Biomedical Engineering Reference
In-Depth Information
project develops standard terminologies and a formal, extensible, knowledge-
based description of biological assays. Bio-curation is also an integral compo-
nent of the BAO project to systematically annotate PubChem bioassays with
standardized terminology describing assay concepts. The BAO project also
develops software tools to query and explore a data set in the context of the
ontology.
BAO follows an established ontology development methodology using a
combination of a top - down domain expert - driven and bottom - up data - driven
approach [60]. The scope of the ontology, knowledge acquisition, software
requirements, and specifi cations were driven by use-case scenarios presented
to domain experts (workshop given at the Society for Biomolecular Sciences,
Phoenix 2010 [60]) and by derived competency questions. The ontology will
unify knowledge-related HTS and other types of screening, including the
concepts described below. One initial goal of the BAO ontology and software
is to enable researchers to query the bioassay repositories and to retrieve
relevant data. For example, in the case of PubChem, even seemingly trivial
queries such as biochemical versus cell-based assays or luciferase reporter
gene assays are currently not possible, because the assays do not have explicit
annotations that capture this type of information. Other relevant queries can
include the identifi cation of nontoxic kinase inhibitors, promiscuous luciferase
reporter gene compounds, or compounds that may interfere with fl uorescence
intensity assays. To enable researchers to retrieve quantitative information
that is meaningful across multiple assays of interest also requires the standard-
ization of the endpoints that are reported as the primary (most important)
outcomes of screening experiments, for example, percent inhibition, IC
50
,
K
i
,
and so on. BAO defi nes the meaning of common endpoints and can relate
different types of endpoints to other concepts in the ontology. A later goal of
BAO is to semantically integrate screening data with other publicly available
biological databases, such as biological pathways, human diseases, known tox-
icities, adverse drug reactions, and possibly also predictive models.
BAO will make use of existing ontologies where appropriate, for example,
cell lines [61] or Gene Ontology [54, 55]. It will facilitate integration with other
databases such as biological pathways via BioPAX [62, 63] and it will support
inference. The software development component of the BAO project makes
use of Semantic Web technologies, such as Jena [64] and Vivo [65]. The ontol-
ogy is also being implemented using the ontology management application
framework Protégé 4.1 [66] to support the design of the structure of the assay
ontology. It is being developed using Web Ontology Language (OWL) 2.0 [48,
67], which is currently the most expressive description logic (DL) language.
This is in contrast to most of the OBO ontologies.
BAO includes several high-level concepts related to biological screening,
including perturbagen, metatarget, format, technology, analysis, and endpoint.
Perturbagens deposited in PubChem and the other screening data sources
mentioned above are mostly small molecules, but perturbagens can include
various other perturbing agents that are screened in an assay. The metatarget
Search WWH ::
Custom Search