Biomedical Engineering Reference
In-Depth Information
researchers of potential collaboration opportunities. Another powerful
data-mining strategy that only becomes possible within the context of a
working ontology used within an environment that hosts both public and
private data sets is the suggestion of similar data that may be relevant to
the researcher.
A major outstanding problem is the lack of a coordinated strategy for non-
ontologists (e.g., experimentalists) to be suffi ciently incentivized to mark up
data for ontological binning upon importing screening data into databases. We
will take a two-pronged approach consisting of both pull and push mecha-
nisms. On the pull side, CDD Collaborate allows researchers to compare its
private data with public data and private data with collaborators' data and,
with the new technology envisioned, has the option to “opt into” alerts for
notifi cation when others are working on similar/complementary compounds,
targets, and so on. Adopting the ontology will be a pull incentive by rewarding
scientists with potentially complementary data and new collaborators. On the
push side, CDD already has required fi elds in the database such as the type
of assay (enzyme, cell, animal, etc.) and requirements for selecting a date for
a run of a screen. Similar requirements will be added to accurately annotate
new defi nitions within an assay ontology, such as selecting a target from a
preloaded ontology. The combined pull and push mechanism will rapidly lead
to a very large set of ontology-compliant data, greatly facilitating both human
and automated connections between public and private data.
21.6.1.2 Collect, Mine, and Share Multiple Types of Data in CDD
T o
date the CDD database has solely focused on the small- molecule community.
There is clearly an enormous opportunity to greatly increase the size of the
researcher community using the collaborative software by expanding to larger
molecules and biological materials. These larger molecules and biologics are
equally important to fi nding treatments for neglected diseases as well as of
broad commercial and academic interest. This will engage a greater percentage
of the research community, for example, those doing fundamental biology
research or working on vaccines. Furthermore, importing related chemical and
biological data sets opens up the possibilities for evaluating combinations of
small and large molecules.
We will create the capacity to archive, mine, and collaborate with generic
objects within CDD. We allow researchers to customize the database because
they will be able to change the Molecule fi eld to, say, a Sequence fi eld. Simple
renaming is not enough, so additional details will be engineered into domain-
specifi c modules. We are aware of at least one pharmaceutical company that
has developed a macromolecular structure notation, editing, and registration
tool (Tianhong Zhang, personal communication, 2009) using the ChemAxon
components (Marvin Sketcher, Marvin viewer, calculator plugin and Marvin
API already integrated within CDD). The pharma application is currently
unavailable to researchers in academia, other foundations, or companies and
represents a signifi cant investment.
Search WWH ::
Custom Search