The research programme is an important move in the right direction, but more work is needed before it can detect unknown GMOs in food and feed. Report: Claire Robinson
In October a US-based consortium of researchers from industry and the US government’s intelligence community announced that they had developed new technologies that can detect when an organism has been genetically engineered.
The researchers are focusing primarily on “biothreat” detection of genetically engineered viruses and bacteria that could cause new pandemics. But the new technologies could flag up the use of genetic engineering in a wide range of organisms, including new gene-edited GM plants and animals in the food and feed supply. In time and with further work, the technologies could identify unknown, unauthorised and “hidden” GMOs, as well as the relatively easily identifiable known and authorised GMOs.
The development has been met with a flurry of excitement at a time when governments and regulatory agencies across the world are moving to exempt “new GM” plants from regulatory oversight partly based on the claim that they cannot be distinguished in laboratory tests from plants produced by conventional breeding and random mutagenesis breeding. If FELIX develops its work with the collaboration of the GMO industry and regulators, the days when that claim can reasonably be maintained might be numbered.
The FELIX programme
FELIX stands for Finding Engineering-Linked Indicators. The research programme is run by the Intelligence Advanced Research Projects Activity (IARPA), the research and development arm of the US Intelligence community. To develop the programme, IARPA partnered with the biotech company Ginkgo Bioworks and the engineering nonprofit Draper. Ginkgo has developed new computational tools and Draper has developed a new experimental platform to help detect and identify when samples include genetically engineered organisms.
At their press conference on 17 October, available on YouTube, the FELIX researchers emphasised that all GMOs carry scars and signatures of genetic engineering processes that could enable detection and identification. But that doesn’t mean that all GMOs can be detected right now, for reasons explained in this article.
In GMWatch’s assessment, the researchers have done important work that will contribute to enabling the detection and identification of all GMOs, including unknown and new gene-edited ones. This work also complements that of other scientists who are working towards the same goal. However, our analysis – compiled with the help of scientists working in the field – suggests that much political will, focused investment, and collaborative work on the part of detection laboratories, regulators and GMO developers will be required before these and similar technologies can be used by public GMO enforcement laboratories for the detection and identification of unknown and unauthorised new GMOs in the food and feed supply.
Known and unknown GMOs
The field of GMO detection and identification is divided into two main areas:
1. Detection of unknown and undeclared GMOs, which in the EU is done by the public GMO detection laboratories, ENGL.
2. The routine cheap detection of known and authorised GMOs (potentially including gene-edited ones), which is done by private and enforcement laboratories. Such detection processes can use just one scar or signature out of the total that have been identified.
The first area – unknown GMOs – is the main focus of FELIX. But the genetic sequences that the programme gathers in its database could, in principle, be used for detecting both unknown and known GMOs.
How it works
FELIX’s method is similar to existing screening methods for older-style transgenic GMOs, where scientists look for certain genetic sequences known to be a feature of these altered organisms.
The FELIX system involves:
* creating a database of DNA sequences known to have resulted from genetic manipulation, as well as a database of non-GMO and natural organisms
* comparing the GMO genetic sequences against the sequences of known natural and non-GMO “reference” varieties, in order to isolate the genetic “fingerprints”, “scars” or “signatures” of genetic modifications; and
* analysing samples in order to identify any potential GMO and assign an origin.
It’s clear from the press conference and accompanying press release that the researchers’ priority is detecting dangerous viruses and bacteria that could cause new pandemics (“biothreat detection”). But as they point out, the technologies work on any type of organism, including yeast, fungi, plants, and animals. Ginkgo’s press release confirms that the potential applications include “environmental monitoring and food inspection”.
So in principle, they could be used for the detection of genetic engineering in undeclared gene-edited GM crops and foods. Currently advocates of deregulation claim that these cannot be identified as GM because the intended changes made by gene editing are not distinguishable from those that can occur naturally.
However, more work is needed before the FELIX technologies can be relied upon for this purpose.
Accuracy: A work in progress
The researchers report that their tests correctly detected about 70% of every 100 genetically engineered samples, with no more than one false positive in each batch tested.
While impressive, a 70% success rate isn’t good enough on its own to provide definitive or even highly probable proof of a GMO’s presence, for instance, in a legal case about contamination. But in time, as the FELIX database expands and software improves, the success rate will improve.
Even in the current state of development, Dr Yves Bertheau, honorary INRA research director at the Muséum national d’Histoire naturelle in Paris, France, who is an expert in GMO detection and identification, believes that if FELIX analyses were added to other information, such as that gathered by GMO enforcement laboratories, the total body of evidence could constitute such proof.
Unknown GMOs: Still a problem
The technology developers aim to identify both known and unknown GMOs. But London-based molecular geneticist Dr Michael Antoniou explained that in order for either type of GMO to be detected, it would have to contain one or more sequence that is already present in the reference database and identified as a scar of a genetic or epigenetic modification. If a genetic modification-related sequence is not in the reference database, the FELIX tests will miss it and fail to detect the GMO.
The FELIX researchers stated in the press conference that their database contains 8,000 known genetically engineered sequences. But they conceded that when it comes to plant genomes, the database isn't yet comprehensive enough to encompass the large range of variation for natural non-GM plants.
This isn’t just a problem of the FELIX database – it’s common to all sequence databases. Attempts to collect GMO sequences are hampered by the failure of GMO developer companies to provide access to good quality reference material (in the case of GMO plants, this is normally crushed seeds or DNA) to the EU’s detection laboratories.
The FELIX researchers acknowledged that the incompleteness of their database will make their system less able to detect the use of gene editing in GMOs with a “small” intended change of the type that developers claim could happen naturally, such as a change to a single DNA base pair.
The problem of detecting such small intended changes could be solved, the researchers stated, by expanding their reference database to include more genetically engineered sequences.
Dr Bertheau commented that the FELIX database could benefit from the scientific literature and patent surveys carried out by ENGL on the intended sequence changes of those GM traits for which information is available. He added, “There are plenty of software pipelines to improve the detection of changes to single DNA base pairs.”
Dr Bertheau also said that many of the scars left by GM techniques, including “new” techniques like gene editing, are already known and can form useful markers for detection of unknown GMOs. He has published a book chapter giving references to papers detailing such markers.
Large-scale gene editing changes are more easily identified as GMO
Dr Antoniou said that some larger-scale changes brought about by gene editing can be obviously attributed to gene editing technology, such as gene insertion (SDN-3) applications of gene editing. But he added that even some applications of SDN-1 (gene disruption) gene editing, where no foreign genes are intentionally inserted, might carry an obvious signature of genetic modification. For example, in a GM gene-edited SDN-1 low-gluten wheat, several dozen genes and gene copies in the wheat genome were changed. These amount to massive changes that would not happen naturally and could be easily spotted using readily available analytical methods.
Furthermore, Dr Antoniou said, “Each gene-edited wheat plant will carry a different pattern of unintended insertions or deletions (indels) of DNA. Also, a wheat plant has six copies of each gene and you won’t get the same insertions and deletions with each gene – again, they will be different. Such patterns of indels are extremely unlikely to happen in nature and will be an obvious sign of genetic modification.”
It's vital to take into consideration the unintended as well as intended changes in order to obtain a complete and unique signature of each GMO.
Access to GMOs a problem
The main challenge to applying detection technologies like FELIX to gene-edited plants and livestock animals can only be overcome if the GMO industry undergoes a revolution in transparency or, more likely, if regulations force transparency. That challenge is access to the necessary reference materials – samples of the GMO and a non-GMO counterpart. In the case of plants, these are normally in the form of crushed seeds or DNA, as mentioned previously.
Thanks to a 2018 ruling by the European Court of Justice, in the EU, new gene-edited GMOs are officially GMOs and are subject to the requirements of the GMO regulation. This means that for the purposes of traceability, the GMO developer must provide the EU Reference Laboratory for Genetically Modified Food and Feed (EURL-GMFF) with access to the relevant reference materials, as well as information on the genetic change made in the GMO. The developer must also supply a detection method that is then assessed and validated by EURL-GMFF.
The irony is that the EU authorities need only allow this current situation to persist in order to ensure that all declared GMOs, new and old, can be detected, without years of expensive effort by the FELIX researchers, enforcement laboratories, or anyone else. But perversely, the EU Commission is trying to deregulate (remove regulatory controls from) new GMOs – which could scrap the requirement for the reference material and detection method to be provided to the EURL-GMFF.
The importance of the GMO developer supplying the reference material, sequence information and detection method is confirmed in an EU Commission DG SANTE report, which noted responses to a questionnaire sent to laboratories that develop and validate methods for detecting GMOs. Most laboratories said the biggest challenge in detecting organisms made with new GM techniques (which the report calls new genomic techniques or NGTs) will be “the lack of reference material and event-specific [specific to certain GM transformations] methods, as well as the lack of sequence information and of validated detection methods.”
If the Commission follows its current trajectory in deregulating new GMOs, we could all be left in the dark about their presence in the food and feed chain.
Origin of small changes in question
Another obstacle identified by the detection laboratories is that “there is currently no analytical approach that makes it possible to distinguish with absolute certainty whether a single nucleotide variation (SNV)” – a small change in the genome – “is the result of the use of NGTs, natural mutation or conventional plant breeding.”
The laboratories said that this would decrease the judicial strength in relation to the detection of NGT products, should a case be taken to court.”
This is where FELIX’s expertise could help, given further work.
The Commission’s report noted that in Germany several projects are working on developing methods for detecting and identifying new GM products, as well as databases to share information on these products.
Companies should come clean on unintended changes
In a truly science-based system, GMO developers would have to supply details of the unintended changes (the mutational signatures) in their gene-edited products to regulatory and enforcement authorities for detection, identification, and risk assessment purposes, as well as to prove that their GMO is man-made and therefore patentable. But current law doesn’t require that this information is provided to EURL-GMFF for detection and identification purposes.
For risk assessment purposes, applicants for GMO authorisation are supposed to tell the European Food Safety Authority (EFSA) about unintended changes. However, GMO developers use inadequate screening methods to look for them. They are not required to use in-depth methods such as long-read whole genome sequencing, which could spot them.
Therefore, as things stand, they will miss many unintended changes, which will not be assessed for their ability to cause harm to health or the environment.
Commission inaction undermines EU agriculture and producers
The EU Commission has justified its push for deregulation for gene-edited foods and crops by arguing that they cannot be identified as GMOs as distinguished from natural products. Critics respond that this is because the Commission has always refused to fund ENGL to develop detection and identification methods for new GMOs. Thanks to the Commission’s inaction and self-fulfilling prophecy, the EU has been left far behind in this area of innovation.
It is obvious that GMO developers are able to identify their patented varieties and tell them apart from natural varieties, or they wouldn’t be able to enforce their patents. Indeed, in the FELIX press conference, the researchers predicted that GMO developers will be interested in using the new technologies to protect their patents and detect the “theft” of their intellectual property.
Dr Bertheau said that the FELIX work “contradicts the assertions of the JRC (Joint Research Centre, the Commission's science and knowledge service), the Commission, and some scientists that it is impossible to detect and identify genome-edited products”.
Dr Bertheau believes that the Commission’s failure to act on detection methods constitutes a “lack of foresight”, resulting in the undesirable situation in which “European seed producers and farmers will be subject to the claims and decisions of foreign companies”. He said that the FELIX results, provided they can first be confirmed on real products of new GM techniques, show that non-EU countries “are not only giving themselves the means to detect unknown potentially dangerous GMOs, but that they are also giving companies producing patented varieties the means to defend their patents. European seed producers and farmers could be sued for patent infringement, for example, following contamination by patented GMO seeds.”
This, he said, “further reduces the competitive advantages of European non-GMO products, which are important for our exports of seeds and food commodities. This is why the application of the European GMO regulation to NBT [new breeding techniques, a euphemism for new GM techniques] products is fundamental, as it will enable the detection and traceability of NBT products by obliging companies to provide reference material and identification techniques that they can no longer deny being able to develop.”
If the Commission succeeds in its attempts to deregulate gene editing technologies, this could deny consumers and farmers knowledge of the genetically modified status of a crop or livestock animal. On the other hand, if the EU keeps them regulated under the current laws, all gene-edited GMOs that reach the market will be detectable and identifiable. That will be a nightmare for developers of the products, since it enables farmers and shoppers to avoid them. By pursuing deregulation, the Commission is kowtowing to the industry but making a rod for the EU’s own back and increasing public mistrust in the food and farming system. It will also increase mistrust in the GMO seed and livestock industry and its products.
The whole GMO
Dr Bertheau emphasises that effective detection and identification of new gene-edited GMOs must entail analysis of the whole DNA sequence – not only the intended edit, but also the unintended DNA damage, or characteristic “scars”, that the gene editing process creates. Taken together, the intended and unintended changes make up the unique “signature” of the GM gene-edited plant that enable identification. For known GMOs, only a part of the sum total of the scars and signatures can be used for screening and identification.
Claims that new gene-edited GMOs cannot be detected as GMOs are based on a perverse and scientifically unjustifiable determination to confine the analysis to only the intended edit. In some cases, this intended edit could be mistaken for something that occurred naturally.
Dr Bertheau says that once the whole spectrum of intended and unintended changes have been analysed, then, if the analytical results need to be validated, other non-laboratory-based evidence can be added in a so-called matrix approach. In this way, detection of gene editing in a plant – and identification of the particular GM plant – is feasible, for unknown and unauthorised GMOs. He also points to evidence that such analysis could lead to the identification of the laboratory that developed the plant and the particular GM techniques used to make the plant.
So if the FELIX researchers continue to develop their reference database of scars and signatures, and add analytical methods for the changes deliberately introduced to bring about the intended traits, they could succeed in being able to detect both known and unknown GM gene-edited plants – and distinguish them from naturally occurring and non-GM plants.
The “model” GMOs that FELIX tested may not represent real-world ones
An important limitation with the FELIX system, as pointed out by Dr Bertheau, is that the FELIX tests have hitherto only used “model” GMOs created by the research partners. These are not necessarily representative of the commercial GMO varieties on the market and in the commercialisation pipeline.
Dr Bertheau said this is a missed opportunity, since “Gingko Bioworks has been working since 2017 with Bayer and created a joint venture with Bayer on 18 October, the day after the press conference, so the project had the opportunity of a real proof of concept”.
Dr Bertheau added, “It would have been interesting and useful if the FELIX contractors had been closer to the competitors of the GEA challenge” (Genetic Engineering Attribution challenge, a data science competition to identify the original source of engineered DNA).
He explained, “This would have increased the sample size (the number of tested organisms in the reference database), which is still small, as an IARPA representative regretted, and would probably have provided useful feedback.”
On the positive side, the FELIX technologies can detect GMOs even in complex mixed samples that contain many organisms' DNA, highly diluted samples, and very small and “precious” samples.
Public access uncertain
At the press conference, the FELIX researchers said they were keen to see the technologies used by both private and public sector bodies, with the latter including NGOs engaged in biosurveillance.
However, no scientific publication describing the technologies in detail was announced. This suggests that access, for example, for public GMO detection laboratories to perform inexpensive analyses, remains uncertain.
No foreign DNA needs to be present
No foreign DNA need be intentionally present in the samples to detect the presence of GMOs, as the FELIX system works by analysing the sequence of the DNA and comparing it with a database of the DNA of natural organisms.
However, Drs Antoniou and Bertheau agreed that the very idea that a GMO can possess “no foreign DNA” – an idea that underpins some governments’ decisions to deregulate such GMOs – is somewhere between highly unlikely and impossible.
GM gene-edited plants and animals can and do contain foreign genes and fragments of DNA in their genomes, either by intention or inadvertently as a result of the limitations and inherent imprecision of the gene editing process. This is especially so since up to now, the vast majority of gene-edited organisms have been made using the old techniques used for transgenic GMOs, and Agrobacterium insertion is still the most efficient delivery system for the gene editing tool. It can leave behind traces of foreign DNA from the plasmids (circular pieces of DNA which encode for the CRISPR/Cas gene-editing tool), as well as from the Agrobacterium genome.
Foreign genetic material in all GMOs, including gene-edited ones, is already detectable using widely available deep and ultra-deep short- and long-read whole genome sequencing methods. So what can the FELIX programme contribute?
It could investigate how common a problem contaminating foreign DNA is and identify the unique scars and signatures of any given GM process, including gene-editing processes. It could also improve the reference databases and the accuracy of the sequencing process. Dr Bertheau says it could also identify the signatures of certain companies' laboratories, just as elite varieties’ genetic background and markers (specific sequences often used for genetic engineering) currently identify the plant breeding companies that produced certain GM plants.
Conclusion and recommendations
In GMWatch's view, IARPA, Ginkgo Bioworks and Draper have come up with a promising suite of technologies to identify man-made genetic changes. In the case of known (authorised) gene-edited GMOs, when analysis of the scars and signatures of any given GMO is joined to the detection of changes conferring the intended trait, FELIX could readily identify both the use of genetic engineering and the specific GMO, using inexpensive processes.
FELIX also provides the tools for carrying out further work to detect the use of genetic engineering in unknown and undeclared new GMOs, including gene-edited ones with small intended changes.
However, these techniques will only become ready and available to use given political will, focused investment, and cooperation from GMO developers. Enforcement laboratories could provide knowledge of genetic sequences but are hampered by their lack of access to reference material (seeds and DNA) and absence of funding. For each GMO product, GMO developers should provide good quality reference material, along with information about the GM processes used to develop each product. And the EU Commission should provide funding for ENGL to collaborate with the FELIX researchers in developing detection methods for all new GMOs. Such cooperation would lead to the expansion of reference databases for GMO and natural varieties and the consequent rapid evolution of detection methods.
Read Dr Yves Bertheau’s detailed comments on the FELIX programme here.
The FELIX researchers carried out an analysis of the SARS-CoV-2 virus to look for evidence of genetic engineering and concluded that there was none. Does this cast doubt on the rigour and integrity of their programme? Read our article.