SCIENTIFIC NETWORKING AND COLLABORATIONS - Collaborative Computational Technologies for Biomedical Research - page 154

example, already has a formidable array of telescopes operating at different
wavelengths on earth and in space. These instruments are streaming enormous
amounts of data to astronomers worldwide that can in turn spend years ana-
lyzing the data at a cost that is greater than the cost of the telescopes them-
selves. A good example of the implementation of e-science is provided by the
“virtual observatory” project where observational data from different sources
are collated for analysis by astronomers at different geographical locations
(http://www.us-vo.org/). Ambitious future plans for massive astronomical
surveys plus ongoing scientifi c activities such as particle physics and climate
science are creating huge demands for hardware and software that can cope
with the deluge of data.
Until quite recently, data-intensive science in biology was limited to tax-
onomy. Genome sequencing, expression analysis, and pharmaceutical research
have changed all that. Thousands of DNA sequencing runs are being per-
formed worldwide using templates from many different forms of life. Meta-
genomics is a powerful but data-intensive approach of reconstructing individual
sequences from heterogeneous biological samples containing many different
organisms. A recent example is the sequencing of the human gut microbiome
which contains 10 times more bacterial cells than there are cells making up
the human body [14].
Despite the ongoing efforts to sequence as much plant, animal, and micro-
bial life as possible, the main focus of biomedical interest remains the human
genome. As new technologies allow the cost of sequencing to approach the
fi gure of $1000 for a 3-gigabase genome, the output of digital data will be easily
on a par with the output from the particle physics laboratories and astronomi-
cal surveys and there will be some tough choices that will have to be made
regarding how much raw experimental data should be archived and how much
permanently deleted.
To conclude this section, interactive Web technologies are providing the
physical and virtual infrastructures required for collaboration and networking
in many scientifi c activities, the life sciences included. Some (but not all) of
this activity involves large-scale sharing and analysis of data, the fourth para-
digm of Jim Gray. Before considering networking in the life sciences in more
detail, we will examine some of the issues that are being addressed in order
to turn these visions into reality.

10.3.2 Computing Infrastructures

Modern computers are orders of magnitude more powerful than the primi-
tive machines that were developed at the end of the World War II. This is
a statement of the obvious to all readers of this topic, but it is worth refl ecting
on just how big this magnitude difference actually is. As noted by Nielsen
in a recent Nature review [15], the electronic numerical integrator and com-
puter (ENIAC) was a state-of-the-art machine for modeling complex phenom-
ena in 1946. Now, the large hadron collider at the European Organization for