Review of the Department of Energy's Genomics: GTL Program (2006)

As noted in Chapter 1, the Department of Energy (DOE) is charged with promoting scientific and technological innovation in support of its overarching mission to advance the national, economic, and energy security of the United States (DOE, 2005a). DOE is a key supporter of non-health-related biological research, and DOE’s scientific strategic goal emphasizes both the production of new knowledge through science and the creation of new research capabilities.

It is the assessment of the committee that the goals of DOE’s Genomics: GTL program are consistent with both elements of DOE’s scientific strategic goal. But as currently envisioned, Genomics: GTL is focused almost exclusively on microorganisms. Although work on microbial systems is well justified, plants also represent a major pathway to the production of bioenergy, they play an important role in carbon sequestration and global nutrient cycles, and they are potentially useful for bioremediation. Thus, the absence of targeted research within Genomics: GTL on relevant aspects of plant biology is a serious omission. Consistently with that view, the Energy Basic and Applied Sciences Act of 2005 calls for an emphasis on both plants and microorganisms. DOE already has a modest investment in energy-related aspects of plant biology in its Energy Biosciences Program. The committee suggests the inclusion of plant biology research in the Genomics: GTL program where appropriate.

In Genomics: GTL, DOE proposes to use systems and synthetic biology approaches to achieve a predictive understanding of microorganisms and to mine these untapped resources. The approaches are well matched to DOE’s history as the founder of the Human Genome Project (Roberts, 2001) and represent a logical extension of DOE’s scientific vision and capabilities. No other federal agency

is playing a lead stewardship role with respect to microbial systems and synthetic biology. In addition, Genomics: GTL’s somewhat unique practice of funding large teams with large grants is essential, and the committee believes that it should be maintained. The Genomics: GTL program has produced valuable scientific results (DOE, 2003, 2004, 2005c), and its planned research promises to generate additional important and useful results. The committee enthusiastically endorses DOE’s plan to enlarge the program to $200 million per year for basic research and further endorses the focus of this research on long-term goals for energy production, environmental remediation, and the mitigation of global climate change. In particular, the committee notes that the ability to produce cost-competitive ethanol from cellulose and hydrogen biophotolytically from water or fermentatively from other carbon substrates, the development of biological solutions to the many recalcitrant problems of legacy wastes, and the attainment of an increased understanding of the role of microbial communities in global carbon cycling to enable the development of carbon-sequestration techniques for addressing climate change are all worthy goals that are highly suitable to DOE’s missions. The committee wholeheartedly supports those goals and the conclusion that the best way to achieve them is through a systems biology approach. We further endorse, with enthusiasm, the ambition of the Genomics: GTL initiative to place DOE at the forefront of systems biology research, as it has been in genomics.

Recommendation 1: The committee recommends that DOE and the nation give high priority to genomics research aimed at achieving DOE’s mission goals.

However, the committee disagrees with the plan to create four facility types—for protein production and characterization, characterizing and imaging biomolecular machines, proteomic analysis of microorganisms, and modeling of microbial community cellular systems—sequentially. In place of that plan, we propose a set of integrated, problem-oriented genomics-enabled facilities that will focus on pioneering technologies rather than duplicating existing technologies.

The committee agrees that the technologies associated with each of the four proposed facilities are critical to providing fundamental and quantitative understanding of biological processes relevant to the long-term missions of DOE and the sought-after predictive capability for biological behaviors at all levels from molecules to whole microorganisms. However, we believe that to advance DOE’s missions and to strengthen the nation’s systems biology research capacity, a

parallel and integrated approach to creating facilities should be taken instead of a sequential approach.

First, DOE’s argument for the sequential creation of the four new facilities is flawed. It rests on an assertion that the facilities will greatly speed the achievement of the program’s long-term goals for research to advance energy production, environmental remediation, and carbon sequestration. In their absence, DOE asserts that the achievement of the goals could take as long as a century because the scientific advances needed to reach productive applications of new technologies to the problems will be severely retarded by the lack of appropriate tools. Department officials have expressed to the committee the view that the creation of the facilities could direct and speed the research efforts to achieve the desired goals in far less time—for example, in 30 years. But in the current plan, the time for each facility to be designed, constructed, and come on line for operation is estimated to be about 6 years because DOE assumes that budgets for the facilities will be appropriated one at a time. That suggests that the complete program capability cannot be achieved until the completion of the fourth and final facility roughly 24 years from now. If for any reason there is a delay in the appropriation of funds for the creation of any of the four facilities, even 24 years may not be enough time to reach full program capability. The committee therefore finds DOE’s estimation of the time for the program to come to full capability and produce the desired result highly unrealistic under the current plan. The reductionist approach, moving from simple to increasingly complex systems in a period of 24 or 25 years, is not an efficient way to achieve the program’s goals. Building four individual facilities in sequence is fraught with too many uncertainties and requires much too long a time. Even if we assume that the conglomerate of different facilities will, in the end, be able to come together to achieve useful progress toward solving the cutting-edge problems being addressed by the Genomics: GTL program, the time to completion is so long that there is considerable risk that the systems biology train not only will have left the station but will be at some other station when everything is finally on line. In this fast-changing era of genome-enabled science, it is not sufficient to require that many years be spent in pulling together the tools needed to make progress. That approach is unlikely to place DOE in a leadership position, either intellectually or in terms of research results.

Second, the proposed configuration for the Genomics: GTL facilities is based on an underlying assumption that large-scale generation of reagents and data is the best way to apply a systems biology approach to the grand challenges of the Genomics: GTL program—bioremediation, bioenergy, and carbon sequestration. That approach first parses the organisms that might be subjects for study into their components, assembles a parts list, inventories the parts, and then reassembles them into interacting complexes, networks, and pathways. Underlying the plan is the belief that data and reagents abstracted from the individual organ-

isms will enable DOE’s missions. The implied assumption for the current facility model is that the best way to solve the problems in systems biology is to operate several separate facilities that concentrate on specific aspects of the question. The committee finds that assumption questionable. Systems biology is a cutting-edge field because the properties of organisms and groups of organisms cannot be predicted from simple considerations of the behavior of their parts. Therefore, obtaining much more information about the parts is not the best way to reach the goal of systems biology. A more consistent view would be that the properties of complex systems can best be understood by studying complex systems. Such studies need to be conducted by interdisciplinary teams of experts that include genomicists, geneticists, physiologists, biochemists, biophysicists, computer scientists, engineers, and mathematicians.

The rationale for the current plan appears to be a direct analogy with genomics, especially large-scale sequencing, as practiced during the last decade. The committee finds, however, that the analogy is weak at best. The economies of scale that made large genome projects successful have not yet been obtained in work with proteins or in most aspects of systems and synthetic biology. It is not clear when technological advancements that deliver such economies of scale will become available for protein analyses and systems and synthetic biology. Hence, DOE should be cautious about embarking on construction of large-scale infrastructure that assumes that such economies are readily available.

We also find a lack of equivalence, for example, between the high-throughput production and characterization of individual proteins as a means to advance systems biology and the effect of DNA sequencing on characterizing the genome. Understanding what the various parts of an organism do is one aspect of understanding the organism as a whole; and knowledge of component functions is useful, and perhaps essential, for modeling the behavior of the organism, but it is insufficient for predicting how various component functions work and interact together. Hence, if the short-term goal of the Genomics: GTL program is to be able to predict the behavior of systems so as eventually to engineer microorganisms to serve the nation’s energy-related needs, this model does not address the most important needs of such a program, nor does it push the research frontier forward. The goals of the protein-production facility are modest compared with its size and budget and can easily be met much less expensively either by several components in different types of facilities or by outsourcing to academic and commercial facilities (Box 3-1). The proposed protein production facility would be doing today’s research, not tomorrow’s.

It is true that the Human Genome Project was both important and timely because it provided reams of data, but that was before we had sequence information. Now that we have it, additional data-gathering and reagent production, although necessary for the field as a whole, will not propel the field forward in the same way that genome sequences did, because the cutting-edge questions are very different. Furthermore, sequence information can lead to immediate leaps in

understanding through analysis of individual genomes but especially through comparison of genomes. Such understanding does not come from a shopping cart full of proteins, however useful they may be to bench scientists. The payoff from such resources is a long-term one, and this greatly lessens the intellectual impact of DOE’s contribution. Simple service functions provide useful tools but will not themselves advance understanding in a major way. They reduce DOE to the role of enabler rather than leader. As stated above, the knowledge gaps in systems biology go beyond an understanding of the biochemical role of each protein in an

organism and beyond an understanding of how the proteins associate with one another and how the associations modulate individual function. The greatest progress in systems biology can be achieved only by solving the problems of integration of function at the level of pathways, organisms, and colonies and communities of disparate organisms. The committee concurs that for the Genomics: GTL program to achieve its mission goals, the facilities model must directly aim to address those needs.

The sequential-facilities model appears to be based on a belief that the barriers to achieving the short-term and long-term goals of the Genomics: GTL program are related primarily to the ability to obtain material (such as proteins) for study. As a consequence, DOE’s proposed plan is technique-driven. It is the view of the committee, however, that the real barriers are related to our ability to study and understand complexity and that a more problem-oriented approach is needed. Our list of key barriers includes barriers to measuring metabolite flow and other biological characters in vivo; to an understanding of interactions in communities of microorganisms that would enable prediction of the effects of introducing new metabolites or new or engineered organisms into the community; to a predictive understanding of how organisms respond to the introduction of new enzymes and pathways; to our ability to model microbial behavior in many developmental states, including stationary phase and sporulation, and in proliferation; to an understanding of transient states; to the acquisition of tools, including mathematical models and concepts, to enable prediction; to the use of general principles; and to adequate characterization of microbial diversity in the target ecosystems. The timely execution of the Genomics: GTL program and the achievement of DOE’s mission goals are much better served by investing in comprehensive research programs that drive technology development than by investing primarily in infrastructure.

Because the currently proposed model is technique-driven, the facilities proposed for it are fairly specialized, single-purpose facilities. We believe that it is incorrect to assume that an interdisciplinary problem can be addressed best with a set of cooperating, but independent, specialized facilities, especially if they are created sequentially. We believe that interdisciplinary problems require interdisciplinary approaches from the outset.

To understand what an optimal model might look like, we consider the facilities needed to meet the short-term Genomics: GTL goal of predicting the properties of microorganisms. Examining several aspects of alternative models in the light of that mission provides a framework for a different model.

• Vertically integrated facilities vs. reductive specialization. It seems inherently contradictory to assume that a coherent approach to modeling the properties

of microorganisms will “emerge” of its own accord from the activities of a set of autonomous facilities. If we assume that such an interdisciplinary problem requires, at a minimum, good communication among researchers from many backgrounds, that alone would favor a vertically integrated structure in which such scientists all work on a common set of problems at the same facility. Another advantage of integrated facilities is that basic scientists would be working with engineers, so discoveries would be translated into applications efficiently.

• Institute vs. service model. As anyone who has watched the transition of a startup biotechnology company from a research-oriented to a customer-driven business can tell, a service facility is concerned primarily with solving day-to-day technical problems and meeting the demands of those it serves. If its technology becomes outdated or less important, the service facility is marginalized in its field. An institute, in contrast, is concerned with the big picture, can contain people of many backgrounds, is not wedded to a particular technology, and can adjust its strategies more easily as its field progresses. It seems logical that moretalented people will want to work in the institute model and that an agency that operates the institute will be more prestigious and have more influence than one that runs a service center.

• Partnerships and integration with a local academic or research community vs. an autonomous DOE-only model. One of the biggest obstacles to progress in any organization is the “not invented here” effect. The tendency of many institutions to look inward restricts progress and retards the integration of new technologies. Thus, a facilities model in which entities outside DOE participate is likely to be more open-minded, flexible, and innovative than one that draws its personnel, tools, and infrastructure exclusively from one source. In addition, a DOE initiative that is isolated from the broader biological community may not be taking full advantage of outside expertise or of opportunities to communicate its own expertise to others. The committee suggests that the optimal placement for Genomics: GTL facilities would be close to strong academic or industrial research establishments. That should be a key criterion in selecting the location of one or more of the proposed Genomics: GTL facilities; proximity to a community of cutting-edge biological research is essential to the long-term success of the Genomics: GTL program. The exchange of ideas and information among academe, industry, federal agencies, and other research entities will leverage human, scientific and financial resources. Although it might be more feasible for DOE to locate facilities near a national laboratory, the committee notes that not all national laboratories are close to strong academic or industrial research establishments.

• Parallel vs. sequential. The advantages of a sequential model are that one can learn from mistakes and build on success. But we conclude that the disadvantages of the sequential model outweigh those potential advantages. One disadvantage is the impossibility of correct timing. Built into a sequential model are assumptions about the pace of advances and the changing needs of a field over

time. No one can predict either with any confidence. Thus, there is a great risk that by the time the first facility finally comes on line, it will be obsolete technically or irrelevant to the needs of the later facilities, which will have changed because of outside developments over which the program has neither influence nor control. Another disadvantage is that only the last facility in such a model interacts with all the others from its inception; the growth of many facilities in isolation from others breeds a culture not of cooperation but of independence. “Parallel” facilities would avoid those pitfalls.

What alternative model follows from those considerations, and why? The committee proposes a model that has the following features:

• It consists of institute-like facilities, each focused on a particular problem or theme in systems biology that will advance both the short- and long-term goals of the Genomics: GTL program. Some possible research areas include remediation of plumes of toxic waste, understanding and predicting the consequences of adaptive evolution (including that originating from horizontal gene transfer), modeling microbial communities, and developing a systems-level understanding of microbial stationary phases, sporulation, and other non-proliferative developmental states.

• Each facility develops technology to support its research. Technology development is driven by research needs.

• The model is vertically integrated to address a problem in concert with many techniques, some of which are developed there. Some work will be collaborative, and some will be contracted. Outsourcing to acquire materials and technology could be a cost-efficient strategy that allows more rapid acquisition of critical capabilities. It is important that a vertical focus directs priority-setting toward specific ends in a manner that is likely to be more or less peculiar to individual challenges.

• Each facility is near major academic, private, or federal research centers. Being close to institutions that have excellent biology, mathematics, and computer science programs is important because access to researchers with related interests allows the facility to draw more completely on and leverage the expertise and interests of surrounding universities, institutes and industries. It involves public-private partnerships with both academe and industry. And it integrates the Genomics: GTL program into the larger biological research community.

In facing the reality of rapid scientific progress and the timeliness of the proposed approaches, the committee strongly encourages DOE to rethink its user facility construction plans and create institute-like facilities each of which combines the capabilities of the original four planned types in a vertically integrated manner. That would allow each facility to tackle all aspects of a problem or small set of problems in parallel and potentially to achieve goals more quickly. Table 1

in Section 5 of the Genomics: GTL roadmap document provides three “conceptual science roadmaps” for microbial energy and environmental processes, and Table 2 in Section 5 provides three more for natural systems (DOE, 2005b). The first vertically integrated facility could reduce its focus from all six conceptual science roadmaps and concentrate on only one or two. For example, the first facility could concentrate on fuels by working on the systems biology of cellulose conversion and the production of hydrogen and high-hydrogen fuels by using sunlight. That focus would be consistent with the committee’s view that investment in bioenergy is inadequate despite the urgent need to find alternatives to petroleum-based fuels.

A concentration on the science underpinning fossil-fuel replacements would have two advantages. First, technologies would be developed with a greater emphasis on the needs of alternative-fuel research than on serving the scientific community as a whole. The synergy between researchers and technology developers should also speed technology outcomes and keep them at the cutting edge. Selection of organisms to be studied should be based on their application to bioenergy production, irrespective of whether they are common or rare. Therefore, DOE could take a lead role in identifying and developing key organisms into model systems for systems biology research relevant to bioenergy production.

Second, experience gained from building and running the first facility could be used to refine approaches and improve planning and execution of the second facility. The second facility would focus on one or two additional aspects of the items listed in the tables in Section 5 of the roadmap document. A third and perhaps fourth facility, if deemed necessary for continued progress, could be planned and constructed by using further design and execution refinements in the longer term. In that way, work on some of the initial problems could be reaching long-term goals long before the estimated 24 years proposed for full design and construction of consecutive facilities had elapsed.

It would be necessary for DOE to select problems for each facility in the chain to focus on as it comes on line. The committee suggests that there are compelling reasons for DOE to give top priority to the creation of a “bioenergy institute” that also focuses on carbon cycling in the context of bioenergy. First, the U.S. (and global) economy is increasingly vulnerable to oil shocks caused by political unrest, terrorism, and natural disasters. A recent analysis conducted by Securing America’s Future Energy and the National Commission on Energy Policy concluded that even “small incidents” that reduced global oil supply by 4 percent would cause oil prices to increase dramatically to more than $161 per barrel (SAFE and NCEP, 2005). Second, as several National Research Council reports have concluded, we need to act now if we are to have any chance of stabilizing greenhouse-gas emissions (NRC, 1992, 2003a), given the 100-year residence time of carbon dioxide (CO₂) in the atmosphere and the 30- to 50-year lifetime of capital stock in the energy industry. Because reduction of CO₂ in the

atmosphere and carbon cycling are related to the use of bioenergy, the bionergy institute should have a secondary focus on carbon sequestration. Third, recent advances in biology—such as rapid sequencing, directed evolution, and whole-genome synthesis—may enable us to design biological systems that can generate affordable, carbon-free energy (see Chapter 2), and thereby to reduce the costs of mitigating CO₂ emissions (IPCC, 1997). Later facilities could have carbon sequestration and bioremediation as their primary foci.

The revised model would greatly improve the cost effectiveness and efficiency of DOE’s investment and optimize the achievement of useful scientific results. The committee agrees that there is a sound scientific case to be made for systems biology at DOE, but the approach to the order, scope, and scale of the facilities needs rethinking. The committee’s alternative model could deliver some of the hoped-for scientific output more quickly and more efficiently.

Recommendation 2: DOE should revise its plans for creating four single-purpose technology-driven facilities in sequence. Instead, DOE should create up to four institute-like facilities that each contain all the capabilities of the original planned facility types—protein production, molecular imaging, whole-proteome analysis, and systems biology—in a vertically integrated manner. Each facility should focus on one or two of the DOE mission objectives and develop short-term, medium-term and long-term goals to chart a course for the program. Short-term milestones should be used as a metric for independent evaluation.

The committee concurs that having a single physical space for the first integrated facility is important, although some of the investigators should be encouraged to participate from remote sites. Proximity will enable a large-scale research program with efficient, coordinated, and complementary activities among a team of investigators who have diverse expertise and no physical barriers. Day-to-day interactions among the assembled scientists in various disciplines (biology, chemistry, physics, nanoscience, computing, and engineering) are essential to overcome cross-disciplinary barriers and to generate novel approaches to and productive outcomes for the intended goals. The Institute for Systems Biology and the Department of Systems Biology of the Harvard Medical School are two examples of integrated facilities that facilitated and advanced biomedical sciences.

Later phases of facility implementation will be based on the experience gained in the first facility and on progress in the proposed research at the time of funding. It is conceivable that there will be more than four facilities at the end, with different emphases and strengths and consistent with DOE’s goals. Some of the future facilities should be flexible enough to include more than one site per

facility so as to leverage additional resources and partnerships that would broaden the portfolio of the Genomics: GTL program.

Our proposed alternative plan of implementation is based on the following merits.

• Vertically integrated facilities would establish the Genomics: GTL program in a leadership position to launch a world-class, comprehensive, integrated research and training program in systems and synthetic biology. They would create a paradigm shift in biological research that will integrate data from a broad spectrum of spatial and temporal scales to advance understanding of biological phenomena to be able to predict or alter capabilities for optimal performance under field conditions. The facility would provide an intellectual and physical environment for both multidisciplinary teams and individual-based research.

• The research programs of the facilities would be built on overarching biological themes relevant to the DOE missions in energy production, environmental remediation, and carbon sequestration. They would involve diverse disciplines, including genomics, genetics, physiology, biochemistry, structural and computational biology, nanoscience, and engineering. The facilities would provide an intellectual and physical environment for both multidisciplinary teams and individuals pursuing research in relevant missions.

• New technologies would be developed in the facilities on the basis of well-justified scientific problems. The technologies will be aimed at particular ends rather than being ends themselves.

• The successful development of the integrated facilities would attract investigators around the country to use them. The resulting scientific discoveries and technology development can be expected to benefit not only a subset of biologists but a broad spectrum of scientists and engineers in different disciplines. Because of the diverse disciplines of the investigators, the integrated facilities are likely to have complex organization charts. That will leverage the experience of DOE to administer this new research enterprise.

• Modeling plays a central role in studying and understanding complexity. New computational approaches and tools would be developed in the facility to promote synergy between modeling and experimentation at both bench and field level.

• The first facility constitutes a pilot to validate the hybrid systems approach and to identify roadblocks to be addressed in the later facilities. All facilities can be designed so that they will not be outdated by the rapid pace of scientific discovery and technological development.

• The vertically integrated facilities lend themselves to a staged investment with expandable bases, flexibility to shift directions without losing prior investments, leverage, and open-source positioning.

The committee has concerns about other aspects of DOE’s plan. The locations for the user facilities should be selected in an open and all-inclusive competitive process that provides adequate opportunities for universities and industry to partner with DOE and its national laboratories. The committee strongly believes that such an open and inclusive competition will provide DOE with a robust and enabling facility that could strongly support systems biology research. The criteria for the selection of contractors should include innovations of the project plan relevant to DOE missions, investigator qualifications, management organization, educational outreach, technology dissemination, strategy for maintaining the interface between Genomics: GTL research and industry and other entities that conduct translational research, intellectual-property management plan, proximity to a concentration of high-caliber participating scientists, and possibly provision of matching funds by the applicant institutions.

The committee encourages DOE to consider cost sharing by applicant institutions for design and construction because timely establishment of the proposed facilities is crucial in the fast-moving field of systems biology. For example, allowing the successful applicant to fund new construction or renovation of an existing facility upfront could greatly speed up the process. As federal funds become available through the appropriations process, DOE could then “lease-purchase” the facility to eventually acquire it from the private sector partner. If DOE lacks the authority to support that kind of transaction, the committee recommends that it request a waiver from current policy. That could be consistent with Congress’s recent conference-report language directing DOE to accelerate the deployment of all four Genomics: GTL facilities.

The committee notes that there is much vacant space in buildings in localities that were once targeted for major development by the biotechnology industry. Reuse of existing space, rather than new construction, would have the advantage of speeding the establishment of new facilities. It could also help to improve access to the facilities for academic and industry scientists by locating the facilities on private land off DOE reservations. The Joint Genome Institute, which is on non-DOE land in Walnut Creek, California, is an excellent model for a facility with an appropriate degree of openness to encourage the scientific community to regard it as a user facility. The committee feels that the greatest mistake would be to create new user facilities behind the fences of some of the more remote existing DOE laboratories (for example, Hanford and Los Alamos) that lack proximity to major centers of biotechnology research. The Genomics: GTL program will not achieve DOE’s mission goals unless it is embedded in a culture of strong basic biology and innovative biotechnology. To ensure the program’s success, DOE should consider locating the facilities close to universities or federal or

private research institutions that have established centers of excellence in biosciences and biotechnology.

Recommendation 3: DOE should consider locating user facilities on private land off DOE reservations to allow an open-access policy and close to research institutions that have established programs or centers of excellence in biosciences and biotechnology. The locations for the user facilities should be selected in an open and all-inclusive competitive process that provides adequate opportunities for universities and industry to partner with DOE and its national laboratories.

The committee recognizes that the proposed integrated Genomics: GTL facilities would have a logical linkage to various large-scale enterprises, such as the nanoscience and high-performance computing programs supported by DOE, NSF, the National Institutes of Health (NIH), and the National Institute of Standards and Technology. Those programs are developing novel technologies that can be of immediate value to the proposed integrated Genomics: GTL facility. For example, breakthroughs in nanofabrication techniques for engineering single-molecule confinement devices coupled with optical interrogation systems are making it possible to explore life processes on a new scale of physical and biological reality (Box 3-2).

The committee recommends that the major effort of an integrated Genomics: GTL facility should be to pioneer new technologies. For example, purifying a soluble protein from a bacterium is now relatively straightforward. Therefore, a facility should focus on developing experimental protocols for understanding and manipulating proteins that are resistant to purification. Membrane proteins will be an important target of investigation because over 30 percent of the human genome is thought to code for them, but membrane proteins have not been well characterized; they make up only about 0.3 percent of the solved high-resolution structures. Such proteins play critical roles in various cellular processes, including signal transduction, ion and metabolite transport, and maintenance of chemical and electrical balance inside the cell. Membrane proteins have been of immense interest to biotechnology companies because of their potential use as drug targets. Their biophysical characterization is urgently needed.

A facility will enable scientists and engineers to address important and intricate metabolic subsystems that are used in such activities as converting sunlight into cellular energy. Manipulation of such an intricate metabolic subsystem inevitably involves understanding of a subset of protein components and coupled reactions that define a major subsystem. The identification and purification of those complexes, which may be transient and labile, would require a major research effort to extract them as functional units for various biophysical-biochemical characterizations. A full understanding of a complex will allow scien-

tists to implement a rational engineering design to make the subsystem more predictable, controllable, and efficient. To introduce the engineered subsystem with a higher efficiency into a cell system, it is equally important to understand how a particular subsystem interacts with other subsystems inside the cell. A whole-cell approach will be necessary to understand the networking among different subsystems in a living cell. That argues strongly for the importance of simultaneously studying molecular components in the cellular context.

All the projects in the proposed integrated facilities would involve multiple investigators, so a user-friendly data-management system across domains of in-

vestigation will be essential to keep track of all the new information and will not only guide next experiments but allow the tracking of progress. It is a challenging project in itself to generate an electronic notebook and database for the heterogeneous data and to cope with the volumes of data that will be generated. We encourage computational engineers and investigating scientists to codevelop the laboratory data-management system to meet specifications of the research. Constant and iterative discussions and software testing between the computational engineers and the scientists will be necessary to produce a practical product.

Because there is no such vertically integrated facility dedicated to studying biofuel, bioremediation, and carbon management, there is no direct competition with other agencies or anyone else. On the contrary, it is conceivable that a successful applicant will be able to partner with other investigators to leverage some of the existing but smaller-scale operations of various modules of technologies in protein and machine purification, mass spectroscopy, crystallography, electron cryomicroscopy, light microscopy, laboratory information management, and computer simulation and modeling.

The recent NIH roadmap initiatives for medical research have a number of programs—such as structural biology, computational biology, bioinformatics, molecular imaging, nanomedicine, building blocks, and biological pathways and networks—that would be complementary to the Genomics: GTL Program. Although NIH’s missions are peculiar to human health, many of the methods developed with NIH support are generic and adaptable for studying other organisms of interest. NSF has also been supporting nanoscience and technology research centers and programs, which have technology-development components relevant to the Genomics: GTL program. Neither NIH nor NSF is supporting a large-scale and integrative approach as Genomics: GTL is planning to do. A synergistic coordination among those funding agencies would have the potential to push the technologies more efficiently in advancing our fundamental understanding of life and in improving our quality of life and environment. DOE has had collaborative programs with other agencies in the past (for example, the Human Genome Program with NIH, and the maize-sequencing program with NSF and USDA). For Genomics: GTL, DOE should be strongly encouraged to coordinate with and leverage the programs of the other federal agencies with common interests in microbial biology (NSF), bioremediation (Environmental Protection Agency), biofuels (U.S. Department of Agriculture), and genomics (NIH) (DOE, 2005d). The complex and often labile nature of the biological systems to be studied will present new challenges for any of the technologies and will justify investment in further development and refinement.

At the same time, it will be important for Genomics: GTL to coordinate effectively with related programs in DOE. For example, as Genomics: GTL

acquires additional emphasis on relevant aspects of plant biology, recommended by this committee throughout this report, it will be critical to achieve close coordination with plant-related research funded through DOE’s Energy Biosciences Program. The DOE Energy Biosciences Program has a long history of supporting basic research in plant and microbial biology. Although the program is small, the quality of the research programs supported by it has been very high. Because both the Energy Biosciences and the Genomics: GTL programs are administered by the Office of Science, it should be easy for these two programs to cooperate. In fact, as the Genomics: GTL program begins to add plant biology to its research portfolio, it makes sense to use the close connections with academic plant scientists already established by the Energy Biosciences Program, as opposed to reinventing such expertise in Genomics: GTL.

In addition to collaborations within the United States, DOE should also consider international collaborations with other countries that have similar genomics programs—for example, Genome Canada and the Netherlands Genomics Initiative. Moreover, bioenergy and carbon management are subjects of global interest, so international collaborations would reduce duplication of effort and leverage each country’s resources and expertise.

Because of the rapid development of various technologies in the proposed Genomics: GTL facilities, part of the role of the Genomics: GTL program is to sponsor regular workshops and symposia to disseminate the new experimental and computational methods to the broad community. Such activities have been well conducted in the DOE laboratories in various disciplines—for example, contractor-grantee workshops of the Human Genome Project and Genomics: GTL program and Sandia National Laboratory’s Workshop on Computational Molecular Biology.

The proposed Genomics: GTL facilities could provide unique and unusually rich technology environments for junior scientists to be introduced into research, to get excited about research related to our environment and energy production, to learn specialized techniques, and to develop research careers in microbial biology. Therefore, training components that reach out to high-school students, undergraduates, graduate students, and postdoctoral fellows should be included in the Genomics: GTL program. The training program will focus on recruiting the most talented young people from various educational and ethnic backgrounds.

Such training programs would benefit not only the trainees but also the research staffs of the Genomics: GTL facilities. Because the proposed facilities are not didactic teaching institutions, the presence of trainees will provide an intellectual environment in which senior researchers will be confronted by students who question and challenge the basic premises of a scientific approach. Often research ideas can be generated from simple questions born of curiosity.

Such training components will add important intellectual dimensions to the proposed facilities.

DOE is already sending brochures about its programs to high schools (for example, Your World: Biotechnology and You), but a selective and well-constructed summer training program for high-school students and teachers will not only reach out to the students but also enhance the quality of science education. Immersion, in which teachers join scientists to conduct experiments, has been identified as a key strategy in the professional learning of teachers (Loucks-Horsley et al., 1998). DOE already has the Laboratory Science Teacher Professional Development program (DOE, 2006), which provides an immersion experience for highschool teachers. Genomics: GTL facilities could be included in the settings for the program. Training for students and teachers would allow the Genomics: GTL program to identify talented young people and ensure the nurturing of their talent. Student participants could become leading scientists, and their training would help to ensure the quality and leadership of bioresearch in DOE and other bioscience enterprises in both academe and industry.

The undergraduate research experience is valuable for students who are exploring career options. Many graduate students who enroll in highly competitive graduate programs in the United States have already had research experience in their undergraduate years through specialized summer research programs or through faculty-supervised research in universities. Both DOE and NSF have sponsored competitive and successful summer research programs—for example, Science Undergraduate Laboratory Internships at DOE and Research Experiences for Undergraduates at NSF. The Genomics: GTL program should create a prestigious undergraduate summer research program that would introduce students to the importance and excitement of microbial systems biology as applied to energy or environmental problems. Such experience may attract this pool of talented undergraduates to select a research career path consistent with the vision of the Genomics: GTL program.

DOE’s national laboratories have a long tradition of providing laboratory space and research resources for graduate students to conduct their PhD thesis research in physics and chemistry aimed at degrees from academic institutions hundreds or even thousands of miles away. Similar arrangements could be adopted in the Genomics: GTL program. The senior staff of the Genomics: GTL facilities would play the role of comentors of the graduate students enrolled in participating universities. NIH has recently implemented the Graduate Partnerships Program (GPP), which links NIH with graduate programs of various universities around the country (NIH, 2005); students can work in one of the intramural laboratories at NIH to conduct part of their thesis work for a period of weeks to months. NSF’s Integrative Graduate Education and Research Traineeship (IGERT) program also seeks to train PhD candidates to become scientists and engineers who have the interdisciplinary background and the technical, professional, and personal skills needed to address the global questions of

the future. GPP and IGERT are models that DOE could consider adapting to its own purposes. Programs similar to those will allow the Genomics: GTL program researchers to gain access to a broad pool of talented graduate students. A successful graduate training program is an effective way of disseminating the Genomics: GTL resources and technologies to the broader biological community.

A postdoctoral fellow is a scientist at the stage of developing his or her own independent research. Postdoctoral work takes place during a critical period of typically 3-5 years during which one must make the transition from graduate student to independent investigator. The productivity of postdoctoral fellows is generally high. Because systems biology is a developing field, many universities and industries will probably be seeking qualified scientists with relevant expertise to fill their new faculty and staff positions. The Genomics: GTL program will have the opportunity to become a primary source of such people. Supplying talented scientists to the marketplace of ideas is one of the best ways to establish the credibility and reputation of the Genomics: GTL program.

The committee suggests that DOE consider innovative arrangements to encourage the use of the Genomics: GTL program and its facilities as training grounds for the next generation of scientists. This is an especially important aspect of a program that promises to deliver a high degree of interdisciplinary cooperation.

In addition to training students and postdoctoral follows, the Genomics: GTL facilities should provide long-term and short-term sabbatical fellowships for faculty from universities and industries. The constant influx of visiting scientists will provide expertise and perspectives complementary to those of the Genomics: GTL staff. The intellectual contribution of well-established investigators not only will benefit the science done today but also will generate the new approaches and ideas of tomorrow in the Genomics: GTL program.

Recommendation 4: DOE should consider partnering with universities and other federal agencies to develop programs that use Genomics: GTL institute-like facilities as training grounds for the next generation of scientists.

The grand challenge articulated in the Genomics: GTL program mission is to understand the “molecular machines of life” that underlie key processes in bioremediation, carbon sequestration, and biofuel production. The more immediate goal of the program is to develop predictive models of system function. If robust, such models would enable efficient re-engineering or optimizing of molecular machines to solve the nation’s energy and environmental challenges. The committee wholeheartedly supports the goal of the program and concludes that the best way to achieve them is through a systems biology approach. We further

endorse, with great enthusiasm, the ambition of the Genomics: GTL program to place DOE at the forefront of systems biology research as it was in genomics.

The committee finds, however, that the proposed configuration of the user facilities is not optimal for achieving the goals. Worse, it risks leaving DOE as a follower, rather than a leader, in systems biology. We offer, as a constructive suggestion, an alternative model that we feel would better serve the Genomics: GTL mission. The alternative model would consist of several problem-oriented facilities that are vertically integrated and institute-like (Box 3-3). Each facility would work on research problems or themes chosen to propel the field of systems biology as a whole, and its applications to the grand challenge in particular, rapidly forward. To the greatest extent possible, the facilities would be built in parallel and brought on line rapidly. Although technology development would be an important part of all the facilities, it should be driven by the scientific questions being addressed, not by the need to produce reagents in bulk to serve a wider community. Our model places the facilities close to and in collaboration

with other centers of front-line biological research so that they can draw on and exchange expertise with colleagues in the wider community.

DOE has done an exceptional job in leading the country in many fields of research that were at the cutting edge, in particular the Human Genome Project. DOE now has an opportunity to become a world leader in systems biology through the Genomics: GTL program and by integrating or connecting fundamental research data to other programs in DOE and other national and international agency programs. The committee commends DOE for its development of the Genomics: GTL program and encourages the DOE administration to consider the committee’s proposed alternative plan for the Genomics: GTL facilities. The recommendations in this report, if implemented, will enhance DOE’s potential for success in its three critical mission areas: bioenergy, bioremediation, and carbon sequestration.