Summary of Findings
I. Information Technology and Biotechnology are Key Sectors for the 21st Century. The information and biotechnology sectors are each very important to America's economy, security, and well-being.1 Although some expect the 21st century to be dominated by the Biotechnology Revolution, the Information Revolution begun in the 20th century has not run its course. Advances in information technologies remain central to economic growth; they will also be critical to progress in the Biotechnology Revolution itself. Reaping the health benefits of sequencing the human genome depends on processing and making sense of enormous amounts of data that, in turn, will be made possible by advances in computing and networking technologies. At the same time, information technology will advance based on a better understanding and use of the principles of biological systems and mechanisms. Already, NASA has developed flight control software for the F-15 aircraft based on neural network principles.2 Biological models also may
lead to improved computer hardware architecture and other technologies, such as MEMS and nanotechnology.3
II. Multidisciplinary Approaches Are Increasingly Needed in Science and Engineering Research. More and more, progress in research depends on multidisciplinary efforts.4 Complex research problems require the integration of both people and new knowledge across a range of disciplines. In turn, this requires knowledge workers with interdisciplinary training in mathematics, computer science, and biology. Bioinformatics is a key example of a highly multidisciplinary field requiring workers with interdisciplinary training.5
Biotechnology R&D and information technology R&D each provide tools and models useful for the other. Interdependencies also exist among chemistry, physics, and structural biology, and among mathematics, computer engineering, and genomics. Further examples of these interdependencies can be found in the complementary roles of the physical sciences and engineering in nano-scale semiconductor work, and in the overall important role engineering research plays in providing new research tools and diagnostics in all of these areas.6
III. Government-Industry and University-Industry Partnerships Have Often Been Effective in Supporting the Development of New Technologies. Symposium participants noted many examples of partnerships involving the federal government, universities, and industry that have contributed to the development and strengthening of leading U.S. industries.7 Mechanisms to
support partnerships include the Engineering Research Centers and Science and Technology Centers funded by the NSF, NIST’s Advanced Technology Program, and the DOD’s former Technology Reinvestment Program. They also included activities funded by Cooperative Research and Development Agreements (CRADAs).8 SEMATECH, initially funded by government as well as industry, has played a key role in improving manufacturing technologies for the U.S. semiconductor industry.9
A Multidisciplinary Approach: Partnerships between universities and companies that cut across disciplines, though often challenging, are increasingly important to progress as biotechnology and information technologies become dependent on contributions from other fields. While projects led by individual investigators remain vital to general scientific and engineering advancement, solving complex problems in new areas such as bioinformatics and next-generation computing requires larger, multidisciplinary collaborations among scientists and engineering researchers. Both government agencies and industry groups can help overcome these institutional problems by funding multidisciplinary research projects focused on important complex problems.
Federal laboratories also offer important capabilities and institutional lessons for dealing with complex research problems. Historically, NIH has not directly supported industry R&D, but this is changing. In 1998, NIH laboratories entered into 166 CRADAs, and in 1999 NIH’s Small Business Inno-
vation Research Program awarded more than $300 million to small companies.10 This is expected to rise to $410 million in 2001.11
The availability of genome sequences has also prompted research on a new scale, leading to new types of partnerships in the biotechnology, pharmaceutical, and biomedical areas. Understanding the functioning of cells at the system level, or how thousands of genes and proteins interact, requires collaborations among large numbers of researchers with interdisciplinary backgrounds, including those with sophisticated computer skills.
Multidisciplinary research, with teams from different departments and organizations is needed to attack issues at the periphery of standard disciplinary boundaries.12 The advent of such organizations as the SNP Consortium, the Alliance for Cellular Signaling, and the Mouse Sequence Consortium—not to mention the organizational effort mounted to sequence the genome itself—illustrate the need for and possibilities of partnerships in biomedicine.13
IV. Federal R&D Funding and Other Innovation Policies Have Been Important in Supporting the Development of U.S. Industrial Capabilities in Computing and Biotechnology. The history of the U.S. biotechnology and computing sectors illustrates the value of sustained federal support, although federal roles and contributions have been different in the two areas:
-
Computing and Semiconductors. In computing and semiconductors, the federal government has played four important roles since World War II.14
-
Antitrust. Antitrust policy played an important role in the creation of the semiconductor industry by requiring the pre-1984 AT&T to license broadly its patents in transistors.15
-
Procurement. Government procurement in the early days of both semiconductors and computers presented a valuable stimulus to technological development and industry growth; in addition to providing generous contracts, the government served as a reliable first customer, providing companies with early markets.16
-
R&D Funding. Federally funded R&D was important in developing certain niche areas (e.g., gallium-arsenide chips) and certain new areas of computing (e.g., ARPANET and 3D graphics). Reflecting the semiconductor industry’s dependence on research and its special needs, the federal research effort has been complemented by private support, such as the industry-created Semiconductor Research Corporation,17 and later through the SEMATECH Consortium—the highly regarded federal-industry partnership.18
-
High Performance Computing. The government has long played, and continues to play, an essential role in high-performance computing. There is usually only a small commercial market for the most powerful supercomputers; the government itself continues to act as the main purchaser of leading-edge machines for a variety of national missions. As a result, federal R&D support and procurement remain critical to progress in this field.19
-
Biotechnology. It is widely recognized that federally funded research has played a central role in the development of the biotechnology industry. The federal government funds 90 percent of university-based health research in the United States.20 Many of the key discoveries and techniques of
-
modern biotechnology have come out of university research funded by the National Institutes of Health (NIH). While the NIH continues to play a crucial role, corporate research also broadens the technology base. Indeed, other non-biological technologies, such as computing, have become increasingly important to the continued progress of the biotechnology industry.21
-
Patent Policies. Federal patent policy also has a significant impact on these fields. Covering, as it does, areas as diverse as biotechnology and e-commerce, current federal patent policy is sometimes a source of uncertainty and, thus, could prove to be an obstacle to innovation.22 The federal government could help further stimulate innovation by clarifying patent rules (e.g., regarding biotechnology research tools) to reduce risks of litigation over patent issues.23
V. Limitations of Venture Capital. The U.S. venture capital market—the largest and best developed in the world—often plays a crucial role in the formation of new high-technology companies. Even so, the nature of the role and constraints posed by venture capital markets need to be better understood. The provision of venture capital, with its informed assessment and management oversight, is but one element of a larger innovation system. It is not a substitute for government support of long-term scientific and technological research. Indeed, the venture capital industry is not designed to support early-phase research and rarely does so. Venture capitalists do provide vital support to business efforts that show the potential of exploiting new technologies and business ideas.24 As such, venture funding tends to
21 |
See the presentation by Mark Boguski on “Computing and the Human Genome” in this volume. See also the presentation of Jeffrey Schloss in National Research Council, The Advanced Technology Program: Challenges and Opportunities, Charles W.Wessner, editor, Washington, D.C.: National Academy Press, 1999, pp. 56–59. To hasten the development of computational tools in biotechnology, the NIH established the Biomedical Information Science and Technology Initiative in April 2000. The initiative “is aimed at making optimal use of science and technology to address problems in biology and medicine.” More information on the initiative is available at the NIH Web site: http://grants.nih.gov/grants/bistic/bistic2.cfm. |
22 |
The STEP Board is currently conducting a study on these and related issues led by Richard Levin, President of Yale University, and Mark Myers, Senior Vice President (retired) of Xerox, now with the Wharton School. A major conference was organized to explore this topic in February 2000. A report on the meeting, entitled Intellectual Property Rights in a Knowledge-Based Economy, is presently undergoing review. See also the National Research Council Web site on this topic: www.nationalacademies.org/ipr. |
23 |
Robert Blackburn, “Intellectual Property and Biotechnology” in this volume. |
24 |
See the statement by David Morgenthaler, past president of the National Venture Capital Association, in Panel I of the Proceedings of National Research Council, The Advanced Technology Program: Assessing Outcomes, Washington, D.C.: National Academy Press, 2001. Morgenthaler notes that “[the ATP] is an excellent program for developing enabling, or platform, technologies, |
be concentrated, rather than evenly spread out across different sectors.25 As a result, areas with recognized long-term potential such as biotechnology may experience periods of very low venture capital funding. In part, this may be because they do not offer the returns perceived by investors in other areas such as information technologies and new Internet businesses. Consider, for example, that in 1999, U.S. venture capitalists invested approximately $59.5 billion, up from $4.9 billion in 1993.26 However, eighty percent of venture capital in 1999 was concentrated on Internet and related businesses, with perhaps $1.2 billion invested in biotechnology companies.27 Government support—through funding for promising technologies with less-assured returns—can partly compensate for this concentration.
VI. Biotechnology and Information Technology Present New Technological Challenges and Opportunities. There are a number of research areas in the two fields in which federal R&D funding could be very valuable. These include:
-
Lithography. The semiconductor industry needs new lithography technologies. Within a few years, the size of features on chips will become so small that traditional optical lithography will no longer work. The Department of Energy and the national laboratories have conducted research on extreme ultraviolet (EUV) lithography. Funded originally for defense purposes, this research is having valuable spillover applications in the semiconductor industry. To capture this potential, a government-industry partnership, in the form of a CRADA led by Intel, is collaborating with DOE researchers to advance this technology.28
25 |
Joshua Lerner discusses this “herding tendency” in the venture capital industry. See National Research Council, The Advanced Technology Program, Challenges and Opportunities, Washington, D.C.: National Academy Press, 1999, Introduction et passim. For a broader analysis of the venture capital industry, see Paul Gompers and Joshua Lerner, The Venture Capital Cycle. Cambridge, MA: The MIT Press, 2000. For a discussion of the requirements of early stage financing—in particular institutional requirements—and the limitations of the venture capital industry in this regard, see Lewis M.Branscomb and Philip E.Auerswald, Taking Technical Risks, Cambridge, MA: The MIT Press, 2001, Chapter 5. |
26 |
See the Web site of the National Venture Capital Association, www.nvca.org. |
27 |
Ibid. |
28 |
See David Mowery, Rosemarie Ziedonis, and Greg Linden, “National Technology Policy in Global Markets,” in Innovation Policy in the Knowledge-Based Economy, Maryann P.Feldman and Albert N.Link, eds., Boston, Dordrecht and London: Kluwer Academic Publishers, 2001. |
-
High Performance Computing. The future in high performance computing may belong to machines that are based on standard microprocessors rather than expensive specialty chips and architectures. However, harnessing the computing power of large numbers of microprocessors creates special technical problems. Federal support at DOE and NASA is proving valuable in addressing these issues.29
-
Genomics. The genomics revolution has created a pressing need for improvements in computational biology/bioinformatics. Progress in this new field will depend, in turn, on the ability of universities and others to build and fund multidisciplinary research teams. Yet, the current university environment does little to encourage such teamwork. Along with research needs in bioinformatics, there is also a great need to train people with skills in both biology and computing. In addition to the computing aspects of bioinformatics, important opportunities exist in combining biology and semiconductors into so-called “gene chips,” or DNA analysis chips.
-
Interdisciplinary Instruction and Multidisciplinary Research. Increasingly, the domains of both biotechnology and computing cover more than one disciplinary field. Given this fact, federal policies that encourage research collaboration within universities and among universities, federal laboratories, and companies may be particularly helpful in promoting future U.S. innovation.
VII. The Allocation of Federal Funding among Fields of Research Has Shifted Sharply in the Last Decade. In recent years, federal research funding has increased substantially for a few important and promising fields—e.g., biomedicine and computer science—while it has decreased in real terms for research in the physical sciences and much of engineering. Biomedical research in 1960 accounted for approximately one-quarter of federal non-defense R&D. Today it accounts for approximately half. Meanwhile, research support for most areas of the physical sciences and engineering has fallen sharply in real terms since the early 1990s. For example, federal spending on research in the life sciences increased by 28 percent between 1993 to 1999. Federal support of mathematics and computer science also increased by 45 percent over the same period, while spending on the physical sciences was 17 percent less than in 1993. Federal funding of engineering research
was 2 percent more in 1999 than in 1993, but chemical, electrical, and mechanical engineering research support fell more than 25 percent.30
It is wise, indeed necessary, to invest in a broad portfolio of research, because it is impossible to predict where breakthroughs will come, or how advances in one field will benefit another.31 Research is also becoming more multidisciplinary. The declines in government resources for the physical sciences and certain fields of engineering run counter to continued progress in these fields and to the broader goal of effective multidisciplinary research; they also run counter to the immediate needs of biomedicine.
The development of new medicines is one of the most prominent products of medical research. The development of new medical techniques and new drugs relies heavily on contributions from a variety of sciences. The traditional method of random prospecting for a few promising chemicals has largely been replaced by methods based on molecular structures, computer-based images, and chemical theory. Similarly, synthesis of promising compounds relies on new chemical methods able to generate pure preparations in a single molecule or collections of subtle variants.32 The exploitation of these new possibilities requires the contributions of many disciplines such as mathematics, physics, and chemistry. In short, advances in health depend on a broad range of disciplines, not just the biomedical sciences.
A case in point might be the enormous progress seen in medical diagnostics in the past two decades. Medical diagnostics now rely on imaging technologies such as ultrasound, positron-emission tomography, and computer-assisted tomography. Magnetic resonance imaging, for example, considered by many to be one of the great advances in diagnosis, is the product of atomic, nuclear and high-energy physics, quantum chemistry, computer science, cryogenics, solid state physics, and applied medicine.33
VIII. Priority Setting or Random Disinvestment? For the most part, the shifts in federal research spending shown in Figure 1 have not been the result of a conscious national debate on priorities.34 The R&D budgets of most agencies were cut in real terms in the 1993–1997 period in response to
the end of the Cold War and the national priorities in federal deficit reduction. The agencies cut research programs selectively. For example, DOD, the agency with the steepest cuts in R&D (more than 25 percent), increased funding of oceanographic research and held funding of research in computer science constant. Yet, it cut research investments in other fields, often substantially.35 However, the net impact on the overall pattern of federal investment in research was not planned.
Agencies such as DOD and DOE reduced R&D expenditure in response to overall post-Cold War reductions. As they provided the majority of support for certain disciplines—electrical, mechanical, and materials engineering and computer science in the case of DOD, and physics in the case of DOE—changes in their priority-setting contributed to a national-level shift in federal funding from the physical sciences and engineering to the life sciences and computer science.
Indeed, shifts in agency disbursement priorities were not the result of considered government-wide reviews, either by the Executive Branch or by Congress.36 Declining support for the physical sciences and fundamental engineering worries many observers, including many in biomedicine who know how important chemistry, physics, and mathematics have been to recent advances in molecular biology and genomics.37 Central advances in health and communications necessarily depend on more than increased funding for biomedical research and computer science in themselves.
35 |
DOD cuts in physics and chemistry were 63 and 27 percent, respectively. Funding of electrical, chemical, mechanical, civil, and materials engineering was reduced by 40, 60, 52, 44, and 28 percent, respectively. See Michael McGeary and Stephen A.Merrill, “Recent Trends in Federal Spending on Scientific and Engineering Research: Impacts on Research Fields and Graduate Training,” Appendix A in STEP, Securing America’s Industrial Strength. Washington, D.C.: National Academy Press, 1999. |
36 |
Such a government-wide review has been recommended by various groups examining federal R&D priorities. See, for example, NAS [Press Committee], Allocating Federal Funds for Science and Technology. Washington, D.C.: National Academy Press, 1995, and most recently, United States Commission on National Security/21st Century [Rudman-Hart Commission], Road Map for National Security: Imperative for Change, March 15, 2001. |
37 |
See Harold Varmus, “The Impact of Physics on Biology and Medicine.” Plenary Talk, Centennial Meeting of the American Physical Society, Atlanta, March 22, 1999. At: www.mskcc.org/medicalprofessionals/presidentspages/speches/theimpactofphysicsonbiologyandmedicine,html. |
Recommendations
The recommendations that follow are based on the Committee’s commissioned research, the presentation statements and suggestions made at the conference, and the Committee’s considerable experience and deliberations in assessing government-industry partnerships.
I. Government and industry should expand support of research partnerships and other collaborative arrangements within and among sectors (government, industry, university, and nonprofit) and take other steps to facilitate multidisciplinary research leading to advances in biotechnology and information technology.
-
Support partnerships to exploit the genome. As a first step, the Steering Committee strongly recommends that NIH and other federal agencies work with industry and university experts to identify what technical steps are needed to ensure that the United States can fully exploit the nation’s investment in the development of genomic information. For example, what kind of computing power, what advances in software, what new analytical tools (such as DNA-analyzing computing chips), and what improvements in structural biology will the U.S. need? And what research agendas in science and engineering follow from these findings? What are the “grand challenges,” the most important complex problems, that arise in these fields? Several existing analyses provide good models for preparing these new roadmaps, including the semiconductor industry’s technology roadmaps and
-
those prepared by several industries in partnership with the Energy Department’s Office of Industrial Technology.38
-
Increase government and industry support for university research in microelectronics and related disciplines. The semiconductor industry has maintained, and even accelerated Moore’s Law.39 This has led to a proliferation of computing power throughout the economy, resulting in increased productivity and economic growth.40 To maintain Moore’s Law, significant increases in funding for physical sciences and engineering—including material sciences, chemistry, physics, and electrical engineering—are needed to build greater understanding of properties of nanostructures underpinning tomorrow’s information industries as well to capitalize on advances in biotechnology.41
-
Address unresolved questions about research partnerships. Government-university-industry research partnerships have been very productive in both the computing and biotechnology fields. However, as Congressman Boehlert pointed out at the conference, we now face several important questions regarding partnerships.42 For example, do current partnerships help companies draw upon federally funded university basic research, or are partnerships today increasingly pulling universities away from basic research and towards applied projects of interest to established companies and to startups? How are the new intellectual property rules altering
38 |
See www.semichips.org for the semiconductor industry roadmap. See also www.oit.doe.gov/industries.html for information on the Department of Energy efforts. |
39 |
See Gordon E.Moore, “Cramming more components onto integrated circuits,” Electronics: 38(8) April 19, 1965. Here, Dr. Moore notes that “[t]he complexity for minimum component costs has increased at a rate of roughly a factor of two per year. Certainly over the short term, this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. That means by 1975, the number of components per integrated circuit for minimum cost will be 65,000.” See also, Gordon E.Moore, “The Continuing Silicon Technology Evolution Inside the PC Platform,” Intel Developer Update, Issue 2, October 15, 1997, where he notes that he “first observed the “doubling of transistor density on a manufactured die every year” in 1965, just four years after the first planar integrated circuit was discovered. The press called this “Moore’s Law” and the name has stuck. To be honest, I did not expect this law to still be true some 30 years later, but I am now confident that it will be true for another 20 years.” |
40 |
See Dale Jorgenson and Kevin Stiroh, “Raising the Speed Limit: U.S. Economic Growth in the Information Age,” Brookings Papers on Economic Activity 0(1), 2000. Also see National Research Council, Measuring and Sustaining the New Economy, Washington, D.C.: National Academy Press, forthcoming. |
41 |
See Wesley M.Cohen and John Walsh, “Public Research, Patents, and Implications for Industrial R&D in the Drug, Biotechnology, Semiconductor, and Computer Industries,” in this volume. Also in this volume, see Kenneth Flamm, “The Federal Partnership with U.S. Industry in U.S. Computer Research: History and Current Concern.” |
42 |
See Congressman Sherwood Boehlert, “Opening Remarks” this volume. |
-
relations between universities and companies and within universities? What is the impact of these new partnerships on education within universities? These are important questions, and both Congress and the Executive Branch might encourage further research on these topics.
-
Support mechanisms for conducting multidisciplinary research. Government agencies and the research community also urgently need better information on what types of research arrangements can best support and perform the new type of multidisciplinary research essential to progress in biotechnology, computing, and other “complex fields.”43 Federal support through NSF and other concerned agencies, e.g., DARPA, should encourage and support innovative university-industry-government research partnerships to address emerging problems and reduce gaps in current U.S. educational capabilities. There are promising models that can and should be evaluated. Federal laboratories, for example, have long experience with multidisciplinary teams addressing complex research problems.44 Within academia, new models, such as those being pursued by the University of California at Berkeley, seem promising.45 For example, several universities are building new centers that will physically co-locate researchers from many disciplines to work together on common problems in areas such as genomics. Important questions remain, however, such as:
-
How best to train people in interdisciplinary fields such as bioinformatics;
-
How generally to get researchers in science and engineering to work together effectively;
-
How best to involve industry researchers in these university projects;
-
How federal agencies can work together to support research in fields such as bioinformatics that cut across their traditional boundaries.
-
Federal agencies, scientific and technical societies, and industry associations all have a stake in developing good answers to such questions.
-
Develop public-private technology roadmaps. Federal agencies should be encouraged to work with industry associations and the scientific and engineering communities to develop technology roadmaps in important interdisciplinary fields such as genomics and bioinformatics, nanotechnology, and advanced information technology, including optoelectronics. The country needs to know what major technical barriers exist and what types of research are thus needed in order to ensure continued progress, and what investments in scientific and engineering research will be required.46
II. The scientific community, U.S. industry, and the federal government should explicitly examine the implications of recent shifts in the allocation of federal investment among fields, especially the decline in federal funding for non-defense fundamental research in the physical sciences and engineering, and address possible solutions.
-
Conduct a comprehensive review of the federal research portfolio. Given that future progress in complex fields (such as genomics and next-generation electronics) will depend increasingly on whether the United States is strong in a wide range of scientific and engineering disciplines, the Administration should commission a comprehensive review of how post-Cold War budget trends have affected U.S. research capabilities and personnel in key areas of science and engineering. In the post-Cold War and budget-constrained environment of the early and mid-nineties, many agencies cut or increased their research budgets for mission-related reasons. Those decisions, made independently by each R&D agency in line with its particular priorities, have had major impacts on the nation’s overall research capabilities in science and engineering. The Administration could conduct this review internally, either through the appropriate White House coordinating councils, or by contracting with an outside group. In either case, the review should actively solicit analyses and documentation from experts in science and engineering, from U.S. industry, and from government experts. The review must include a review of needs in engineering research as well as scientific research, given the importance of new engineering work in providing the instruments, analytic tools, software, and other new technologies
-
essential to progress in fields such as genomics, nanotechnology, and computing.
-
Establish an “alert system.” The nation would benefit from an “alert system” that would notify government when important disciplines are getting below critical mass.47 “Benchmarking” projects now underway at, for example, the National Research Council is a possible start in this direction.48
-
Make the case for increased support. If scientific and engineering societies and industry believe the federal government should increase funding for the physical sciences and engineering, they will need to make a strong case to the public and Congress for why such increases are in the nation’s interest.
III. Federal policymakers should support an infrastructure and create an environment conducive to research partnerships and other collaborative arrangements.49
-
Build interdisciplinary competence for multidisciplinary research. Federal agencies can readily undertake some steps that would help build the interdisciplinary competence of researchers in fields such as bioinformatics.50
-
For example, “glue grants,” which encourage researchers from different disciplines to work together on complex problems, can be a low-cost but effective way to encourage multidisciplinary research.51 Short summer courses that bring biologists and computer scientists together also would be valuable. Any resulting initiatives in bioinformatics should explicitly consider the training of bioinformatics experts as well as R&D needs and opportunities.
-
Increase support for interdisciplinary training of graduate students. A key element of national success in new, complex areas such as bioinformatics is training adequate numbers of high-quality graduate students in these fields.52 One problem is that while financial support for graduate students is generally available in the life sciences, such support is becoming more rare in important engineering disciplines. One result is that major U.S. companies are now establishing engineering research centers overseas largely because not enough Ph.D. engineers are available now in the United States. In the future, another result could be a lack of good engineering students and faculty to work on problems such as bioinformatics and nanotechnology. NSF and industry associations should work together to gather detailed information on these issues and consider appropriate steps.
-
Review the impact of patent decisions on technological progress. Congress and the Executive Branch also may wish to participate actively in the NRC review of the impact of recent patent decisions (PTO decisions and court decisions) on innovation in both biotechnology and information technology.53 Uncertainty concerning the scope of new patents in areas such as biotechnology research tools, as well as risks perceived as arising from the proprietary nature of such patents could become deterrents to future innovation in these industries.54
CONCLUSION
The Committee believes that if the government, in partnership with industry, universities, and nonprofit research institutes, adopts these recommendations, the nation will experience greater progress in the research that underlies innovation. This progress will occur not only in the fields of biotechnology and information technology, but in other fields as well, because of the interconnectedness of scientific knowledge. The Committee believes that, in addition to changes in the organization of research, the nation must provide greater support for research across a broad portfolio of fields and disciplines in order to capitalize effectively on existing research investments and to ensure continued benefits for future generations.