Thank you very much.
My name is Rene Van Acker. I'm a professor in the Department of Plant Agriculture and associate dean of the Ontario Agricultural College at the University of Guelph, Canada. I thank the committee for the opportunity. I was also previously a professor of weed science and crop management, from 1996 to 2006, at the University of Manitoba in Winnipeg. My research areas include weed seedling biology and ecology, robust cropping systems, coexistence of genetically modified and non-GM crops, and trait movement from crop to crop.
My trait movement work has led to international collaborations, presentations, and consulting work in Denmark, Germany, Austria, Australia, Switzerland, and the United States, including membership on the scientific advisory committee for the international conferences on the coexistence of GM and non-GM crops in the agricultural supply chain, which has hosted conferences in Denmark, France, Spain, and Australia.
I grew up on a farm in southwestern Ontario. I hold B.Sc. and M.Sc. degrees in crop science and weed management from the University of Guelph and a Ph.D. in crop-weed ecology from the University of Reading in the U.K.
I thank you for the opportunity to present. My presentation is intended to draw attention to the challenges that may exist in trying to ensure that one type of crop does not contaminate another type of crop, and in particular how challenging this is in the context of preventing novel traits from appearing in crops in which they are not intended or wanted, especially when the threshold of presence that can cause harm is very low. If there is a regulatory consideration of potential harm due to the unintended presence of a given trait, it has to be realistic in that regard.
Most risks related to the release of crops with novel traits are related to novel trait movement, both from crop to wild type, for weeds, and from crop to crop. This is especially true for the movement of traits within and among farming systems and agricultural supply chains.
The issue of containing novel traits and/or transgenes and making sure they do not show up where they are not intended or wanted is a key point in debates about the desirability of certain novel traits. Coexistence is typically discussed in the context of accepted threshold levels of adventitious presence, but it is important to recognize that traits that are regulated must be fully contained to prevent escape and that the threshold for the presence of regulated traits is zero. This is the policy in Canada, as it is in the United States, Australia, Japan, Korea, and all EU countries currently.
In North America we have well over a decade of experience of commercial production of GM crops that contain distinct and easily traceable novel traits, and this experience provides us a wealth of examples and evidence that bear on the consideration of trait containment.
In a review I co-authored in 2005, I provided information to support and emphasize two important points in this regard. The first is, when crops of novel traits are grown commercially outside for any length of time, the movement of those traits beyond their intended destination is virtually inevitable. The risk of escape increases with scale of production and of associated equipment and as the number of participants in the production and handling increases. The second is, once a given trait has escaped into the environment, which includes the agricultural supply chain, retraction is difficult if not impossible, and as such, in situations where the escape is a problem, the problem becomes persistent and likely permanent.
These points support the need for great caution and care in the production and testing of novel traits that require containment or that can cause harm, if they appear where they are not wanted or expected. The challenges in managing trait containment are many, and they include the fact that the traits are often invisible and their monitoring requires effective detection methods.
Traits can move via either pollen or seed. That movement occurs within a complex of subpopulations across the landscape, which include crop, volunteer, and feral subpopulations. Trait movement can occur via equipment or via human handling during planting, harvesting, seed cleaning, seed handling, and seed storage. Each piece of equipment and each human participant can act as a sink or a source for traits, often as seed. In this respect, each piece of equipment or human operator can be considered an additional subpopulation for a given trait or latent populations of seed.
Traits can move among these subpopulations, which taken together act as a meta-population or an overall population with respect to a given trait. In this context, responsible containment efforts must take into account all possible subpopulations and possible pollen and/or seed movement opportunities between them. In particular, it's highly dependent on detection and eradication at reception points, in the receiving crop. This is a critical consideration, because the trait reception points may occur in fields, farms, equipment, and business operations of people who are not involved and perhaps not even aware of the containment effort. So that's a difficulty.
The required stringency of a given trait containment system depends on the threshold level and the facility of trait escape and movement. The latter depends on the nature of the crop species and the complexity of the crop production and handling system. To be effective, these plans need to extend beyond individual fields or farms, and the plans must reflect a healthy respect for the challenges of containment.
Since commercial production in 1996, we've had long experience with glyphosate-tolerance canola, for example, in western Canada, and it shows that volunteer canola can exist as a meta-population with respect to the Roundup Ready trait. This is after unconfined release. We have published work recently that shows the accumulation of novel traits in roadside canola populations. For the Roundup Ready trait in canola, trait containment would have required—although it wasn't required—a plan that encompassed the entire region. Management for containment within a given field and for a given crop alone would have been insufficient and unrealistic.
Given the number of mechanisms leading to trait escape and the fact that escapes can self-replicate and self-disseminate and persist, those who hope to prevent it must employ all methods available. A redundancy of methods is fundamental, because even low levels of trait escape into a seed lot can easily result in significant levels of trait presence in the harvested product, even for species that are primarily self-pollinated and have very limited seed persistence, such as spring wheat in Canada.
Physical isolation is one traditional means for limiting pollen-mediated gene flow; however, it does not assure protection from trait invasion, and those working to contain traits must take into account that traditional isolation distances were established to assure seed purity, not necessarily absolute absence of a given trait or genetic purity. Isolation distances must be suited to the nature of the species and the tolerance threshold.
Another containment method is temporal isolation, often used by plant breeders and seed growers; however, traditional seed purity assurance systems were not designed to deliver the type of seed purity levels required for containing explicit trait movement. We reported in a peer-reviewed study released in 2004 that certified seed lots of canola tested from western Canada had unintended GM traits in 97% of the seed lots, in some cases at levels as high as 4.9%.
A famous example of failure of trait containment is StarLink in the U.S., where corn engineered to express insecticidal protein was approved for animal but not human consumption. There was insufficient segregation oversight between food and feed streams in the U.S. bulk commodity handling system, and the insecticidal protein was found in a number of processed foods in 2000. Three years after this discovery, and after the execution of a massive recall effort, the USDA was still finding traces of StarLink within both food and feed handling streams in the U.S.
The StarLink case showed not only that insufficiencies in containment protocols resulted in problematic trait escape, but that full retraction of traits and their products from complex and massive commercial food and feed systems is extraordinarily challenging and maybe impossible.
A more recent example is the LibertyLink rice case in the U.S., in which regulated GM rice events escaped contained field trials and were eventually found in many elements of the U.S. commercial rice supply chain, including certified seed, mills, and final consumer products in key U.S. rice export markets, including several European countries. The economic impact on U.S. rice farmers has been estimated to be in excess of $1 billion. The final cost to farmers will not be known until the nearly 3,000 cases filed against the GM rice developer have been settled.
These and other cases highlight the potential impact of trait escape and the pervasiveness of escape when it becomes part of a large supply chain.
In summary, trait development in crops is in a new era, an era that includes any and all possible traits, including traits that can have true potential human health or environmental risks, traits that can affect farming system costs, and traits that are being considered and deregulated at varying rates around the world, leading to asynchronous deregulation and balkanized farm commodity export markets. In this era, economic harm could occur when traits appear where they are not expected and/or wanted. In addition, trait movement from crop to crop across diverse agricultural landscapes and within large integrated agricultural supply chains is very complex and challenging, and if there is an escape, trait recall is difficult and could be impossible in some cases.
It is important, therefore, that if there is a regulatory consideration of potential economic harm, it be realistic with respect to realities of trade movement and trade containment.