Biomedical Engineering Reference
In-Depth Information
is that these resources are available on demand without evaluation or previous
agreement by a scientifi c committee. This opens new perspectives to scale up
existing strategies by a factor of 10, 100, or even 1000 or to explore new
approaches without guaranteed success. The change of scale is a major driver
for scientifi c progress. The history of physics shows that scaling up in energy
and in luminosity was the key to discover new phenomena.
Our group was involved in two biomedical projects where the capacity to
think bigger was exploited and led to very signifi cant scientifi c results.
15.3.1.1 Example of WISDOM Drug Discovery Platform
The pharma-
ceutical research and development (R&D) enterprise presents unique chal-
lenges for information technologists and computer scientists. The diversity and
complexity of the information required to arrive at well-founded decisions
based on both scientifi c and business criteria are remarkable and well recog-
nized in the industry. Drug discovery is the process by which drugs are discov-
ered and/or designed. Drug candidates are inputs to the drug development
process. Current efforts within the pharmaceutical industry are directed at
reducing the time and costs for drug development. Recent progress in genom-
ics, transcriptomics, proteomics, high-throughput screening, combinatorial
chemistry, molecular biology, and pharmacogenomics has radically changed
the traditional physiology-based approach to drug discovery where the organ-
ism is seen as a black box.
An important step in the drug discovery process is virtual screening, which
is about selecting druglike molecules active on a specifi c biological target by
computing the binding energy of the molecule to the target active site [6]. The
prerequisite for the use of virtual screening is to know the three-dimensional
(3D) structure of both the druglike molecules and the target active site. The
3D structures of more than 3 million chemical compounds are now available
in public databases like ChemBridge and ZINC while the Protein Data Base
provides the structure of more than 50,000 proteins of biological interest [7].
Since 2004, the WISDOM initiative [5] has successfully deployed large-scale
virtual screening computations on grid infrastructures in order to fi nd new
drugs against malaria, avian fl u, and diabetes. Meanwhile, it has also grown into
a multidisciplinary collaboration of biologists, biochemists, bioinformaticians,
and e-scientists from Africa, Asia, and Europe. More than a thousand CPU
years have been used since 2004 on e-infrastructures in France, Africa, America,
Asia, Open Science Grid (OSG), and of course on EGEE [3], which has pro-
vided the majority of the resources. About 20% of the druglike molecules
selected in silico have been confi rmed by in vitro tests to be active inhibitors
and most promising molecules have been patented [8]. Today, WISDOM is a
success story from a grid deployment point of view because it has demon-
strated the potential impact of e-infrastructures for virtual screening.
15.3.1.2 Example of Protein Database Refi nement
During the spring of
2007, a large-scale application [9] was deployed on grid resources in order to
Search WWH ::
Custom Search