CULTURE AND RESEARCH INFRASTRUCTURE 349
10.2.4 Industry,
Industrial interest in the BioComp interface is driven by the prospect of potentially very large
markets in the life sciences—especially medicine. Information-enabled bioscience is further expected to
create large markets for information technologies customized and adapted to the needs of life scien-
tists—accounting in substantial measure for the interest of some large IT companies in this area. Indeed,
according to the International Data Corporation, life science organizations will spend an estimated $30
billion on technology-related purchases in 2006, up from $12 billion in 2001.^25
Life science companies (e.g., pharmaceuticals) view information technology as a (or perhaps the)
key enabler for drug design and treatments that can in principle be customized to groups as small as a
single individual. Consider, for example, the specific problem of finding useful organic compounds,
such as drugs, to treat or reduce the effects of disease. One approach is based on the use of combinato-
rial methods in chemistry, genetic engineering, and high-throughput screening technology. Such an
approach relies on trial-and-error to sift candidate compounds on a large scale to sidestep the complexi-
ties of data in a search for compounds with sufficient potential to be worth the effort of laboratory
testing for useful outcomes; similar techniques can be used for strain improvement and natural product
synthesis.^26
A second approach is to use computational modeling and simulation. Data mining (Section 4.4.8)
can be used in addition to empirical screening to identify compounds that are likely to have a desired
pharmacological effect. Moreover, what the combinatorial and high-throughput empirical approach
gains in expediency, it may lose in insight. For example, causality in combinatorial approaches is often
difficult to attribute; and thus, it is difficult to generalize these results to other systems. Combinatorial
methods are less likely to find solutions when the desired functionality is complex (e.g., when the
biosynthetic route to a product is complicated or when a disease treatment relies on the inhibition,
without side effects, of various pathways). Also, of course, from the standpoint of basic science, predic-
tive understanding is at a premium. Computational simulation is thus used as the screening tool for
promising compounds—a cell’s predicted functional response to a given compound is used as that
compound’s measure of promise for further (empirical) testing. Thus, although granting drug approv-
als on the basis of simulations makes little sense, simulations may be able to predict with an adequate
degree of reliability what drugs should not advance to expensive in vivo clinical trials.^27 Many believe
that information-enabled bioscience and biotechnology have the potential to be as revolutionary as
information technology was a few decades ago.
(^25) E. Frauenheim, “Computers Replace Petri Dishes in Biological Labs,” CNET News.com, June 2, 2003, available at http://
news.com.com/2030-6679_3-998622.html?tag=fd_lede2_hed.
(^26) See, for example, C. Khosla, and R.J. Zawada, “Generation of Polyketide Libraries via Combinatorial Biosynthesis,” Trends in
Biotechnology 14(9):335-341, 1996; C.R. Hutchinson, “Combinatorial Biosynthesis for New Drug Discovery,” Current Opinion in
Microbiology 1(3):319-329, 1998; A.T. Bull, A.C. Ward, and M. Goodfellow, “Search and Discovery Strategies for Biotechnology:
The Paradigm Shift,” Microbiology in Molecular Biology Review 64(3):573-606, 2000; Y. Xue and D.H. Sherman, “Biosynthesis and
Combinatorial Biosynthesis of Pikromycin-related Macrolides in Streptomyces venezuelae,” Metabolic Engineering 3(1):15-26, 2001;
and L. Rohlin, M. Oh, and J.C. Liao, “Microbial Pathway Engineering for Industrial Processes: Evolution, Combinatorial Biosyn-
thesis and Rational Design,” Current Opinion in Microbiology 4(3):330-335, 2001.
(^27) For example, the Tufts Center for the Study of Drug Development estimates the cost of a new prescription drug at $897
million, a figure that includes expenses of project failures (e.g., as those drugs tested that fail to prove successful in clinical
trials). Since clinical trials—occurring later in the drug pipeline—are the most expensive parts of drug development, the
ability to screen out drug candidates that are likely to fail in clinical trials would have enormous financial impact and would
also reduce the many years associated with clinical trials. See Tufts Center for the Study of Drug Development news release,
“Total Cost to Develop a New Prescription Drug, Including Cost of Post-Approval Research, Is $897 Million,” May 13, 2003,
available at http://csdd.tufts.edu/NewsEvents/RecentNews.asp?newsid=29. Of particular interest is a finding reported by
DiMasi that if preclinical screening could increase success rates from the current 21.5 percent to 33 percent, the cost per
approved drug could be reduced by $230 million (J.A. DiMasi, “The Value of Improving the Productivity of the Drug Devel-
opment Process: Faster Times and Better Decisions,” PharmacoEconomics 20(S3):1-10, 2002).