Biomedical Engineering Reference
In-Depth Information
enable multiple personalized virtual biotechs, working on the same or related
diseases, to collaborate and share data, knowledge, and resources. By coordi-
nating their efforts, a network of personalized virtual biotechs can systemati-
cally explore the opportunity space of targets and leads and avoid unnecessary
replication of experiments.
One can conceptualize the ongoing activity of a given personalized virtual
biotech in terms of requests of various sorts ranging in complexity. Examples
of requests may include a call for opinions on a particular scientifi c question
(say, by asking for votes), a request for some data set (or reference to a paper)
that addresses some specifi c scientifi c question, a request for help with some
complex bioinformatic calculation, and so on. Note that the requests form a
hierarchy: At the top a physician might request a drug to treat a particular
patient given the patient's clinical and genomic profi le. This is, of course, a very
broad request, and hard to satisfy as stated, but must be broken down into a
series of simpler requests: a request from pathology for a tissue sample, a
request for a laboratory to culture the sample, a request for a high-throughput
screening facility, such as the National Institutes of Health (NIH) Chemical
Genomics Center (NCGC) (http://www.ncgc.nih.gov/) to run assays on the
tissue model using existing drugs, a request for the gathering together of those
results, a request for experts to interpret those analyses and rank the results
to create a pipeline, a request for a mouse model of the disease, a request for
someone to run the highest ranked molecules through the mouse model, and
so on. Of course, each of these will usually be broken down further into sub-
requests until one reaches small doable tasks or small answerable questions.
This recursive request structure is commonplace in science, and indeed in any
structured problem solving, and can work across a community [4]. What is
unique about operating such a hierarchy of goals in Cancer Commons is that
such requests may be effi ciently fanned out across a wide-ranging distributed
community. This ability to effi ciently leverage distributed resources in a call-
and-response manner is a unique capability of Internet-based technologies [4].
Moreover, the Internet offers methods for distributed decision analysis, which
can be utilized to prioritize opportunities such as targets and leads, to collab-
oratively analyze and interpret data, and to make other complex decisions.
The Cancer Commons platform facilitates this process of distributed call
and response by employing a series of unique technologies. Specifi cally, the
platform includes a registry of services (such as those exemplifi ed just above)
and a hub-and-spoke architecture where the hubs can automatically and
securely communicate with one another through the Web. Each hub corre-
sponds to a community of clinicians and/or researchers organized around a
project, an institution, a disease, or a discipline. Each hub advertises the ser-
vices that are available within that hub's community (i.e., the knowledge, tools,
and expertise of those in that particular community). When requests are made
within a given hub (either by researchers or in the process of operating semi-
automated workfl ows), if the services required to respond to that request
are not locally available, the hub architecture will automatically call out to
Search WWH ::
Custom Search