Biomedical Engineering Reference
In-Depth Information
OPS Semantic Integration Hub
OPS Data Sources
Small-Molecule Databases
Biological Target Databases
Gene Expression Databases
BioBanks
Literature
Patent Database
Wiki Sources
Patent/Physician Blogs
Data sources fully interoperable
and available for the querying of
relationships between the
cardinal assertions to allow for
prediction of new hypotheses
that are targeted to specific
audiences
In-Text Semantic Support
(e.g., Knowledge
Enhancer)
Mobile Nanopublication
Reviewer
Data Extraction
Nanopublication Store
Human and Automated
Concept and Triple Creation
Cardinal Assertions Store
Mapping to ConceptWiki
UUIDs for disambiguation
Providence/Evidence Store
Reasoning with
Cardinal Assertions
Figure 26.2
The OPS process moving from disparate data sources to an interoperable,
searchable resource.
sible homonyms, but of course the use of these will never disappear completely
due to legacy literature. Thus, human disambiguation for notoriously problem-
atic symbols will be an integral part of the OPS, aided by in-text semantic
support systems, such as the knowledge enhancer.
As described earlier, concept triples or assertions typically have the format
of three concepts, namely, an object, a predicate, and a subject. In classical
RDF triples, the predicate is not necessarily seen as a concept. However, in
the OPS, the predicate will be defi ned as a concept and have a UUID. Based
on this approach, each unique assertion can be defi ned as a unique three-
UUID combination of subject, predicate, and object. The OPS interoperabi-
lity layer is in fact a very rich triple store or nanopublication store, ideally
containing all relevant assertions in the pharmacological space, richly anno-
tated with provenance metadata. Obviously, many nanopublications may
make the same basic assertion and only differ in metadata. This is due to the
fact that many assertions in the biomedical and chemical space are frequently
repeated over and over in articles or database records subsequent to their
original publication. Mapping of all identical subject-predicate-object combi-
nations in the OPS to create unique assertions is a crucial step to enable
applications that are suitable for retrieval, browsing, and reasoning. It is impor-
tant to keep track of the frequency of the repetition of the assertions, as it will
be refl ected in an evidence score catalogued in the OPS nanopublication store.
Search WWH ::
Custom Search