Click for next page ( 2


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
Executive Summary Since genetically engineered organisms (GEOs) were introduced into the environment nearly 20 years ago, questions have been raised about the consequences of the escape of those organisms and their engineered genetic materialtransgenesinto natural and managed ecosystems. Ecological research has shown that some GEOs are viable in natural ecosystems and can cross with wild relatives. There also are instances in which transgenes from one domesticated variety can move to others. As a result, there is interest in developing methods to confine certain GEOs and their transgenes to specifically designated release settings. Many confinement methods, including induced sterilization and other methods, are biological in nature, whereas others rely on physical restrictions such as greenhouses or aqua- culture pens. This report refers to these biological methods as bioconfine- ment. Although bioconfinement of GEOs is still largely in the conceptual and experimental stages of development, some methods already have been applied to nonengineered organisms. The primary mechanism in the United States for regulating GEOs and the products derived from them is the 1986 "Coordinated Framework for the Regulation of Biotechnology Products." This framework apportioned jurisdiction over transgenic products by using existing legislation and al- lowed the U.S. Environmental Protection Agency, the U.S. Department of Agriculture, and the U.S. Food and Drug Administration to work together in assessing the safety of the process and products of genetic engineering. In May 2000, the federal government conducted a six-month interagency re- view of its oversight of biotechnology products. The review explored the boundaries of the framework by focusing on several products that were not 1

OCR for page 1
2 BIOCONFINEMENT OF GENETICALLY ENGINEERED ORGANISMS developed until the 1990s, and that therefore were not included in the 1986 framework. The review acknowledged that ensuring confinement could become one of the regulatory requirements for approval and commercial- ization of some GEOs. In 2001, the U.S. Department of Agriculture re- quested that the National Research Council's Board on Agriculture and Natural Resources (BANR) and Board on Life Sciences (BLS) review and evaluate bioconfinement of GEOs. BANR and BLS organized the Commit- tee on Biological Confinement of Genetically Engineered Organisms. The committee's charge was to review and evaluate bioconfinement methods and report on their application in confining transgenic crop plants, grasses, trees, fish, shellfish, and other organisms. The committee's report was to focus on genetic mechanisms, such as induced sterility, but it also was to identify and discuss other available or possible bioconfinement methods. The committee was asked to examine the following questions: What is the status of scientific understanding about various bio- confinement methods for genetically engineered organisms? What methods are available, and how feasible, effective, and costly are these methods? What do we know about when and why methods fail, and what can be done to mitigate those failures? When these methods are used in large-scale applications, what pro- cedures can be used to detect and cull individuals for which the bio- confinement methods have failed? What is the cost effectiveness of these mitigation, detection and culling procedures? What are the probable ecological consequences of large-scale use of bioconfinement methods (e.g., deployment of sterile organisms) on wild populations, biological communities, and landscapes? What new data and knowledge are required for addressing any of these important questions? Although not a focus of the report, the social acceptability of biocon- finement methods is discussed in the introduction and as context for the technical analyses. This report examines bioconfinement of genetically engineered plants, animals, microbes, and fungi. Particular attention is given to transgenic fish and shellfish, trees and grasses, and microbes, because many of those species have been successfully engineered and currently are under federal regula- tory evaluation. Because the committee was not asked to evaluate govern- mental practices or policy, it has limited its discussion to the scientific and societal components that are brought to bear on the process of choosing and applying bioconfinement of GEOs.

OCR for page 1
EXECUTIVE SUMMARY 3 This report consists of six chapters. Chapter 1 defines terms and intro- duces concepts used throughout the report and briefly overviews the history and social acceptability of GEO confinement. Chapter 2 addresses the questions of when and why bioconfinement should be considered and it provides context for the need for and the application of bioconfinement methods. Chapters 3, 4, and 5 review and analyze bioconfinement methods for plants, animals, and microbes, respectively. Chapter 6 reviews the bio- logical and operational opportunities and constraints for bioconfinement and examines bioconfinement failures and their mitigation. Chapter 6 con- cludes with a look to the future, exploring unanswered research questions that will establish better methods for the bioconfinement of GEOs. RATIONALE FOR BIOCONFINEMENT In many cases GEOs will not require bioconfinement, but in some cases they will. The need for bioconfinement should be evaluated on a case-by- case basis. The predominant factors for consideration involve the risks associated with the dispersal of a transgene or transgenic organism into a place, a population, or a biological community for which it was not intended. Significant research efforts on new categories of transgenic plants, insects, microbes, and animals are under way and many of those organisms are being considered for use or release into the environment. Species that dis- perse easily can pose particular risks because of the inefficacy of physical confinement methods and because of the potential for escapees to interact with and harm wild populations. Currently, the most publicized environmental risk associated with transgene dispersal involves the evolution of increased weediness or invasiveness as a result of the sexual transfer of plant crop alleles to wild relatives. When domesticates and wild relatives live in proximity, it is not unusual for natural hybridization to occur. Spontaneous hybridization between nontransgenic crops and their wild relatives already has led to the evolution of several weeds and invasive species such as weed beets in Europe and weed rye in California. It is possible that some engineered genes that confer pest resistance or otherwise improve a crop plant might contribute to the evolution of increased weediness in wild relativesespecially if the genes escape to an organism that already is considered a weed. A transgenic organism itself can become an environmental problem if the transgenic traits it expresses alters its ecological performance such that it becomes an invasive or nuisance species. Many crop plants pose little hazard because the traits that make them useful to humans also reduce their ability to establish feral populations in agricultural or nonagricultural habi- tats. However, feral and naturalized populations are well known for some crops and domesticated animals. If transgenes confer the ability to over-

OCR for page 1
4 BIOCONFINEMENT OF GENETICALLY ENGINEERED ORGANISMS come factors that limit wild populations, the resultant genotype might be significantly more weedy or invasive than is its nontransgenic progenitor. Other concerns about transgenic organisms include their effects on non- target populationsincluding humansand the potential for transgenes to disperse and spread before becoming deregulated in particular regions or nations. Gene flow from GEOs can greatly increase the extent to which nontar- get species are exposed to novel proteins. In food crops, transgenes that code for "novel" pharmaceutical or industrial compounds could be candi- dates for bioconfinement if dispersal from their site of production is possible. METHODS OF BIOCONFINEMENT Many bioconfinement methods have been proposed for limiting the effects of transgenes. Although those approaches necessarily are tailored to specific organisms, and the terminology used to describe them is varied, all bioconfinement methods can be conceptually divided into three general categories: those that reduce the spread or persistence of GEOs; those that reduce unintended gene flow from GEOs into other organisms; and those that limit expression of transgenes. Plants Several existing methods target sexual and vegetative reproduction in plants. For example, sexual reproduction of genetically engineered plants can be blocked by including a gene that renders the organism either perma- nently sterile (nonreversible transgenic sterility) or sterile until the applica- tion of an appropriate trigger is available, such as the use of a chemical spray on a plant (reversible transgenic sterility). Other methods target pol- len to confine pollen-mediated gene flow. Those include methods that achieve male sterility (the inability of a plant to produce fertile pollen) and those that transform chloroplast DNAwhich usually is inherited mater- nallyrather than nuclear DNA. An alternative approach reduces the ef- fects caused by unwanted transgenes by activating a transgenic trait through a specific artificial stimulus, such as a chemical spray (trait-genetic use restriction technology). A few of these and other methods are based on existing agronomic or horticultural practices and have been tested already; many others are newly developed or are only working hypotheses. Although the efficacy of some of the approaches is known, most are untested. Animals Confinement approaches in aquatic species can be achieved through physical confinement, through methods that prevent or disrupt sexual

OCR for page 1
EXECUTIVE SUMMARY 5 reproduction, or through methods that prevent GEO survival in the wild. The induction of triploidy is more established as a technique for finfish and mollusks than for crustaceans. Triploidy is a method that creates, in an organism, the state of having three sets of chromosomes in each cell nucleus, rather than the two typically found in most animal cells, which prevents successful cell division and reproduction. Triploidization is fairly successful and inexpensive, but like all bioconfinement techniques, it cannot guarantee 100% sterility. If only one sex of the GEO is used in the production opera- tion--usually the femalethen the likelihood that a self-sustaining feral population will become established is further reduced. All-female lines often are used for certain commercial species, and their use in conjunction with sterility techniques offers great promise. The use of single-sex lines is not a confinement system on its own, however, if related species that could mate with the GEOs are found nearby. If GEOs are crossed with related species, possibly sterile, interspecific hybrids would result, although thorough test- ing is required to ensure that sterility is close to 100%. Finally, several approaches could reduce the survivorship of GEOs by making them depen- dent on humans, either by genetically engineering the organism so that it requires an anthropogenic substance for its survival or by genetically engi- neering the organism so that it cannot live without an anthropogenic com- pound that blocks expression of the harmful gene. There has not been much research on the bioconfinement of insect species, and so the subject is not well understood. Sterility is relatively easy to produce in insects by radiation, and the techniques used to produce and then quantify sterility are well characterized. However, sterility induced by radiation also can reduce the fitness of individuals. Any significant reduction of fitness would likely render the bioconfinement strategy ineffective within the target population, so transgenic approaches could be used to ensure sterility in insects without the loss of fitness. The large number of insects in any population, however, could make even a small failure rate of sterility techniques problematic. Microbes The two major bioconfinement methods used in microbes are pheno- typic handicapping and the induction of suicide genes. The energetic cost of expressing the genetically engineered trait after phenotypic handicapping causes a loss in those organisms' ability to compete well with indigenous bacteria in soil and aquatic environments. Microbes multiply rapidly and can mutate, however, so subsequent generations of these bacteria might be better adapted to the environment than the original GE strain and be able to coexist with the indigenous populations. The ability to coexist depends in part on the highly variable external environment. Handicapped fungi

OCR for page 1
6 BIOCONFINEMENT OF GENETICALLY ENGINEERED ORGANISMS appear to act similarly to handicapped bacteria in that they cannot compete with indigenous populations. The effectiveness of this mechanism as a con- finement measure is unclear. Another major confinement method for bacteria and fungi is the use of suicide genes. In the case of bacteria, numerous systems have been devised to significantly reduce the population of bacteria upon the addition or removal of a chemical or with a change in the environment. In the case of fungi, several mechanisms have been devised but none have been tested even in the laboratory. In no case has any method been field tested. Suicide gene systems have not been used for viruses. Collectively, the methods remain completely theoretical or have been used in the laboratory, in small test plots, or in laboratory microcosms. ENSURING BIOCONFINEMENT EFFICACY Typically, precommercial evaluation of GEOs starts on a small scale and then is expanded to larger scales before release. It is likely that appro- priate bioconfinement methods will be evaluated in a similar way. Never- theless, these methods will vary in efficacy, depending on circumstances. Each is likely to work well with a specific organism, genotype, or environ- ment, to work poorly with others, and to be inappropriate in yet others. Each is expected to work best on a small spatial scale, and the probability of failure increases with the number of individuals involved and the size of the area they occupy. Likewise, each is expected to work best over short periods of time, and the probability of failure increases with the amount of the time that the organisms are in the open environment. It is likely that no single method can achieve complete confinement on its own. The efficacy of a bioconfinement method will vary depending on the organism, the environment, and the temporal and spatial scales over which it is introduced. Most of the bioconfinement methods for GEOs discussed in this report are in the early stages of development and much is yet to be understood. It is clear that there is a great need for additional information on how well bioconfinement methods work separately and together. The effectiveness of any method will vary with genotype and the environment. Thus, the effi- cacy of combining confinement methods should be tested in representative genotypes that are under development and in environments into which a GEO is likely to enter to ensure that the plan is effective. Before field release, the reproductive biology of the novel genotype should be measured relative to its progenitor to evaluate changes that might affect its rate of gamete and progeny production and dispersal. New geno- types generally do not have reproductive phenotypes that are different from

OCR for page 1
EXECUTIVE SUMMARY 7 those of their parents, but any changes that do occur could have important consequences. Changes in the reproductive biology of a GEO might not be anticipated because of its transgenes, but unanticipated phenotypic effects of a single allele are not rare. To evaluate changes in reproductive biology, the novel genotype should be compared with its progenitor before field release. For long-lived species, such as trees, it may be necessary to begin field tests before such comparisons are possible, with a realistic plan to mitigate any unexpected and dramatic increase in reproduction. Confinement techniques should be experimentally tested, separately and in combination, in a variety of appropriate environments and in repre- sentative genotypes under development before their application. In the case of long-lived organisms such as trees, they should be tested in conjunction with the release or scale-up of GEO products that are considered safe. The evaluation of whether and how to confine cannot be an after- thought in the development of a transgenic organism. Safety must be a primary goal from the start of any project. Furthermore, it is important to consider the dispersal biology and the opportunities for the unintentional movement of transgenes in determining the best choice of an organism for use in creating a GEO. The constant and iterative evaluation of confine- ment options during the development of a GEO should optimize both the efficacy and the cost-effectiveness of the confinement options once they are deployed. Hurried consideration of confinement just before the deployment of a GEO will create a makeshift and expensive plan that might work better in theory than in practice. The need for bioconfinement should be considered early in the develop- ment of a GEO or its products. Many opportunities to mitigate the effects of a bioconfinement failure can be put in place during the earliest stages of development. For example, the act of choosing which GEOs to develop is in fact one form of bio- confinement. An organism that is typically grown to produce a common and widespread food product probably would be a poor choice as a pre- cursor for an industrial compound unless that organism were to be grown under stringent conditions of confinement. This is an important issue for any novel compound or GEO for which zero tolerance of bioconfinement failure is needed. Engineering organisms that are not otherwise used for food or feed could be an effective way to prevent a transgenic compound from entering the human food chain. Alternative nonfood host organisms should be sought for genes that code for transgenic products that need to be kept out of the food supply.

OCR for page 1
8 BIOCONFINEMENT OF GENETICALLY ENGINEERED ORGANISMS Given that no single bioconfinement technique is likely to be com- pletely effective, the use of multiple techniques with different strengths and weaknesses will decrease the probability of failure. Many bioconfinement techniques are in the early stages of development, and some will be unacceptable in various circumstances. Therefore, before a GEO is released, its bioconfinement techniques should be tested in appropriate environments and in representative genotypes under development, and the reproductive biology of the GEO should be elucidated relative to its progenitor. If a bioconfinement method is applied, the committee proposes that a new approachan integrated confinement system (ICS)should be used. ICS is a systematic approach to the design, development, execution, and monitoring of the confinement of a specific GEO. Among its features are a commitment to confinement by top management; the establishment of a written plan for confinement measuresand their documentation and remediation (in case of failed confinement); training of employees; assign- ment of permanent employees to maintain the continuity of the system; development and implementation of standard operating procedures; use of good management practices; periodic audits by an independent entity to ensure that practices are in place; periodic review adjustment to permit adaptive management of the system; and reporting to an appropriate regu- latory body. For ICS to be effective, it is essential that it is supported by a rigorous and comprehensive regulatory regime that is empowered with inspection and enforcement. An integrated confinement system that is based on risk assessment (including the risk of human error) is recommended. There also is a need to defineearly onwhat constitutes adequate bioconfinement. This requires an evaluation of failures, their effects, and their probabilities under worst-case scenarios. It also entails the assumption that escaped genes have the opportunity to multiply. The stringency of the integrated confinement system, including bio- confinement, should reflect the predicted risk and severity of consequences of GEO escape. Because methods can fail, a single confinement method will not neces- sarily prevent transgene escape. For most GEOs, their escape will not pose a risk. In some cases, however, stringent confinement could be warranted, which a single method would not provide. Redundancy involves applying two or more types of safety measures to product design and use, each with fundamentally different strengths and possible vulnerabilities, so that the failure of one safety measure would be countered by the integrity of another. The choice of redundant confinement techniques, including bioconfinement, should consider a list of methods whose characteristics will combine to

OCR for page 1
EXECUTIVE SUMMARY 9 produce the best results. In many cases, this will involve the application of a mix of biological, physical, and physicochemical confinement measures tailored to specific GEOs. In other cases, it may be possible to combine two barriers of the same type but whose failures would be independent events, such that a failure of one barrier does not trigger a failure of the other. It is unlikely that 100% confinement will be achieved by a single method. Redundancy in confinement methods decreases the probability of failing to attain the desired confinement level. The development, testing, and use of GEOs is increasing worldwide. GEOs can move across national borders by a variety of mechanisms includ- ing natural phenomena and trade. No country can manage all of the confinement issues that could affect its environment. An assessment of bioconfinement in any country will require attention to the efficacy of a given method and to concerns about its likely consequencesnot just within that country but in other places as well. Regulators should consider the potential effects that a failure of GEO confinement could have on other nations, as well as how foreign confinement failures could affect the United States. International cooperation should be pursued to adequately manage confinement of GEOs. A bioconfinement scheme will be effective only if it is fully imple- mented, and several factors affect compliance. The efficacy of bioconfine- ment will vary with the human processes involved in applying the technique; the confinement method itself; the characteristics of the GEO; the cost of compliance; the characteristics of the organizations involved; the regulatory system in place; and public transparency. The majority of the bioconfinement methods discussed in this report are in development and have not been used in conjunction with commer- cially available GEOs. Consequently, the public has had little opportunity to develop opinions regarding this aspect of biotechnology. Nonetheless, GEOs or their products can have social significance or be infused with symbolic, social, and aesthetic values that might present important challenges for determining the need for or application of bioconfinement methods. In order to enhance public trust and acceptance of a given confinement strat- egy for a GEO, a sound science-based risk assessment might need to be coupled with a clear and public articulation of any potential ethical concerns. Broad social and ethical values should be considered in assessing the stringency of the integrated confinement system which includes bio- confinement.

OCR for page 1
10 BIOCONFINEMENT OF GENETICALLY ENGINEERED ORGANISMS The public's right to informationoften called transparencyand its right to participate in decision making are fundamental to the practice of democracy. Each right complements the other. Appropriate transparency and public participation can improve the effectiveness of confinement, for example, by informing decision makers about otherwise unknown facts about the environments in which confinement would be implemented, and can increase the acceptance of bioconfinement measures (and of the GEOs being confined) by building trust in the decision-making process. Transparency and public participation should be important compo- nents in developing and implementing the most appropriate bioconfinement techniques and approaches. DETECTING AND MITIGATING BIOCONFINEMENT FAILURE Failures in the bioconfinement of GEOs have not been documented to date, in part because so few methods have been implemented. However, given the imperfections of methods under development and those of methods that have been applied to nonengineered species, it is likely that failure will occur. The degree to which failed confinement events can be monitored and managed depends on whether the GEOs are easily detected, the scale at which they are released into the environment, the GEOs' subsequent popu- lation dynamics, and the degree to which they can hybridize with related species. Early detection of failed methods will be important for mitigating bioconfinement failure, especially if the confined transgenes are likely to spread. Even if a failure is detected early, effective mitigation might not be feasible. Some limited options are available for detecting individuals and culling them after failed bioconfinement. In plants, a failure might be signaled by a distinctive phenotypic trait, such as the presence of flowers on plants that have been engineered to lack them, so workers could cull abnormal plants from small fields. The failure of many bioconfinement methods, however, will be much more difficult to detect. For example, elaborate experiments would be needed to determine whether a repressible seed-lethal transgene is functioning properly. Also, many bioconfined plants will be grown on such large areas of land that repeated comprehensive inspections will be imprac- tical. In the future, DNA "fingerprints" could be linked to bioconfined transgenes to function as "bio-barcodes" that could be detected and used to cull GEOs. Remote sensing approaches might also be available to detect GEOs. It is feasible to detect and then cull individual fish in which triploid sterilization induction fails before they are transferred from secure hatcheries to much less secure facilities, such as outdoor ponds or open-water cages. Economies of scale and possible automation could reduce the cost of such

OCR for page 1
EXECUTIVE SUMMARY 11 efforts. A similar approach can be applied to oysters and shrimp. To detect and cull failures in bioconfinement of fish, shellfish, or insects, one could screen for proteins expressed by the key gene involved or for a co-inserted marker gene. Nonlethal detection might be possible for larger organisms or with such marker genes as green fluorescent protein; detection in smaller organismsespecially insectswould be more likely to require lethal sam- pling. It is not currently possible to detect or cull microbes if bioconfinement fails. The committee did not speculate about cost-effectiveness because genetic engineering-based bioconfinement methods are theoretical or at an early stage of development. Current methods for detecting and culling individual GEOs after a bioconfinement failure are very limited, and they depend on the organism and scale of the original release of the GEO. For large-scale GEO releases, effective monitoring will be essential for mitigating failure. Currently, monitoring is difficult because it involves searching for what often will be a rare event over a potentially large area. In the future, organisms might be purposefully transformed with additional constructs for monitoring. Ideally, monitoring methods would be devel- oped that could identify escapes with remote sensing. Monitoring should be seen as a complement to confinementnot a replacement for it. That is, the act of monitoring should not result in complacency about the possibility of a bioconfinement failure. Easily identifiable markers, sampling strategies, and methods should be developed to facilitate monitoring of bioconfined GEOs in the environment. ECOLOGICAL CONSEQUENCES OF LARGE-SCALE USE OF BIOCONFINEMENT Many bioconfinement methods might be successfully used and result in certain GEOs having negligible effects on wild populations, biological com- munities, or ecosystems, but there has been little research on this topic. Some methods have been used in nonengineered organisms in the past, often with other goals, and they were considered in the committee's evalu- ation. Those methods include growing male-sterile crops for hybrid seed production, small-scale rearing of sterile fish, and releasing sterile male insects that mate with wild females as part of biocontrol strategies to reduce pest insect populations. Two related areas of ecological concern about the use of bioconfinement with GEOs were identified: the large scale at which bioconfined organisms could be released and the possibility that even carefully planned, integrated bioconfinement methods could fail. In some cases, the area over which sterile or handicapped GEOs are released could be large enough to affect

OCR for page 1
12 BIOCONFINEMENT OF GENETICALLY ENGINEERED ORGANISMS biodiversity. In salmon and other species, the presence of large numbers of sterile GEOs or those with reduced fitness in some cases could threaten local biodiversity. There is concern that some native populations of animals might lose the ability to compete for food or mate successfully in the presence of more competitive or more attractive, but sterile, GEOs. If this were to occur in small populations, depressed levels of natural reproduction could threaten the long-term survival of native genotypes. In other cases, the large-scale release of sterile GEOs could have the beneficial effect of alleviating existing problems, such as the loss of genetic diversity that can occur when modernand often genetically uniformcrop plants (or hatchery- raised fish) interbreed with rare wild relatives or locally adapted varieties. A more general problem with all bioconfinement methods is that occa- sionally they could break down, especially if they are intended to confine millions of free-living individuals. Depending on the original reasons for using bioconfinement, the ecological consequences could be serious. If a bioconfined GEO can become a pathogen or an invasive species after the breakdown of an Integrated Confinement System, the decision to release it on a large scale should be scrutinized with extreme caution. If the reason for using bioconfinement is mainly commercial, the ecological effects of bioconfinement failure could be of no consequence. It is difficult to general- ize about the ecological effects of large-scale releases of bioconfined GEOs, and further research should address these questions in relation to specific realistic conditions. Research is needed to characterize potential ecological consequences of bioconfinement methods and to develop methods and protocols for assess- ing environmental effects should confinement fail. CONCLUSIONS The current lack of quality data and science is the single most signifi- cant factor limiting our ability to assess effective bioconfinement methods. In many cases GEOs will not require bioconfinement, but when they do the need for bioconfinement should be evaluated case by case, considering worst-case scenarios and the probability of their occurrence. The evalua- tion of whether and how to confine a GEO should be an integral part of its development, and the need for bioconfinement should be considered early in the process. It is unlikely that any single bioconfinement technique will be completely effective, and using multiple techniques with different strengths and weaknesses will decrease the probability of failure. Further- more, many bioconfinement techniques still are in the early stages of devel- opment, and the possible unintended consequences of some bioconfinement methods mean that some technologies will be unacceptable under certain circumstances. Therefore, before a GEO is released, the techniques to be

OCR for page 1
EXECUTIVE SUMMARY 13 used should be tested in a variety of appropriate environments and in representative genotypes under development, and the reproductive biology of the GEO should be understood relative to that of its progenitor. If a bioconfinement method is applied, an integrated confinement system should be put in place. Such a system must be supported by a rigorous and compre- hensive regulatory regime empowered with inspection and enforcement. Finally, in order to implement effective bioconfinement of GEOs, the committee recommends support for additional scientific research that characterizes as completely as possible the potential ecological risks and consequences of a failure in bioconfinement develops reliable, safe, and environmentally sound bioconfinement methods, especially for GEOs used in pharmaceutical production designs methods for accurate assessment of the efficacy of bio- confinement integrates the economic, legal, ethical, and social factors that might influence the application and regulation of specific techniques models (using models that are calibrated and can be verified experi- mentally) the dispersal biology of organisms targeted for genetic engineer- ing and release, where sufficient information does not exist. Interdisciplinary research will improve the future of biotechnology by developing new confinement methods that minimize the potential for unin- tended damage to human health and the environment. The success of these efforts will do much to bolster public confidence in the continued growth, development, and opportunities presented by biotechnology.