Capitalizing on New Needs and New Opportunities: Government-Industry Partnerships in Biotechnology and Information Technologies (2001)

This analysis is based on the belief that biotechnology and information technology are the two most innovative industries in the United States today, and that, while each is distinct in character, they display important complementarities. Indeed, the links and synergies in their innovation processes are such that progress in one sector increasingly depends in many ways on progress in the other. Innovation, which maintains U.S. technological leadership and helps to sustain economic growth, depends in turn on a policy framework that encourages basic, applied, and multidisciplinary research. Continued U.S. technology leadership will depend on the ability of the government, universities, and industry to collaborate effectively in the development of appropriate policies, especially for these innovative sectors.

It is for this reason that issues in biotechnology and information technology have been integral to the STEP Board’s broader program of study of Government Industry Partnerships for the Development of New Technologies since the inception of that effort. This section reviews the new needs and opportunities in biotechnology and information technology as discussed by experts from industry, academia, and government. We begin by noting that the U.S. government has historically supported biotechnology and information technology in different ways. While this has led to significant technological progress, this success has necessarily created new issues requiring policymakers’ attention. The second section looks at some future technologies that rely on complementarities between biotechnology and information technology and identifies some of the challenges that must be overcome if we are to realize the full potential of these technologies.

Government-industry and university-industry partnerships have a long history in the information technology sector.¹ In contrast, the biomedical sector has seen a greater division of labor, with government funding basic research and industry concentrating on applying these research results to develop new pharmaceuticals and other products.² Even so, as the distance from fundamental research to applications has shortened in biomedical research—already a well-established trend in the information technology sector—the demand for partnerships and other forms of close collaboration has grown. Partnerships have also been a mechanism for promoting multidisciplinary research—another strong trend underlying research in both sectors. Finally, differing uses of patent protection provides another perspective for reviewing the impact of public policy on the biotechnology and information technology sectors. In the sections below, the roles played by partnerships, funding, and patents in each sector are highlighted by various participants at the conference.

In his opening remarks, Gordon Moore gave a personal tour of the evolution of government support to the semiconductor and computing industries. He emphasized that the government aid to these industries was through general support of science and technology, as well as through serving as an early market for innovative devices. He also underscored the importance of the government’s policy on intellectual property protection for the growth of the semiconductor industry. As an example, he cited the government’s requirement that AT&T license its semiconductor technology as part of a 1956 antitrust decree, observing that this was a key catalyst for the growth of the industry. He noted that the government still plays an important role in the computer industry through the purchase of high-performance computers.

Direct federal support of R&D, Dr. Moore noted, did not play as large a role in the semiconductor industry’s early evolution as in computers, although the government contributed substantially as a reliable early buyer of high-end semiconductor devices.³ As the U.S. semiconductor industry matured, it adopted an innovative means of addressing the need for industry-related research by establishing the Semiconductor Research Corporation in 1982. This program initiated cooperative research efforts between the industry and selected universities.⁴

In response to severe qualitative and pricing competition from Japan, the industry campaigned and secured agreement to create a consortium to redress its competitive position. Called SEMATECH, the consortium represented a substantial and innovative effort by a fiercely competitive U.S. industry. It began in 1987 with matching funds from the federal government and industry, with each contributing $100 million a year. Reflecting the tremendous growth and commercial success of the American industry, SEMATECH’s membership decided, in 1996, to “sunset” the federal contribution and fund the consortium from its own resources—albeit at a reduced level.⁵ In 2000, it became International SEMATECH with the inclusion of members from Korea and Europe.⁶

Government-industry cooperation on research and development continues through a variety of mechanisms, noted Dr. Moore. For example, cooperative R&D agreements have proved effective mechanisms for encouraging cooperative research and development with the National Laboratories. Research and development in priming and lithography technologies—such as that in Extreme Ultraviolet Lithography (EUV) —are likely to prove critical for the future growth of the world’s semiconductor industries. EUV promises a means of continuing the remarkable pace of advance in the capacity of computer chips, as captured in

Moore’s Law. EUV Lithography is being developed under a Cooperative Research and Development Agreement (CRADA) between the Sandia and Lawrence Livermore National Laboratories and the leading SEMATECH companies, such as Intel, Motorola, AMD, and Infineon, to which industry contributes $100 million a year.⁷

Expanding on Dr. Moore’s discussion, Kenneth Flamm provided evidence of recent trends in government funding for computers.⁸ The computer industry benefited from significant federal support in the years following the Second World War, with the government encouraging private sector initiatives, serving as the primary market for most early machines, and developing much of the basic architecture of today’s computer.⁹

As the commercial market for computers grew in the 1960s, the government’s role in supporting R&D evolved, with the government continuing to push the leading edge of technology through support for advanced research and as a consumer of high-performance machines.

Recent federal support for innovation in this sector, however, has declined. Dr. Flamm reported that, since 1990, government’s support for computer hardware R&D, already at historically low levels, declined substantially. Further, industry investment also fell, both in absolute terms and as a percentage of computer sales. At the same time, both federal and industrial R&D investments have shifted over time from long-term basic research to near-term targeted projects.

In concluding, Dr. Flamm argued that the information technology sector is not yet a mature industry, that innovation in the computer industry remains an important source of economic growth, and that government-industry collaboration, highly productive in the past, should be encouraged in the future.

In contrast to semiconductors, where the government has employed a variety of agencies and mechanisms, the biotechnology industry has relied heavily

on a single government agency to underwrite basic research support. In this partnership of government, industry, and universities much of the technology commercialized by the biotechnology industry has been developed in the university setting, with funding from the National Institutes of Health. This relationship, however, is changing rapidly as the industry evolves.

Whereas the first 20 years of the biotechnology industry were dominated by two technologies—recombinant DNA and techniques to produce monoclonal antibodies—new technologies now are based on a wider range of scientific disciplines. These are becoming important to both the biotechnology and pharmaceutical industries.¹⁰ DNA chips and the growing output from the Human Genome Project are just two examples of recent developments that require new responses from government, industry, and universities.

Describing trends in bio-pharmaceutical research, Leon Rosenberg emphasized the interaction among industry, government, and university. Although the government and industry fund about 90 percent of bio-pharmaceutical research, universities conduct much of this research. Universities act as an important source of R&D output, in addition to providing trained scientists. Continued advances in bio-pharmaceutical research will depend on advances or applications of new information technologies, notes Dr. Rosenberg, and it is important that academic R&D retain its emphasis on basic research. Even with the growing role of industry in funding university research, Dr. Rosenberg affirmed that academic research has remained focused on basic research rather than on applications.

The intellectual property regime, which structures the incentives to innovate and the ability to capture innovation’s fruits, is another important piece of the policy puzzle. There have been a number of changes in intellectual property policy and practice in recent years.

Research by Wesley Cohen and John Walsh, presented at the conference and included in the report, shows the varying importance of patenting across the biotechnology and computing sectors. Their research looks at the impact of the 1980 Bayh-Dole patent and trademark act, which extended patent protection to

publicly funded research, such as that conducted in universities.¹¹ They found that patenting is especially important in the biotechnology sector, because it is often the sole asset of a young start-up venture. On the other hand, patenting seems to be less important in the computing and information technology sectors, because companies in these industries tend to rely on trade secrecy and long lead times to protect intellectual property. In that situation, giving universities exclusive rights to research results will have little effect and may even inhibit commercial exploitation of publicly funded research, noted Drs. Cohen and Walsh. This is because it can inhibit the free flow of information between universities and firms. They conclude that policymakers should take into account the various methods that firms in different sectors employ to protect intellectual property when considering changes in intellectual policy law.

In exploring the patenting activity of universities in the post-Bayh-Dole era, Maryann Feldman of Johns Hopkins University found that patenting activities at some universities seem to be governed more by past practices and cultures than by Bayh-Dole. For most universities—including those most actively generating patents—a few patents generate most of the license revenue. Therefore, she found that widespread patenting efforts by universities may yield only modest benefits in terms of licensing revenue. At the same time, Dr. Feldman noted that such patenting practices raise the transaction costs of university-industry collaboration and thus can have the effect of discouraging collaboration.

Dr. Feldman also found that universities and companies have developed a broad range of mechanisms for exchanging information in the post-Bayh-Dole era. In addition to exclusive licenses, arrangements include sponsored research, consulting arrangements, recruitment of students, equity arrangements, support for start-up companies, and even gifts to university endowments. As Drs. Cohen and Walsh also observed, the types of transactions depend on the sector, technology, and size of the company.

Apart from reviewing past support for the biotechnology and information technology sectors, conference participants and paper authors identified a variety of policy concerns related to the federal research portfolio and patent laws.

Speaking at the conference, William Bonvillian, from the staff of Senator Joseph Lieberman, noted that overall, federal support for R&D fell nearly 9 percent in real terms from 1992 to 1997. These cuts, however, were not evenly distributed, he noted, with the impact in some disciplines (including some fields related to information technology) more severe than in others. For example, from FY 1993 to 1997, in real terms, mathematics was down by 6 percent, electrical engineering by 36 percent, and physics by 29 percent.¹² Federal support for health and biomedical research has increased, while support has declined for research in the physical sciences, mathematics, and many fields of engineering, on which advances in biomedical research has depended and will depend, he noted.

Although the allocation of federal research funding among scientific fields should shift, ideally, with changes in needs and opportunities, Mr. Bonvillian argued that the recent shifts in funding were due more to organizational changes than to changes in scientific opportunity. For example, the steep decline in the defense budget for R&D—a major source of funding for the physical sciences and engineering—has played a central role in these shifts in federal funding. At the same time, the fact that many of the most important scientific advances in biotechnology and information technology come from cross-disciplinary research argues for sustained funding across a broad range of fields. He recommended that policymakers develop an “alert system” to signal when R&D funding levels for certain disciplines fall below critical levels.

The analysis by Michael McGeary (in this report) —which documents recent trends in federal funding of research related to biotechnology and IT—supports this view. The data indicate a substantial shift in the federal research portfolio over the 1990s. McGeary raises “the question of whether the federal research portfolio has become ‘imbalanced.’” He finds that federal support for biomedical research has expanded substantially relative to that in most fields of the physical sciences and engineering. Noting that overall federal funding for research was flat in real terms from 1993 to 1997,¹³ McGeary finds that the Administration and Congress over that period increased the budget for the National Institutes of Health—which provides more than 80 percent of the federal support for the life sciences. While this development is in itself positive, the research

budgets of DOD and DOE—which together provide the majority of funding for research in electrical engineering, mechanical engineering, materials engineering, physics, and computer science—have suffered substantial reductions in the same period.¹⁴

At the conference, Tom Kalil, formerly of the White House National Economic Council, described how the Clinton administration’s effort to enact a 28 percent increase in spending on information technology research for FY 2000 was unsuccessful, despite support from House Science Committee Chairman James Sensenbrenner. While the administration won some increases for the NSF and DARPA, Congress chose not to fund research for high-end computing in the Energy Department. Mr. Kalil said the administration had hoped to raise the level of funding for information technology researches in FY 2001. Indeed the current administration has proposed a $2.6 billion increase in funding for R&D in FY2001, including major initiatives in nanotechnology and information technology (including development of a second terascale computer for civilian researchers), as well as substantial increases for NSF and NIH.¹⁵

In explaining this trend, Leon Rosenberg observed that the bio-pharmaceutical community has done a good job in building a constituency for research funding among policymakers in Washington. He noted that the NIH budget doubled in the 1990s, and the increase of 15 percent in FY 1999 marked the first year of a five-year effort, supported by most members of Congress, to double the 1998 NIH budget by 2003.¹⁶ Mr. Bonvillian, in this regard, observed that the biotechnology industry has done an excellent job in taking its message to Con-

gress. He pointed out that their success stands in stark contrast to the more limited success of proponents for research support in information technologies.¹⁷

Mr. Bonvillian went on to note that this growing support for the biomedical enterprise has two interconnected explanations. First, it may be that the focus of much of the biomedical research activity of one agency—the National Institutes of Health, (which also benefited from exceptionally able leadership through the nineties) —contributed to the ability of proponents to focus policymakers’ attention. Second, a series of high profile research efforts, notably the Human Genome Project, captured the public’s imagination in the 1990s, much as the exploration of space did in the 1960s. This initiative has helped provide support from the general public for increased funding.

In contrast, noted Mr. Bonvillian, support for computer research is scattered among many government departments and agencies with diverse objectives, constituencies, and modes of operation. In addition, investments in health R&D are meant to lead to cures for diseases that afflict us all, including members of Congress. The fact that much of the NIH research effort is extramural and widely distributed among the nation’s universities may tend to augment its already broad-based support.

The pace of innovation in the biotechnology industry has caused a strain in the patent system, noted Robert Blackburn of Chiron Corporation. He recounted that in some instances, courts have been applying patent law doctrine developed for the synthetic chemical industry to biotechnology—including patents on DNA sequences and other things found in nature. This practice is problematic when applied to biotechnology, he noted, since biotechnology directly concerns research tools, rather than consumer products. Furthermore, patents covering research tools are difficult to enforce and cause friction between university and industrial researchers. In conjunction with limited resources of the Patent Office and with only intermittent legislative guidance from Congress, the patent system can sometimes have the unintended effect of discouraging drug development. Mr. Blackburn recommended, therefore, that Congress undertake an extensive review of patent law.¹⁸

A leading theme of the conference was the importance that each field, biotechnology and IT, has for the other. Typically, we think of the critical role that information technology will play in fulfilling the promise of the genetic revolution. Just as importantly, but less well understood, biological systems and concepts have much to offer the information technology and other high-technology fields such as materials and nanotechnology. The spillover effects of scientific advances propelling the biotechnology industry help the computing industries and vice versa. This holds great promise for the economy and for society at large.

Marvin Cassman noted that the tools of molecular genetics, structural biology, genomics, and other fields are creating an enormous amount of information about humans and other living organisms. With the necessary technology now in hand, he noted, scientists can begin to understand biology as a “wiring diagram” —a complex system that is dynamic and richly interconnected. This contrasts with the discipline’s traditional approach to assessing biological organisms using “parts lists.” To understand an organism’s complexity, biologists will have to process the huge amounts of data that will flow from research. This is where bioinformatics and the tools of information technology enter the picture.

As we see below, organizations including the Department of Defense and NASA are already interested in research on small and lightweight high-performance technologies to meet their own needs.

In his luncheon keynote speech, NASA Administrator Daniel Goldin, observed that the marriage of biology and information technology is key to space exploration in the future. Travel to Mars will take many years, and sophisticated medical treatments—drawn from biotechnology and remote communication systems—will be necessary to detect, diagnose, and cure health problems that astronauts might encounter. Administering medical treatment over the immense distances of space will require remote sensing, highly sophisticated software, and advanced robotic technology. Developing such technology is, perhaps, the largest challenge facing biologists and computer scientists today.

In Mr. Goldin’s view, biology has tremendous potential—through biometrics, bioinformatics, and genomics—to change electronics, computational hardware and software, sensors, instruments, control systems, and materials. Developments in biology will also call for new platform concepts and system architectures to integrate these technologies, he noted. They will be key to the development of more complex and self-correcting—that is, more intelligent—

technologies that are also smaller, lighter, faster, more adaptable in changing environments, and less power-intensive.

Even as succeeding administrations and agencies (including NASA, NIH and NSF) lay the groundwork for bridging the gap between biology and computing, others are already looking to the future. In doing so, they help to chart what policymakers must do today so that the full potential of current innovations can be more fully exploited.

Functional genomics, which Mark Boguski discussed, is an example of an approach to closing the gap between biology and computing that will push innovation in computing, while enabling biology to realize the benefits of the Human Genome Project. Functional genomics uses computing technology to model organisms (using data from the human genome) so that these organisms can be understood as a system. Rather than comparing a few genes, functional genomics will allow scientists to compare many genes—or systems of genes—across organisms in order to make inferences about their function in the body and to determine what proteins are involved. This will require, however, a great deal of computing power and the development of sophisticated new visualization and simulation technologies.

Miniaturization of devices is another area that will rely on biology and computing as it offers exciting possibilities in medicine, energy efficiency, national security, and environmental protection. Al Romig of Sandia National Laboratories reported that miniaturization will proceed both by making today’s components smaller and by taking micro-scale materials, including biological materials, and building up from there. Thus, he said, not only will the physical sciences help biomedical research, but also biological research will lead to advances in the physical sciences. He is confident that over the next 10 to 20 years, a growing number of materials and devices on the market will be manufactured using biological methods.

In the field of nanotechnology, for example, nanocrystals have been injected into living cells. These crystals do not damage cells. Rather, they emit light enabling scientists to explore further the properties and behavior of cells. Nanotechnology will also enable the creation of more efficient materials, which should reduce energy consumption with major benefits for the environment.

Timothy Coffey of the Naval Research Laboratory discussed how miniatur-

ization in military applications will contribute to national security. Nanoelectronics and microelectromechanical systems (MEMS) will enable electronic systems to carry out essential battlefield functions that humans do now at their peril. For example, micro-air vehicles are being developed to jam enemy radar systems, and biosensors will detect the presence of dangerous substances in the environment. As battlefield requirements increase, the military will look to biological systems for solutions, but developing those technologies will rely on advances in computing and electronics. In turn, this will require the confluence of a variety of research disciplines, including the biosciences, materials science, computers and information technologies, and electronics.

Other interesting developments at the intersection of biology and information technology were described by Jane Alexander of the Defense Advanced Research Project Agency (DARPA). One such example is an effort to develop an “electronic dog’s nose” that can sniff out explosives. If adequately portable, it can be carried into the battlefield by soldiers. Another initiative would be to intervene at the larval stage of insects to enable them, when mature, to detect explosives. DARPA is also considering putting electronic chips on insects in order to track their hunting patterns for use in developing search algorithms for Department of Defense sensors.

The combination of breakthroughs in computing technologies along with the sequencing of the human genome offers exciting possibilities. Paul Horn of IBM observed that in the future, for example, doctors will prescribe medications tailored precisely to a patient’s problem and genetic makeup. To do this, new information technology will be necessary.

Government support for basic research will be indispensable in creating it. New bridges must be built between basic research and industry, Dr. Horn said, and government-funded partnership programs—such as Engineering Research, Science and Technology Centers, the Advanced Technology Program, and the Technology Reinvestment Program—have and should play a role in speeding the exchange of information between universities and industry.¹⁹ Government can also help by providing the information infrastructure, such as the development of very high-end computers that are needed to turn the vast amount of data information from the Human Genome Project into knowledge and understanding.

Dr. Romig took a similarly positive view, noting that multidisciplinary government-university-industry partnerships are essential for progress in nanotech-

nology research. He predicted that the federal government will have a prominent role as a supporter of research, and as an owner of national laboratories.

Scientific advance is inherently interconnected; neglecting one area of inquiry can jeopardize advancement in other fields of inquiry. Progress in solving research problems requires interaction across disciplines as well as across supporting government agencies, industry groups, academic organizations, and other nonprofit associations. Capturing the full potential of existing investments requires additional multidisciplinary research collaboration among biologists and computer scientists, as well as among those active in related disciplines, such as physics, chemistry, and electrical and chemical engineering.²⁰ While the research community seems well aware of these challenges, more needs to be done to educate policymakers on this issue.

A further implication is the need for interdisciplinary training. This is clearly the case in biotechnology—particularly in the new field of bioinformatics. In fact, one of the challenges to analyzing bioinformation and promoting collaboration is the shortage of bioinformatics specialists. Paula Stephan and Grant Black, in a paper in this report, provide explanations for this problem: They note that while bioinformatics is inherently multidisciplinary, institutionalized practices at universities often inhibit cross-departmental collaborations. Further, they note that information technology professionals, often in great demand in their own fields, find few incentives to undertake additional training in biology.

Encouraging more interdisciplinary training is a concern that is being addressed on various fronts. Rita Colwell, Director of the National Science Foundation (NSF), noted that the NSF recognizes the challenges facing those studying entire biological systems. The NSF now funds efforts to develop more bioinformatics tools and seeks to encourage cross-disciplinary collaboration, she noted.²¹ Dr. Marvin Cassman described the initiatives of the National Institute of General Medicine Sciences, which he directs, in establishing a new Center for Bioinformatics and Computational Biology to support interdisciplinary training in conjunction with biology. The center will assume oversight of the NIH’s Biomedical Information Science and Technology Initiative. Dr. Cassman also described related NIH initiatives—such as “glue grants,” designed to bring biologists, physical scientists, and information technologists closer together—that

are intended to encourage collaboration across disciplines.²² Dr. Penhoet described the Berkeley Health Sciences Initiative—a major effort to promote such cross-disciplinary collaboration. This initiative brings together a variety of disciplines in the same location to address research problems, such as improved gene chips and computational methods for analyzing the data that the gene chips produce.

Building bridges across disciplines is a long-term effort. To succeed, it will require leadership and innovation as well as other new partnerships between government, industry, and universities. It will also require appropriate allocation of federal R&D resources and, in some cases, new institutional frameworks to foster sustained technological.

In his conference presentation, Congressman Sherwood Boehlert focused on the practical aspects of allocating scarce federal R&D resources to help realize the possibilities of biotechnology and new information technologies. He said that the question is not one of whether the government should support biotechnology and information technology research. Rather, the question is one of how and at what level such support is to be realized. These were bound to be “the trickiest questions” in R&D policy, he observed. Mr. Boehlert said that he believed partnerships between government, industry, and academia would be part of the answer.

Mr. Boehlert suggested a careful examination of partnerships—especially their effects on the important role of universities in long-term research and the production of scientists and engineers—to see whether these partnerships produce commercially viable results. He also brought up the question of whether

the current allocation of research support among scientific fields was appropriate. He urged members of the research community to take their views on investment in research not just to members of Congress on science committees, but to all members of the House and Senate.

Several ongoing partnerships illustrate why increased collaboration among sectors in the biomedical area is so important. The Alliance for Cellular Signaling, a government-industry-university partnership, was established in September 2000 with a glue grant from the NIH. Dr. Cassman described this as a new mechanism to increase multidisciplinary collaboration and inter-institutional sharing of resources.

Other current partnerships include the Mouse Sequencing Consortium (MSC) and the Single Nucleotide Polymorphism (SNP) Consortium. MSC, also begun in 2000, supports sequencing at several academic and nonprofit research institutions. Six NIH institutes, along with the Wellcome Trust and several pharmaceutical and biotechnology companies, fund MSC.²³ The SNP Consortium, created in 1999, seeks to map genetic variations within the human genome in order to increase the understanding of their role in causing disease. SNP was formed at the initiative of the Wellcome Trust, which provided substantial funding and convinced 10 competing companies to contribute funding.²⁴ In both cases, as with the Alliance for Cellular Signaling, the results are being made public. The goal is to provide wide access to basic data that not only will speed up progress in research but also disseminate information that industry can use to develop new products.

This summary provides an overview of the issues the Committee sought to address at the conference. It underscores the need for additional analysis and careful monitoring of national policy on resource allocation in the biotechnology and information technology areas. These issues are presented above in an abbreviated form as a means of providing the background and necessary context for the Committee’s Findings and Recommendations as detailed in the section that immediately follows.