Grid Security - Grid Computing: Infrastructure, Service, and Applications - page 318

agreed a priori and coni gured in the portal so that a limited form of query-
ing is possible. Thus a nurse or investigator can only query datasets based
upon the forms in the portal, which can be parameterized.
The results of the given query shown here depict the cancer patients: the
specii c forms of cancer that they have and other information such as their
postcode. The geospatial clustering of diseases is often an important
research area in its own right. Similarly, when undertaking a feasibility
study for a given trial it may be important to know the number of patients
with a given condition. Such scenarios are being explored in the recently
funded ESRC project Data Management through e-Social Science (DAMES)
project [31].
Assuming that a set of individuals have been found with a particular
condition, that they have been recruited to a given trial, understanding
the genetic similarities and differences of these individuals is often
required. Do they have the same gene mutations? Are there genetic differ-
ences that can help shed insight into why this individual developed this
particular disease? To address these kinds of questions requires access to
services that deal with individuals' genetic datasets. One way in which
genetic data are established and analyzed is through microarray analysis.
The BBSRC GEMEPS project developed various services that allowed
secure access to both non-public and public microarray datasets.

12.5.2 GEMEPS Project

The GEMEPS project developed a grid infrastructure for discovery,
access, integration, and analysis of microarray datasets. Through the
GEMEPS infrastructure scientists were able to ask the following ques-
tions and obtain appropriate results based upon their privilege: Who
has run a microarray experiment and generated results similar to mine?
Who has undertaken experiments and produced data relevant to my
own interests; for example, for a particular phenotype, for a particular
cell type, for a particular pathogen, on a particular platform or microar-
ray chip set? Show me the results from a particular collaborator, and show
me the conditions and analysis associated with experimental results simi-
lar to mine.
In all of these scenarios, the premise was that sites should keep and
maintain their own data and dei ne their own security policies on access
and usage. Since scientists are often reluctant to publish their data in pub-
lic repositories until they have published results in recognized journals,
which can, depending on the journal, be a long and protracted affair,
many datasets remain inaccessible to the wider research community.
These issues are discussed in detail in [32].
Scientists in the i rst instance would like to be able to query across a
range of experiments based on any one or more of a variety of search terms.
Thus scientists are unlikely to be interested comparing experimental