Biomedical Engineering Reference
In-Depth Information
26.7 ADOPTION OF TECHNOLOGIES IN OPEN
PHARMACOLOGICAL SPACE
Drug discovery is data intensive, requiring all major pharmaceutical compa-
nies to maintain extensive in-house instances of public biomedical and chemi-
cal data alongside internal data. Analysis and hypothesis generation for drug
discovery projects require careful assembly, overlay, and comparison of data
from many sources, requiring shared identifi ers and common semantics. For
example, expression profi les need to be overlaid with gene and pathway identi-
fi ers and reports on compounds in vitro and in vivo pharmacology. The utility
of data-driven research goes from virtual screening, high-throughput screening
analysis, via target fi shing and secondary pharmacology to biomarker identi-
fi cation. The alignment and integration of internal and public data and infor-
mation sources are signifi cant efforts and the process is repeated across
companies, institutes, and academic laboratories. This represents both signifi -
cant waste and an important opportunity at the same time.
To address these challenges, the Open PHACTS consortium, comprised of
14 European core academic and small and medium enterprises, will partner
with leading pharmaceutical companies to develop an open-source, open-
standards, and open-access innovation platform, the Open Pharmacological
Space (OPS), via a Semantic Web approach. OPS will bring together the data,
vocabularies, and infrastructure needed to accelerate drug-oriented research.
This semantic integration hub will address key bottlenecks in small-molecule
drug discovery—disparate information sources, lack of standards, and shared
concept identifi ers—and be guided by well-defi ned research questions assem-
bled from the participating drug discovery teams (Fig. 26.2). Vocabularies, or
simple terminology systems, contain multiple symbols referring to each concept
contained in the vocabulary. These symbols are essentially synonyms of each
concept. Each symbol used by the OPS will be recognized and mapped to the
correct concept UUID, as done in ConceptWiki. This is not always trivial, due
to vast ambiguity in symbols used. Many symbols have multiple meanings in
the sense that they can refer to multiple concepts. For instance, BSE is an
ambiguous symbol as it can refer to many concepts, including bovine spongi-
form encephalopathy and breast self-examination. To map ambiguous symbols,
disambiguation algorithms are used, usually based on context. Several partners
in Open PHACTS are among the leading research groups developing and
exploiting high-performance disambiguation algorithms in the life sciences
and chemistry.
For the OPS, the growing and curated vocabularies containing eventually
all symbols for all relevant concepts will be based on the principle that each
concept has a UUID, and all symbols known to refer to that concept will be
listed in an identity resolution system, presently represented by ConceptWiki.
Active resolution of symbols to UUIDs will enable specialized concept taggers
to map symbols in text and databases to the correct concepts. Highly ambigu-
ous symbols can be actively discouraged by immediate exposure of the pos-
Search WWH ::
Custom Search