[permaculture] Gene Ecology Guide to: Is confidence in the monitoring of GE foods justified? Was: Re: [SANET-MG] Honeybee Colony Collapse Disorder may be caused by GMO Bt genes
Subject: [permaculture] Gene Ecology Guide to: Is confidence in the monitoring of GE foods justified? Was: Re: [SANET-MG] Honeybee Colony Collapse Disorder may be caused by GMO Bt genes
Gene Ecology Guide to: Is confidence in the monitoring of GE foods justified?
New Zealand Institute of Gene Ecology, University of Canterbury,
Christchurch, 8020, New Zealand
A condensed version of a paper published in Trends in Biotechnology in July
2004.
This paper summarises a critical analysis of the effectiveness of monitoring for evidence of genetically
modified/engineered organisms (aka GMO/GEOs or LMOs) in the human food supply. The focus of the paper is detection of
GMOs, because national and international laws require that they be monitored and quantified. The investigations of GMO
corn in the New Zealand environment since 2000 are analyzed to find the strengths and weaknesses of monitoring food for
GMOs or other contaminants.
The authors conclude that the effectiveness of monitoring is less than needed to prevent or heal a harm that might come
from the contamination of the food supply by a GMO. Monitoring must be improved to meet standards that are set by law or
contractual obligation, or which may have influenced any considerations of risk during an approval process. Monitoring
could be significantly improved by elevating the public’s right to effective monitoring above the developer’s interest
in restricting information about DNA sequences that are considered proprietary information. Enough DNA sequence
information should be made public for any commercial GMO to be identified and distinguished even from novel hybrids.
Moreover, public regulatory agencies must claim and retain ownership of monitoring data, and the data from monitoring
should be made available in a form that can be interrogated by the independent scientific community.
Trends in Biotechnology is rated among the top 4 international biotechnology journals. Trends in Biotechnology publishes
articles on all aspects of biotechnology only after independent and stringent peer-review. It is published monthly by
Elsevier Science London, by an international team of editors.
The original paper was invited by the Editor and is based on a report prepared by one of the authors at the request of
the Local Government and Environment Select Committee on Corngate.
Introduction
Food security is in the long term interests of all countries. Food security is the sustainable capacity to produce safe
food and to monitor trade of organisms that are a direct threat to the food chain or health of citizens. These threats
must be detected while at levels below which they can cause harm or spread by reproduction or infectious transfer.
Page 1
Why monitor GMOs?
Monitoring is required by different national and international laws for organisms that are deemed to be threats to human
health or the environment, including GMOs. Some GMOs are tolerated in the food supply of many countries, but only up to
particular levels above which the food must be labeled as containing GMOs. Monitoring is thus both for detection—the GMO
is or is not present—and a way to measure the quantity—for example, there is 0.05% GMO corn in a batch of corn flour.
The monitoring challenge is not unique to the problem of GMOs. In this regard, attempts to enforce regulations on GMOs
are teaching the scientific and regulatory communities both how effective and vulnerable the food chain is to the
movement of any unwanted, illegal or harmful contaminant.
What is needed for proper identification?
Information is the key limiting ingredient for conducting the best possible monitoring. Some limitations can be
overcome, for example by requiring GMO developers or governments to:
• disclose all DNA sequence information on existing commercial GMOs—enough information to distinguish between known GMOs
and any hybrids;
• standardize best practice for detecting, or measuring the amount of, GMOs;
and
• fully disclose the data of monitoring tests for inspection by the full scientific community (effectively the same
standard as applied by respected journals for publication).
Four error stages
This review considered the challenges to monitoring GMO crop plants, focusing on events in New Zealand from 2000. These
separate stories each conform to a basic skeleton of the monitoring process (FIGURE 1).
(1) Seeds delivered or destined for export in large grain consignments are sampled for GMOs. The number of seeds in the
sample must be large enough to guarantee at a high level of certainty (e.g. 95%) that it will include a GMO seed at the
desired or practical detection limit (FIGURE 2). So for confidence that the consignment were 99%, or 99.9% or 99.99%
GMO-free, corresponding to 1 seed in 100, 1,000 or 10,000, respectively, the number of seeds must be large enough to
ensure that one GMO seed would be in the sample if the consignment had as much as 1%, 0.1% or 0.01% GMOs.
(2) Seeds or other sampled material must then be ground and mixed before DNA is extracted for subsequent testing. The
mixing stage ensures that further subsampling will not exclude any GMO genome present from the test.
(3) The most common method for routine monitoring and quantification of GMOs is the polymerase chain reaction (PCR). For
optimum reaction conditions, a test can usually have no more than 100-200 nanograms of DNA. (A nanogram is 1 billionth
of a gram. For comparison, a single seed of wheat is about a hundreth of a gram and a single kernel of corn is about 3
grams).
Page 2
FIGURE 2
For consignments with an average of 1% GMOs (top), a sample of 300 seeds(60 circles) is the minimum size necessary to be
95% sure that one will includea GMO (solid circle). This is because smaller samples, even up to 100 seeds in size
(shaded boxes) would have a variable number of GMO seeds and, by chance, hold zero in any given sample. In practice,
even larger sample sizes or different sampling strategies are needed because GMO seeds are not randomly distributed in
consignments (bottom), returning a false reassurance that the consignment was GMO free.
This limit in the weight of DNA in a reaction restricts the number of genomes in the reaction. For corn, each test tube
will contain about 18,000 genomes. So if the proportion of genomes from a GMO were 1%, the test tube would have 180 GMO
genomes; 0.1% and the test tube would have 18 GMO genomes; 0.01% and the test tube would have 2 GMO genomes. For crops
with smaller genomes more sensitive limits are possible because a greater number of genomes will be in any given amount
of DNA.
(4) The final source of error comes from:
• mis-interpretation of monitoring data;
• failure to confirm the effectiveness of the PCR; and
• failure to retain rights to the data which, if viewed by the scientific community, might reveal evidence that further
testing is necessary.
Reliable PCR requires having complete descriptions and regular confirmation of key indicator DNA sequences. Without such
information, an apparently GMO-free consignment could have illegally high levels of GMOs. In New Zealand, the primary
data is often not owned by the regulatory authority and is not released because it is considered proprietary.
Sampling
THE ERROR SOURCE: samples too small.
An important source of error comes from assumptions made in taking the initial seed sample. Seed testing certificates
were issued for samples that were too small for the detection limits claimed (i.e. less than 0.1% and 0.05%,
respectively) in the corn consignments investigated in 2000 and 2003, and sampling was based on unverified assumptions
of random seed distribution throughout the consignment (FIGURE 2).
Samples for the still controversial case of GMO corn found in New Zealand in the year 2000, were often of 1,000 seeds
with claims of detection down to 0.1% GMOs. To test to 0.1% GMOs, however, requires a minimum of 3,000 seeds.
Samples from the outbreak of GMO corn in 2003 had as many as 10,000 seeds. These samples roughly correspond to a
detection limit of 0.03% (3 GMO in 10,000 seeds), but New Zealand regulators claimed detection to 0.01% contamination.
The seed samples used in 2000 and 2003 provided about the same accuracy as flipping a coin, or playing “paper, scissors,
rock”, at the detection limits claimed. That is, the probability of incorrectly thinking that the consignment was
GMO-free would have been nearly 50% and 37%, respectively.
The above sample sizes ignore the additional uncertainty caused by errors at the other stages in the process
(sub-sampling and reaction sensitivity discussed below). Therefore, the sample sizes discussed above are almost
certainly still far too small. There is yet no data on how big the sample size should be to overcome error contributed
by uncertainties in seed mixing, sub-sampling and reaction sensitivity measurements.
Page 3
Sub-sampling
THE ERROR SOURCE: the proportion of GMO genomes is the same as the proportion
of GMOs in a sample.
Often it has been assumed that the number of GMO genomes in a test tube was an accurate measure of the number of GMOs in
the larger consignment. For example, if a test tube had zero or one GMO per 1000 genomes, then the number of GMOs in the
consignment was less than or equal to 0.1%.
This assumption is based on:
• the average, or known, sizes of the genomes of the crop from which they are
made; and
• that all crops of the same species yield the same amount of DNA when it is
extracted prior to conducting a PCR.
Both the genome size (in number of nucleotides of DNA) and the amount of DNA that can be extracted from any particular
GMO, determines the number of genomes in each test tube. Even small deviations from the average can have large effects
on the calculated proportion of GMOs that could be detected.
Contrary to expectation, a GMO and its parent can differ in genome size by up to 25%. Moreover, the amount of DNA
extracted from a GMO and its parent can differ by up to 20%.
The distribution of DNA in corn seeds is different in certain important respects in each of the three different tissues
that make up the seed. About 97% of the DNA is held in the embryo and endosperm tissue, and the remainder in the
teguments can be neglected for this discussion. The embryo, which develops into the adult plant, has one complete genome
from each of the seed’s male and female parents. The endosperm tissue has two copies of the female to every copy of the
male parent’s genome. Depending on the plant, the embryonic DNA can vary from as little as 40% of the total to nearly
60%. This means that the endosperm can also contribute as much as 60% and as little as 40%.
This variance in amount of DNA from the different tissues changes the number of genomes from the seed’s different
parents. If a GMO plant were the product of a cross between a male GMO (pollen) and an unmodified female, then the
number of GMO genomes in the sample could under-represent the number of GMOs.
All-in-all, the combination of these errors, neglecting sampling (above) and reaction sensitivity (below) errors, could
overestimate the capacity to detect a GMO by nearly a factor of 3 (FIGURE 3).
Reaction sensitivity
THE ERROR SOURCE: PCR is routinely effective at one or a few genomes.
The most important variable for measuring the efficiency of a PCR reaction is the match between the presumed DNA
sequence of the GMO and the actual sequence. This variable is the most difficult to control because uncertainty arising
from it can only be resolved if the developer provides complete and up-to-date sequences, and if the environmental
sample contains a known GMO.
If the actual DNA sequence varies by as little as a single nucleotide, the reaction efficiency could be reduced by a
factor of 10 or more. The cumulative effect of such errors could result in a false confidence that a consignment was
GMO-free despite quite high proportions of GMOs.
Page 4
Independent scientists have found GMOs that have not been approved for human consumption in common processed foods, but
only after they painstakingly determined the DNA sequence of the inserted genes in the GMOs by intensive trial and error
approaches.
Researchers cannot rely on the registered name of the GMO to accurately inform them of the DNA sequence in their
samples. Commercially produced GMOs of identical name or “event” can have inserted DNA with different sequences.
Interpretation error
THE ERROR SOURCE: Data generated by monitoring techniques are self-evident
and definitive.
Each of the above error-generating steps is magnified by misunderstanding,
and misinterpretation of, the data.
The consequences of interpretation error are not easily quantified, but must
be acknowledged.
Interpretation error may also be minimized by complete disclosure of the defining DNA sequences of GMOs and provision of
genotyping profiles that would distinguish between the commercially approved variety and any hybrids that might arise
post-release.
Making proper and complete identification of GMOs in environmental samples is especially important in a time when GMOs
may be used as bioweapons or for industrial sabotage.
Concluding remarks
An increasing body of scientific evidence supports the suggestion that the capacity to detect and monitor GMOs is below
what is sometimes claimed. It is also falls short of what would be necessary for the purposes of containing GMOs in some
environments or eliminating them from others.
The cumulative effect of errors at the four key stages of monitoring could mean the difference between detecting and not
detecting a GMO. Such errors are large enough in some cases to create a false confidence that the amount of
contamination is below a legal threshold, set, for example, by food labeling laws.
FIGURE 3
Underestimating the proportion of GMOs from mixed samples of unmodified corn (black double helix) and GMO corn (grey
insert double helix) seed DNA. (top) Hybrids derived from a GMO father all are genetically engineered but contribute
only one third as many copies of the inserted DNAinto the mix, falsely devaluing the number of GMOs. (bottom) Some GMOs
yield less DNA. Thus, in any sample of the same totalweight of DNA, the number of genomes from GMOs appears to be less.
Page 5
At this time, changes in policy and procedure would improve the performance of monitoring more than changes in the
science. Consumers and governments could have more confidence in monitoring if:
• there was complete public access to all DNA sequences in approved GMOs and
those in development or field trials;
• developers submitted a genetic profile of all commercial GMOs so that the
identity of contaminants could be made certain;
• the public owned the data generated during testing/monitoring, so that the data could be interrogated by both the
public and independent scientists;
• there was re-confirmation of DNA sequences of released GMOs over time, so that monitoring tools could be modified to
account for any changes;
• regulatory authorities insisted upon—and public and private funding bodies made possible—monitoring capacity that
exceeded or matched the ability of GMOs to be anywhere they were not permitted.
These lessons should not be dismissed simply because they were learned with commercial GMOs. All societies will always
have an interest in monitoring their food supplies for GMOs and other potential contaminants whether or not some GMOs
are legal and welcomed.
Glossary
DNA, nucleotides
A DNA molecule is composed of two polymers, or ‘strands’, consisting of a linear sequence of nucleotides called A, T, G
and C (abbreviations of their chemical names). DNA is normally double-stranded, with the strands bound together by bonds
between complementary nucleotides. The weak bonds between nucleotides on different strands are formed optimally when A
pairs with T and G pairs with C.
embryo, endosperm, teguments
Tissues of seeds. The adult plant is derived from the embryo.
tissue
proproption of seed DNA
parental genomes (father:mother)
embryo
40-64%
1:1
endosperm
36-60%
1:2
teguments
negligble
varies
hybrid
The offspring of crosses between varieties within the same or closely related species. When one or both parents is a
GMO, then the hybrid is a GMO regardless of the number of genomes that are from the modified parent.
polymerase chain reaction, PCR
These principles are the basis of the polymerase chain reaction. In the PCR, discrete and short sequences of DNA from an
organism are selectively amplified by a runaway, or chain, reaction concentrated on those particular sequences. The
reaction is both initiated and bounded by even shorter sequences of DNA (usually <20 nucleotides) called primers.
Primers are synthesized to be complementary to each end of a sequence of DNA to be amplified. The match between primer
and ‘template’ (organism DNA) is a critical determinant of the efficiency of the PCR reaction. PCR efficiency is the
minimum concentration of a target template in a mixed sample that can be reliably amplified.
Page 6
About the authors of the original paper:
Assoc. Prof. Jack Heinemann has a BSc with honours in Biochemistry and Molecular Biology from the University of
Wisconsin-Madison, USA, and a PhD in Molecular Biology from the University of Oregon, USA. He joined the University of
Canterbury in 1994, following five years as a fellow at the US National Institutes of Health. He is the founder and
current director of the New Zealand Institute of Gene Ecology, an adjunct professor at the Norwegian Institute of Gene
Ecology, recipient of the New Zealand Association of Scientists Research Medal (2002) and he serves on the UNEP
Biosafety Capacity Building Package Executive.
Dr. Ashley Sparrow has a BSc with honours and a PhD in Ecology from the University of Adelaide, Australia. He joined the
University of Canterbury in 1993 and was elected Head of the Department of Plant and Microbial Sciences in 2000 and then
was appointed Head of School of Biological Sciences in 2003.
Professor Terje Traavik has a Doctorate of Veterinary Medicine and a PhD in Virology from the University of Tromsø,
Norway. He joined the University of Tromsø Department of Microbiology in 1977 following seven years at the National
Institute of Public Health, Oslo. He initiated the program in virology at Tromsø, and was founding Head of the
Department of Virology from 1983-2004. He is currently the founder and scientific director of the Norwegian Institute of
Gene Ecology, recipient of the Erna and Olav Aakre Foundation Prize for Excellent Cancer Research (1992), and he serves
on the UNEP Biosafety Capacity Building Package Executive.
Published in June 2004 by the New Zealand Institute of Gene Ecology, University of Canterbury, Private Bag 4800,
Christchurch, 8020, New Zealand (www.nzige.canterbury.ac.nz).
Citation: Heinemann, J.A. 2004. Gene Ecology Guide to: Is confidence in the
monitoring of GE foods justified?
Original paper
Heinemann, J.A., Sparrow, A.D. and Traavik, T. 2004. Is confidence in the monitoring of GE foods justified? Trends
Biotechnol. v.22(7) p.331-336.
This publication of the New Zealand Institute of Gene Ecology may be copied provided that the source is acknowledged.
Copies of the original article are available by subscription to Trends in Biotechnology or by request to the authors.
Page 7
[permaculture] Gene Ecology Guide to: Is confidence in the monitoring of GE foods justified? Was: Re: [SANET-MG] Honeybee Colony Collapse Disorder may be caused by GMO Bt genes,
Lawrence F. London, Jr., 04/05/2007