Principles of Plant Genetics and Breeding

Introduction
ISSUES IN THE APPLICATION OF BIOTECHNOLOGY IN PLANT BREEDING 265
Industry highlights
The intersection of science and policy in risk analysis of genetically engineered plants
David A. Lee 1 , and Laura E. Bartley 2 *
1 EPA Office of Research and Development, 8623N, Washington, DC 20460, USA;
2 USDA-APHIS Biotechnology Regulatory Services, Riverdale, MD 20737, USA
In our view as scientists recently entrenched in the policy world, the regulation of genetically engineered (GE) plants illustrates the
challenges of applying science to policy for a number of reasons. The laws and implementing regulations regarding the oversight
of GE products require that regulation be science-based; however, risk analysis always requires some judgment on the part of
decision-makers. In addition, the United States does not have an overarching biosafety law and instead uses existing laws to regulate
GE organisms. Further, despite the objective of US policy that regulation of GE organisms be product-, not process-, based, in some
instances different regulations and standards are used to oversee organisms that are GE as opposed to those that are not. Below we
elaborate on these ideas and discuss the data used in risk assessment. Despite the tensions that exist within the US biotechnology
regulatory system, our educated opinion is that the system is generally working to allow responsible use of this technology.
Risk analysis: a science-based endeavor that relies on values
Risk assessment fits into a larger process known as risk analysis, which has the following three components: risk assessment, risk
management, and risk communication. Risk assessment involves the highly scientific and analytical process of identifying hazards
that might result from an action and evaluating their probability and consequences, as simplified in the following equation:
risk = (probability of a hazard) × (consequences of a hazard). However, human judgments are often required to complete the
assessment of risk because typically gaps exist in the data regarding the probability and consequences of potential hazards. Risk
management is the process in which decisions regarding the action are made and any identified risks are managed, mitigated, or
monitored for. Value-based judgments are particularly required in decision-making to balance the risks of action with those of
inaction. Risk communication is the part of the iterative risk analysis process in which information regarding the assessment and
management steps is made available to, and discussed with, interested parties.
To briefly illustrate risk analysis for plant genetic engineering, an example of a hazard that is often discussed is that of a gene
introduced through genetic engineering being passed to a wild relative of a crop, termed “gene flow”, and causing a change in the
population of the wild relative. In this case, the probability of gene flow occurring has been shown in many cases to be
significantly greater than zero (Ellstrand 2003). However, while geneticists can describe scenarios that would lead to changes in
plant populations due to the introduction of a single gene, the probability of those changes and the long-term consequences of
such changes on ecosystems and for the preservation of crop biodiversity are unknown in most cases. In addition, some would
argue that these risks need to be weighed with the difficult-to-assess risk of not using a particular GE variety, which could lead
to changes in wild relative populations and ecosystems through other means, such as habitat destruction due to inefficient agriculture.
Thus, despite the legal requirement that decisions regarding GE products be science-based, there are many decisions
regarding data gaps and value judgments that must be made in the course of risk analysis.
Law meets science: product- versus process-based regulation and some consequences of the use of particular laws for
the regulation of biotechnology in the USA
The 1986 Coordinated Framework for Regulation of Biotechnology lays out the general US policy on regulation of GE products
and highlights that regulation should be science-based and product-based, rather than treating products created through a particular
process (i.e., genetic engineering) as inherently different. Consistent with this approach, the Coordinated Framework outlines
how the USDA, FDA, and EPA are expected to regulate GE products using existing laws. However, to implement the Coordinated
Framework and apply existing laws to GE products, the agencies established separate regulatory staffs, and USDA and EPA also
found it necessary to write regulations that are in essence, if not in letter, specific to GE products. Thus a process-based system has
been established, even though there is no scientific justification for regulating GE plants differently than conventional plants
(National Research Council 2002). For example, plants that are genetically engineered to be herbicide tolerant fall under the
scrutiny of the USDA; whereas, conventionally bred herbicide-tolerant varieties do not, even if the conventionally bred lines are
resistant to the same herbicide as the GE ones.
Under the different laws and implementing regulations for GE products, the agencies have varying abilities to assert authority
over GE products, require particular data, and use particular information in decision-making. The Animal Plant Health Inspection
* The authors are solely responsible for the content of this piece and any views expressed do not necessarily represent those of the EPA,
the USDA, or of AAAS.