Pirate: is this trend of collecting biotech data a growing part of NTAP's biz?
Future Boy
The New Biology
By Erick Schonfeld
ecompany.com
With the first phase of the Human Genome Project recently
completed, the real work of figuring out how genes influence
disease can finally begin in earnest. But one truth that is already
apparent is that biological research will become impossible without
the assistance of advanced computer technology. There will be so
much data to sift through -- trillions and trillions of possible
gene and protein combinations -- that there will be little
meaningful work done in this field without computing.
An announcement on May 30 by IBM and others that they would create
a joint venture called Blueprint Worldwide highlights the growing
link between the two disciplines. Blueprint aims to bring together
all publicly available information about proteins (their structure,
function, and location, and how they interact) from all available
biomolecular databases, as well as to cull data from 200,000
published papers. During the conference call, Caroline Kovac, the
vice president of IBM's life sciences division, explained the point
of the not-for-profit project: "Blueprint is a part of what we see
as a convergence between biology and computers. The new biology is
a science that no longer can be practiced without computers." There
has been an exponential growth of biological data represented by
genomics, and an even vaster amount of data will be created by
proteomics (the study of proteins).
For many computer scientists today, there are few challenges as
exciting as attempting to delve into -- or even replicate in
silico, as they say -- the events that occur in a biological
environment. Kovac mentioned BlueGene, an IBM project to create the
fastest computer in the world by clustering together more than 1
million microprocessors; when BlueGene is completed in 2004, its
first task will be to try to figure out how proteins unfold.
"Supercomputing today is being driven by biology," she pointed out,
"whereas 10 years ago it was driven by physics."
But Big Blue isn't helping to create this free repository of data
purely out of the goodness of its heart. It needs to learn how to
better apply computer technology to the new biology. Bioinformatics
represents a potentially huge market for IBM's hardware and
software. Kovac estimates that life sciences will be a $40 billion
industry segment by 2004, up from $22 billion in 2000. As
scientists move from mapping out the human genome to mapping out
the protein interactions that take place within human cells, "we
see a tremendous opportunity to bring information systems to these
scientists," Kovac said. IBM will learn from Blueprint by using it
to test novel ways to, say, cluster computers or come up with
software to better annotate data about gene splices.
But IBM is not the only company preparing for the onslaught of data
that the proteomics revolution is about to unleash. Housed in a
seven-story, green-marble, art-deco department store building in
downtown Oakland, Calif., is a startup called DoubleTwist that is
tackling the same challenge. "What is bioinformatics?" asks CEO
John Couch. "It is capturing information coming out of the genome."
But capturing that information is easier said than done. "Every
biotech company is struggling with its information systems," he
says. A veteran software executive from the early days of Apple,
Couch sees a great need to simplify database computing for
scientists, just as Apple simplified personal computing for
non-programmers.
Large drug companies today employ bioinformaticians with roles
similar to those of management information systems employees three
decades ago. Back then, an MIS manager would take queries from
executives about things such as inventory levels or sales patterns,
run them through mainframes or minicomputers, and then report the
answers to the people who actually needed the information to make
business decisions. The same thing is happening today with
bioinformaticians and the biological data that scientists need to
do their research. Couch is addressing this problem by making
DoubleTwist an information resource for bioinformaticians and
scientists alike, and even going so far as to take away the
headache of managing the databases and server farms needed to do
cutting-edge biology.
At first glance, it may appear that DoubleTwist is trying to charge
for the same thing that the Blueprint joint venture proposes to
give away for free: a one-stop digital portal for gene-related
data. Indeed, there will be some overlap, but Blueprint is focused
only on publicly available sources of such data. While that is a
rich vein, it is far from complete. In addition to offering access
to public databases (such as those that contain a map of the human
genome, as well as information about proteins and nucleotides),
DoubleTwist pulls together proprietary databases from other
companies. These may focus on specific sets of genes, such as those
related to the human brain, or allow for comprehensive patent or
literature searches. DoubleTwist also has its own gene indexes and
an annotated version of the human genome (which companies such as
Merck and Hitachi pay good money for). "I want to be the Bloomberg
of this space," Couch says. All in all, he already integrates data
from about three dozen sources, and he sees Blueprint as simply
another one to draw from.
"Data is data," says DoubleTwist president Rob Williamson, "but
information is what people want." It takes computers and computer
know-how to find those valuable nuggets of data that all of a
sudden become "information" -- the elusive philosopher's stone of
our age. DoubleTwist, for one, is making this transformation easier
by adding an intuitive graphical interface that allows scientists
themselves (with a little training) to interact with the slew of
databases available to them. The company runs multiple software
algorithms against multiple databases, and provides scientists with
one comprehensive view of their search results. If anything related
to a gene sequence that a scientist is studying changes in the
public or private databases, the literature, or the patents, the
software informs that scientist automatically. DoubleTwist also
makes it easy to integrate all of this outside information with a
drug company's own closely held internal data (DoubleTwist's
customers can either host all of this software on their own servers
or get access to the external databases through an ASP service on
DoubleTwist's site.)
What both the Blueprint and DoubleTwist efforts point to is an era
in which scientific discoveries are just as likely to come from
parsing trillions of bits of data that already exist somewhere in a
computer as they are to be discovered through finely tuned
experiments in a wet lab. The efforts also signal a much more
collaborative type of science than has ever before existed. As
Kovac put it, "Biology is no longer about researchers in their
laboratories, but a more dynamic community living and breathing in
an age of interconnectedness." If Kovac, Couch, and others can pull
off their grand bioinformatics projects of connecting the most
brilliant minds in biology, they might turn out to be right that
the impact of this new technology during the next decade will be
greater than what we have seen so far from the Internet.
For more information and related links, see the online version of
this story at: ecompany.com |
