Want to study relative risks of GE?

Last Thursday, the United States Department of Agriculture (USDA) and the National Institute of Food and Agriculture (NIFA) posted a new grant – one that readers of Biofortified might be interested to read about. Called the Biotechnology Risk Assessment Grants Program, (PDF) this grant for up to $1 million for each project is for scientists who want to study the environmental risks of genetic engineering in agriculture.
What kinds of environmental risks? Things like basic genetics research, comparing breeding to biotechnology, and downstream effects of environmental release. There is even a section for it you want to submit a research proposal to study co-existence between GE and non-GE crops. You could even study pyramided, or “stacked” GE crops and compare them to single-transgene varieties. So many possibilities.
This call for grant applications is important for several reasons. Research coming out of this program would build upon, and compliment existing research, much of which is listed on this page on our blog. When published, it would also go on this list. (Out of curiosity I called up NIFA to inquire if a project like our GENERA project could be funded under a grant such as this. Sadly, no – the focus is on experiments and not catalogs of them, unless we were to take other people’s raw data and reanalyze under a new algorithm.) Sanely investigating and evaluating the kinds of environmental risks involved in genetic engineering is of utmost importance, and the research must put it in the context of plant breeding and agriculture in general. It is also my fond hope that for whatever projects are funded by this grant program, that the investigators keep in mind how their research would be conveyed to the public.
This is the first time I’ve promoted a grant for scientists to apply for on the blog, and it may not be the last. I thought that we could get some discussion going about what kinds of research we think would be worthy of pursuing? I have two main ideas.
The first is that we have had and heard a lot of discussion about using genetic engineering techniques to move and modify genes between plants of the same species, known as Cisgenics, and more recently, Intragenics. While it seems to be the case that consumers find this to be more appealing than cross-species genetic engineering, I find from discussions with scientists that there is considerable debate about the ups and downs of this distinction. Some see it as little different and merely a way to make an end run around regulations, others see it as a potentially less-disruptive way to alter the genomes of plants. This blog post by Kevin Folta was picked up for one of my department’s journal club discussions, and there were some interesting comments about it, including, how long must a gene be in a species for it to be worthy of being picked up and moved and have you still call it Cisgenics? Can you take Bt out of corn and stick it into another corn?
Correct me if I am wrong, but it seems to me that we have limited data to really answer the question about whether or not moving genes within or between species are inherently different in their risks, versus the same. Some argue that taking a gene from another species is riskier because if opens up new possibilities for interactions between genes because the proteins didn’t co-evolve. On the other hand, some take the co-evolution argument to the other side to point out that interactions also co-evolve and that two genes that have worked together in the same species may be more likely to interact, and there is still the null hypothesis that there is no significant difference at all.
There is some evidence and theory behind each position, but what we would need, then, is an experiment designed specifically to address the question of whether there is a difference, and what kind of difference. I suggest that one way you could go about this is with an experiment that goes something like this. There are genes in many species which serve similar functions due to common ancestry, which are called homologous genes. But, they may have evolved slightly differently in each species, so you could test whether the closeness of the species source matters by simply generating many GE plants with each of these homologous genes, and then comparing the result. You could look at gene expression to see if there are any significant differences between them, for example.
Scientists in the audience might point out that the place in the genome where you engineer the new genes will matter, so you would probably have to generate several transformations in different sites and compare the distributions of effects. You could also look at phenotypes to see if anything odd comes out consistently with one and not the other, or investigate what proteins might interact with the different transgenes. Do this for several sets of homologous genes from progressively more distant organisms and you’ve got yourself a way to test this hypothesis! Add another plant species to insert them into and it will broaden applicability.

Alfalfa by TwoWings via Wikimedia Commons.

I would also like to see some research on coexistence come out of this. What would be some good management practices that will minimize gene flow and spillover effects between neighboring farms? How much time must there be between flowering, or what kinds of borders are necessary, and what would crops that won’t cross with other varieties due to cross-incompatibility genes mean for these practices? Moreover, there is some opportunity for an interdisciplinary research in this area.
As I have shown before, we don’t know a lot about exactly what are the thresholds for consumers when it comes to cross-pollination between GE crops. The Consumers Union did a biased poll with loaded terms, and yet, couldn’t get consumers to care very much about it. Organic groups and exporters are worried about consumer rejection (or rather, processor rejection) if there is cross-pollination, and it seems to me that coming up with thresholds for cross-pollination in a coexistence regime necessitates knowing what consumers think, in a robust and scientific manner. If consumers like the 1% rule, are fine with a 5% rule, or wouldn’t touch a 0.0001% transgenic-pollinated organic crop, those would lead to different situations entirely. Imagine a project where both the methods to achieve coexistence are studied alongside consumer attitudes toward the results of those methods, and you’ve got a nice interdisciplinary project.
What would you like to see get worked on? Any thoughts about what I described above? There’s $4 million of research to be funded – wouldn’t it be grand if an idea started here and made it into a selected proposal?

10 comments

  1. This is pretty good money for seeking risks that aren’t there. Sure, the results might be wielded against the claims of environmentalists etc., but they don’t pay attention to such things.
    This money would be better spent on public-sector research to develop crops with novel traits which will compete against the products of the multinationals.
    I have to wonder if this is a ‘stimulus package’ to fund ‘shovel-ready projects’.

  2. An ideal opportunity for the anti-GM lobby to design some nice experimental studies backed by predetermined and validated statistical methods to test their hypotheses and to ‘once and for all’ show whether their GM concerns are valid.
    I won’t hold my breath.
    Jonathan

  3. There are risks in all things – so there are risks with GE as there are risks with driving a car. The key is to generate useful data to hone the laws and regulations so that it is appropriate for the level of risk.

  4. Possible naïve comment here – I would like to see a multi-generational animal study done to finally prove/disprove/address the concern that “who knows what will happen 75 years down the road.” I have no idea how hard these would be to do, or if there are so many uncontrolled variables that the results would be inconclusive. One of the comments I hear is that “the Industry” limits access to tests because of confidentiality. A multigenerational test may be a way of addressing this. Also – I don’t know if these tests already exist in the Genera database here, but for whatever reason those tests are not recognized by the “anti-GM lobby.”
    Side note and personal observation – we have been feeding chickens/pigs/cattle feed using GMOs for ~20 years now. Does anyone have an idea of how many generations that would be? I don’t see any of the chicken/pig/cattle producers or processors up in arms over harm to their products from GMO feed. Is this not an indication that it is safe over multiple generations?

  5. There are already several multi-generational studies in our list, so far it has been difficult to get people to recognize that they are there. One was I believe ten generations with quails. They all differ in duration, variables measured, etc.
    It could be useful to design a mutually-agreeable feeding study, involving some prominent anti-GE individuals in the process, with an open format for data collection, analysis, publishing, etc. I have thought about that. However, a lot of this hinges on what will happen on either side if the data does not go their way. A similar exercise for climate science was the Berkeley Earth Surface Temperature project, which recently made headlines because it started out as a skeptical venture, and ended up confirming the climate scientists’ work. http://www.berkeleyearth.org/ However, Anthony Watts, a ‘skeptic’ involved in the project, is now crying foul at the results and denying their significance. The question is, if Greenpeace was involved with such a study, would they agree with its conclusions if it found that GE crops were safe?
    Unfortunately, though, such a thing could not get funded by this grant program, as it involves a feeding study and not research on environmental impact.

  6. Karl,
    What Watts and other skeptics are saying is that the BEST data proves what skeptics have been saying for a long time. That there has not been significant warming in the last 12 years and that it remains unknown what role humans play in ‘climate change’.

  7. Richard,
    A big drawback to long-term trials is that influences of factors which cannot be controlled for ‘pile up’ over the course of the study and render the results increasingly less reliable. The best data come from tightly-controlled, short-term studies based on a well-considered hypothesis. Making the study long and taking a ‘we might find something’ approach is not a good idea. You may well find ‘statistically significant’ values but no clear notion of what caused them.

  8. I’m not overly conversant with the data, but isn’t the “not significant over 12 years” claim rather similar to what was done here regarding crop yield increases over time – highly variable metrics can fail to be significant over a short timespan (a decade say) and yet be significant over longer timespans.
    I seem to recall the soy yield data did precisely this, and conversation around the issue was kiboshed when pdiff (I think!) pointed out that breaking a time series into smaller bits and claiming any lack of statistical significance in a single bit was either (depending on your background) misunderstanding, or wilfully misusing the statistics.

  9. That is what I was thinking as well. As I was writing my original reply, I was recalling the study investigating the effects of introducing gmo corn into the diet of farmed salmon. The study said, if I recall, that there were no harmful effects in any of the diets used other than a variation in organ size. Yet the point of the individual introducing the study was that the study was evidence of gmo feed causing harm. This wasn’t even a multiple generation study and it was used (misused) to show harmful effects where there were none.
    I still go back to the anecdotal and non-scientific evidence that after feeding livestock gmo products for ~20 years there is no evidence of any harmful effects.

  10. Richard,
    I would hardly call epidemiological surveillance results, however informal, “non-scientific.” The sheer size of the base makes it likely to capture small effects (I hesitate to use the correct term “number of subjects” because that implies “experiment,” and we know that people equate that with Dr. Frankenstein et al, and react negatively).

Comments are closed.