain-specific vocabulary that
is provided by the taxonomies is used by both service annotations and con-
straints for referring to the types and services in the domain. Accordingly, the
following introduces the type and service taxonomies that are used, before
dealing with service annotation and constraint formulation.
Taxonomi e s
The taxonomies for this example are derived from the EDAM ontology de-
scribed above, which provides terms for the classification of bioinformatics
services, data and resources. The examples presented in this topic are based
on the beta12 version of EDAM, which has been released in June 2011 and
is the EDAM version that has been used for annotating the EMBOSS 6.4.0
release on which this scenario is based. EDAM can simply be used for the
service and type taxonomies by taking the (relevant parts of the) ontology
and sorting the domain-specific service and data types into the skeletal do-
main structure. As detailed in [179], setting up a PROPHETS domain model
based on EDAM involves four major steps:
1. Converting EDAM from OBO (Open Biomedical Ontologies) [285] into
OWL [276] format.
2. Generating the service taxonomy from the Operation term and (transi-
tively) all its subclasses.
3. Generating the type taxonomy from the Data and Identifier terms and
(transitively) all their subclasses.
4. Sorting the available services and their input/output types into the ser-
vice and type taxonomy, respectively.
Steps 1-3 can be executed fully automatic. Step 4 can be automated whenever
EDAM annotations are available for the types and services, which is the case
for the EMBOSS tool suite that is used in this scenario.