ABSTRACT
The rate of growth of available information about biomedical and biological
systems is exponential. In fact, available information has already gone far
beyond human ability to synthesize, analyze, and predict. The biomedical
information torrent is now a continuously growing set of exceedingly intricate
knowledge about complex, large dynamic systems. Against this background,
one of the most urgent challenges to the biomedical research community, with
direct impact on drug development, seems to be to develop approaches to
analyze this extremely large amount of data to discover patterns (1) or useful
information that may be available in the data but not apparent through simple
inspection. This integration aspect has always been a challenge in drug develop-
ment, where data are continuously gathered at a variety of scales and sizes
through the preclinical and clinical development programs. Recently, academic
and health sciences research has also become aware of the importance of this
very challenge. The National Institutes of Health (NIH) Roadmap for medical
research in the 21st century states that:
Today’s biomedical researcher routinely generates an amount of data
that would fill multiple compact discs, each containing billions of
bytes of data. [A byte is approximately the amount of information con-
tained in an individual letter of type on this page.] There is no way
to manage these data by hand. What researchers need are computer
programs and other tools to evaluate, combine, and visualize these
data. In some cases, these tools will greatly benefit from the awesome
strength of supercomputers or the combined power of many smaller
machines in a coordinated way but, in other cases, these tools will be
used on modern personal computers and workstations (2).