Biomedical Engineering Reference
In-Depth Information
network enrichment or network building can be carried out for in-depth
exploration of the data.
An even more interesting workfl ow was enabled when we connected the
manually curated biomedical assertions described previously to Ariadne
Genomics Pathway Studio. In this module the quantitative biomedical asser-
tions are translated into semantic triples by a software algorithm, exported
into the appropriate XML format, and passed on for network visualization to
Pathway Studio (Fig. 16.11 ).
16.8
STRATEGIC CONSIDERATIONS
In building tranSMART, we made several strategic decisions that have proven
very benefi cial. First, we have selected the core technologies such that we can
build on previous efforts in a most effi cient way. After rigorous evaluation, we
decided to build tranSMART on open-source frameworks such as i2b2,
GenePattern, and Haploview to enable data sharing and partnering.
Alternatives were considered such as proprietary coding or several commer-
cial offerings, but we were intrigued by the possibilities of taking the open-
source approach. The result has been an excellent and robust system, access
to an outstanding community of academic partners, and the ability to share
and propagate our success.
Second, a preferred academic partner was selected as core partner for the
program. In this case, we chose the Cancer Institute of New Jersey (CINJ) for
its informatics and medical expertise which we utilized throughout the project.
In return, the institute will receive a full version of the software for internal
implementation. This provides an obvious economic “win-win” for both sides
and serves as a model for further academic and nonprofi t collaborations that
is being established.
Finally, tranSMART is the fi rst Johnson and Johnson application to be
hosted externally on the Amazon Elastic Compute Cloud (EC2). Cloud com-
puting offers the advantages of low entry cost, a pay-per-use system, high
speed, ready accessibility, easier access for external partners, and a potential
to effi ciently integrate with other public available resources hosted on the
same cloud. At Johnson and Johnson, we achieved a 12-fold reduction in
hosting costs, as well as increased “up-time,” without incurring any additional
risks associated with protection of information assets.
16.9
DISCUSSION
The system has been deployed for almost a year at the time of this writing.
During this time we have trained approximately 200 internal scientists and
physicians and the system was queried close to 4000 times. The data warehouse
contains more than 60 internal and public human studies with the access to
Search WWH ::
Custom Search