2018-2020 Intersessional Period

----Message posted on behalf of the forum Moderator---

Dear participants of the Online Forum on Risk Assessment and Risk Management,

Welcome to the first week of the forum discussions. My name is Marja Ruohonen-Lehto, my educational background is on genetics, microbiology and virology, and I have more than 20 years of experience in Biosafety and Risk Assessment.

I have the honor to serve as the moderator of this forum and as such, I will provide support and guidance to the discussions.

Decision CP-9/13 established a process for identification and prioritization of specific issues regarding risk assessment of living modified organisms for consideration of the COP-MOP. It also entails the commissioning of studies to inform the application of annex I to (i) living modified organisms containing engineered gene drives and (ii) living modified fish. The studies are to be presented to the open-ended online forum and later to the Ad Hoc Technical Expert Group on Risk Assessment and Risk Management.

The discussions in the online forum are organized in two topics – one for each study.
This topic will be dedicated to discussing the draft study on engineered gene drives.
The draft study is available on the web page for the online forum: https://bch.cbd.int/onlineconferences/forum_ra/discussion.shtml.

For the discussions, I would like to suggest that your comments are always as specific as possible to facilitate the analysis of the information and understanding among other forum participants. In addition, I would like to encourage you to focus your comments on the substance of the studies rather than on editorial suggestions, and to share information that:

• Could complement the study, which may include: further development of concepts, explanatory comments, relevant resources, bibliographic references, among others;
• Could identify any information gaps or factual errors;
• Is relevant to one or more of the aspects of annex I of decision CP-9/13

The consultants who developed the studies will be following the discussions of the forum, and I will be acting as the main focal point to facilitate the exchange of information between them and the forum as needed.

Having said this, I now invite the participants to the Online Forum to comment on the draft study on engineered gene drives  (topic 1) for the next 2 weeks, and kindly request that you post in reply to this message only, such that all the comments are kept in one single forum thread under this topic.

I wish you all a very productive discussion.

Marja

Thank you Marja for moderating the forum, the Secretariat for hosting the forum and greetings to my colleagues participating in the forum. I look forward to many stimulating interventions.

We have been asked to share information that

(1) Could complement the study, which may include: further development of concepts, explanatory comments, relevant resources, bibliographic references, among others;
(2) Could identify any information gaps or factual errors
(3) Is relevant to one or more of the aspects of annex I of decision CP-9/13

My overall observation of the study is that it has taken a very narrow definition of gene drives and from that extrapolated to some incorrect conclusions. While I recognise that the biotechnology community is currently fixated on a current type of application, that which causes segregation distortions in organisms that are reproducing through meiotic pathways -  and the attention to these drives has been enormous - these are examples of gene drives in particular kinds of organisms but gene drives are not limited to these examples.

In their general and relevant sense, gene drives are any genetic mechanism that can distort segregation of traits. Genetic elements that do this are known in both sexually and non-sexually reproducing organisms. They may be engineered with the techniques of modern biotechnology into both sexually and non-sexually reproducing organisms, and infectious agents of these organisms (e.g. conjugative plasmids, bacteriophage). They may exert a bias in the reproduction of the organism, or in the reproduction of other infectious elements of the organism. In the latter case, they are not necessarily restricted to reproduction of a host organism to cause biased inheritance. I provide a few citations in my specific comments, attached.

Moreover, not all genes in sexually reproducing organisms follow ideal Mendelian ratios. This is because they are not in the nucleus and thus are not processed through meiosis. Such genes are found in organelles such as chloroplasts and mitochondria. Drive mechanisms, such as suppressiveness, are found in organelles. Drives may be engineered into organelles for the range of purposes that they may be engineered into chromosomes but not fit the arbitrary definition of drives in this report because they would not distort Mendelian ratios that relate to markers on chromosomes.

An example of how this leads to incorrect conclusions is a sentence such as this: “However, since gene drives are based on mating potential, the potential for exchange with related species is very species specific”. For drives in non-sexually reproducing organisms, gene drives are not based on mating potential. The kind of risk to be considered is therefore not mitigated by presumptions of species fidelity and from the report these kinds of risks do not seem to have even been contemplated. Thus, the literature review upon which the report is based has not been comprehensive.

Thank you for your patience with my long intervention.
Best regards
Jack

Dear Jack is a pleasure for me to read your comments. They are always very useful and I agree with you.
The risk assessment on living modified organisms containing engineered gene drives, is an issue very current and important because using new technologies that also lead to genetic modifications, without the transfer of genes from other species, so that properly trained multidisciplinary groups are needed. We need to know very well whats happend  with this kind of organisms. Taking into account the examples presented in the study, capacity building programs are needed in our region to identify in each case the possible hazard of these technologies.

Thank you Marja for being the moderator and the Secretariat for hosting this forum. I hope we all can have a very productive interaction.

I am scientist at the University of California Irvine working in developing gene drive for mosquito control, our approach is population replacement, which means that our main goal is to introduce beneficial traits in mosquito populations that can stop the transmission of the malaria parasite.

I totally agree with Dr. Jack Heinemann. The study presented a very narrow definition of gene drives and also approached some conclusions that might not be applicable for all the gene drives and/or their applications.

Gene drive technologies are being developed at an extremely fast pace, the prototype gene drives published a couple of years ago now are probably in a different phase in the lab, having more accurate and controllable systems. For example, for mosquitoes, the first gene drive reported by Gantz et al (PNAS, 2015) showed considerable fitness cost and generation of resistant alleles. Our current gene drive has almost no fitness cost (at least nothing that can impact the gene drive capacity) and the generation of resistant alleles is pretty much neglectable (less than 1%) in cage trials. In the case of resistant alleles, they are probably not anymore a concern for gene drives designed for population replacement in mosquitoes. So it is inaccurate to generalize this type of concern. I think the study must also include not only the gaps or possible risks but examples of how new more stable gene drives are being developed.

I would like to highlight that for all the resources we have in the present (drugs, vaccines, insecticides, etc) there is a risk on them and that most of them (if not all) have some unknowns that have not been answered yet. There is always some uncertainty linked to new products that cannot be answered by clinical/field trials. I think something important for gene drives is trying to determine the maximum acceptable risk for the technology (as it is done for every product). Probably this "maximum acceptable risk" will depend on the type of drive and thereby a case-by-case decision must be done. We as a scientist are making efforts to provide as much information needed to show that our systems are reliable and that will not impose a risk to the environment. We recently developed Target Product Profiles for mosquito gene drives for population replacement (Carballar et al, 2018, Pathogens and Global Health), this kind of information will be helpful to identify new gaps and the characteristics of a gene drive that need improvement to make it more reliable and to comply with standards.

Dear All,

I am an ecologist who is involved in the ecological risk assessment of GMO. I think we may need more knowledge/research on the effects of removing a species from an ecosystem.

Christian Damgaard, Aarhus University, Denmark

My name is Werner Schenkel, I have more than 15 years of experience in Biosafety and Risk Assessment as scientific officer at the German Federal Office of Consumer Protection and Food Safety.

Firstly, I wish to thank for the opportunity to contribute to the discussions in this online forum. Further I would like to thank G.Smets and P.Rüdelsheim and Perseus for conducting the extensive study, including comprehensive literature search, developing and evaluation of questionnaire, and thus for analysing and dissemination of available information on living modified organisms containing engineered gene drives.
My special thanks also to Marja for moderating this forum, and all colleagues sharing fruitful insights.

I think for a productive outcome of this forum it is of high importance, that the participants try to reach a common understanding of the specific issue which is being discussed.
The draft study of G. Smets and P. Rüdelsheim clearly defines the issue analysed in the first two sentence: “Gene drives allow for a trait to be distributed across generations deviating from the laws of Mendelian inheritance. Active in sexually-reproducing species, they are powerful tools to “drive” a gene, i.e. increase its frequency, independent of external selection pressure.”
This understanding of the term “gene drive” as being restricted to sexually-reproducing species is consistent with definitions generally used in scientific literature and publications of sciences academies, biosafety advisory boards and regulatory offices. Here are just a few quotations:

NASEM (2016): „In this report gene drives are defined as systems of biased inheritance in which the ability of a genetic element to pass from a parent to its offspring through sexual reproduction is enhanced.

Australian Academy of Science (2017): “A number of basic criteria are required for a synthetic gene drive to work. Firstly, the organism must reproduce sexually.”

ZKBS (2016):„Gene drive systems are understood to be genetic elements or gene constructs that expedite their own spread in populations of sexually reproducing organisms by being inherited to more than 50% of the offspring“

RIVM (2016): “A gene drive is only effective in organisms that reproduce sexually.”

NEPAD (2018): Gene drives are systems of biased inheritance in which the ability of a genetic element to pass from a parent to its offspring through sexual reproduction is enhanced (National Academies of Sciences & Medicine, 2016; Sinkins & Gould, 2006).

OGTR (2019): “Gene drives are genetic elements that are favoured for inheritance, and which can therefore spread through populations at a greater rate than genes with standard Mendelian inheritance. Gene drives can only spread from sexually reproducing parents to their offspring”

I am not aware that in any discussion leading up to decision CBD/CP/MOP/DEC/9/13 the term “gene drive” was used in context with species or genetic elements that reproduce asexually.
If we open up this definition it will become increasingly difficult to clearly describe the specific issue let alone clearly describe the challenges in addressing the specific issue, as is required in Annex I (d) of CBD/CP/MOP/DEC/9/13 in order to identify specific issues that may warrant consideration.

I strongly recommend to stick to the widely accepted definition of gene drives as systems that result in significantly super-Mendelian inheritance in sexually reproducing species.

Dear colleagues,

Greetings to all of you and all my best wishes for the New Year!

I wish to thank you for the opportunity to contribute to the online forum and the exchange of valuable thoughts through this channel. I also thank the authors of the “study on risk assessment application of annex I of decision CP 9/13 to living modified organisms containing engineered gene drives” for compiling the extensive draft report commissioned by the CBD Secretariat. It is obvious that the authors of the draft report were faced with a challenging task! Many thanks also to Marja for moderating this online forum.

I think there are a few things that could be envisaged for the revision of the draft report:

1. Scope: Given that the gene drive field is evolving at high pace and that a broad array of gene drive applications may be envisaged from a conceptual perspective, it is not feasible to capture all possible scenarios in the draft report. Instead, it may be helpful to limit the scope of the draft report to the most likely cases of gene drive modified organisms moving to practical application. This may help to focus the scope of the draft report, which is too broad and hypothetical at present.

2. Familiarity/experience with existing insect vector/pest control strategies: The draft report only briefly mentions that the use of some gene drive modified organisms has similarities with some well-established insect vector/pest control strategies, including sterile insect releases and incompatible insect technique, and does not refer to classical biological control programs. It would be important to stress more explicitly in the draft report (i.e., in the executive summary) that one can draw on this familiarity/experience when framing the risk assessment of gene drive modified insects.

3. Natural gene drives: It would be useful to specify in the draft report that the study of natural gene drive systems over the last century has provided considerable theoretical and empirical insights into how gene drives work and how they spread, and that in some cases, this could provide useful baseline information for the risk assessment of synthetically engineered gene drives.

4. Comparators: When addressing comparators, it would be helpful to indicate that the deliberate release of any gene drive modified organism should be compared to a range of comparators (including alternative solutions) to allow harms (and hence benefits) to be appropriately considered.

5. Problem formulation process: Due weight should be given to the problem formulation, as risks assessments should begin with an explicit problem formulation where protection goals, plausible and relevant exposure scenarios and the potential adverse effects from those exposures are identified on a case-by-case basis.

6. Applicability of statements: Some of the statements made in the draft report are not applicable to all cases of gene drive modified organisms. To avoid confusion and generalizations, it would be helpful to clarify to which gene drive cases or groups/categories of gene drives statements apply. Ideally, this should be done for each of the statements made in the draft report, as this would better frame the draft report, and ensure it is more in tune with the case-specific nature of risk assessment.

7. Weight of evidence substantiating statements: The lines of evidence used to substantiate some statements in the draft report and their weight are not always obvious (e.g. scientific publications vs. opinion of the authors or an interviewed person). Perhaps this point could be addressed by being more specific about the sources of evidence used to substantiate statements. Ideally, this should be done for each of the statements made in the draft report.

8. Literature search: Insufficient information about the search process and results is currently reported to enable an appropriate assessment of the quality of the literature search performed by the authors of the draft report. Moreover, the wording used in the title of Annex 1 is misleading, as an extensive literature search has been performed instead of a systematic review.

I also wish to take this opportunity to inform you that the European Food Safety Authority (EFSA) has been requested by the European Commission to assess, through a problem formulation exercise, whether: (1) the deliberate release of gene drive modified organisms could pose potential new hazards and thus risks to human/animal health and the environment, considering relevant comparators; (2) the scientific considerations/requirements given in its existing guidelines for the risk assessment of genetically modified animals are adequate for the molecular characterisation and environmental risk assessment of gene drive modified organisms; and (3) there is a need for updated guidance in relation to previous EFSA documents. In line with its policy on openness and transparency, EFSA will soon launch the draft GMO Panel scientific opinion on “the evaluation of existing EFSA guidelines for their adequacy for the molecular characterisation and environmental risk assessment of genetically modified insects with synthetically engineered gene drives” for public consultation (see EFSA’s Public consultations planner).

With thanks,
Yann Devos (senior scientific officer at EFSA)

Dear participants to the forum,

Before providing specific comments to the questions of our moderator, I would start with a brief introduction. My name is Ana Atanassova and I work as a global regulatory policy manager for BASF Agricultural Solutions and have more than 20 years’ experience in research and development of agricultural biotech products and good understanding of the regulatory frameworks around the world governing the use and commercialization of GMOs.

My participation in this forum also continues the long-standing engagement of the plant science industry, through our representative organization – the Global Industry Coalition (GIC) to take part in the discussions on risk assessment and risk management of LMOs, new technologies and other relevant topics under the CBD. Over the years, the GIC has been sharing our practitioner’s perspective in the CBD fora. Our long-term engagement in R&D activities and commercialization of GMOs around the world has provided us with insights of how LMO regulations are implemented in practice in different countries.
Irrespective of whether or not a country is a Party to the Cartagena Protocol on Biosafety, the general principles and methodology described in Annex III of the Protocol are applied routinely by all risk assessors.  Some of the participants to this forum that have also been engaged in previous discussions of RA and RM may well recognize our position that the main challenges with risk assessment is linked to the level of practical experience of the risk assessors. As such, development of additional guidance martials on specific issues may not address successfully the most pressing needs of such risk assessors. Access to the right scientific information and expertise can also be a challenge and the information exchange happening under the CBD is a valuable tool to raise awareness about resources and expertise that is available.
The GIC has contributed to the information submissions on gene drives in 2019 (see https://bch.cbd.int/database/record.shtml?documentid=114561 ) where we also expressed our strong support for the decision by the Parties to establish and use a process for the identification and prioritization of specific issues regarding risk assessment that may warrant the development of additional guidance. We hope that the current discussions of the on-line forum as well as the work of the AHTEG will test the criteria for identification and selection of issues outlined in Annex I of Decision CBD/CP/MOP/DEC/9/13 and that, if needed, these are strengthened and refined by the Parties following the outcomes of the current exercise on engineered gene drives and LM fish.


In terms of concrete and substantive responses to the moderator’s questions, in this post I will provide comments on the draft study of Perseus on GD focusing on the points outlined by our moderator to share information that
1) Could complement the studies, which may include: further development of concepts, explanatory comments, relevant resources, bibliographic references, among others;
2) Could identify any information gaps or factual errors;

In a follow up post, I will address the third specific point and discuss information that is relevant to one or more of the aspects of annex I of decision CP-9/13.


The Perseus report

The report presents a good description of the technology, proposed uses and issues of concern. It raises several concerns that are related to gene drive but are not specific to it including capacity building, protection goals, informed consent and international agreements.

Strong points
• The report affirms the ‘case-by-case’ approach to risk assessment. It acknowledges that ‘gene drive’ comes in a wide range of types for population replacement and suppression and drives that are both low and high threshold, which entail case-by-case treatment. While ‘technical and methodological challenges’ exist (page 6, page 31, lines 1-2), these are largely scientific challenges to providing information rather than risk assessment methodology.
• The report acknowledges that population replacement vs suppression presents very different risk considerations.
• Discriminates between domesticated species that have been modified and wild populations which gene drive intends to alter.
• The issue of potential irreversibility is particularly noted.
• The report repeatedly notes that most interviewees felt existing RA approaches were adequate to address gene drive and correctly identifies that while unique features of gene drives exist, ‘most interviewees anticipated it will be possible to use existing risk assessment methodology’ (page 5, line 27; page 30, lines 11-12, page 31, lines 17-20)
• There is generally a differentiation between (1) risk assessment (RA) methodology and (2) relevant information that may be (in)adequate to prepare a gene drive RA, however this topic can benefit from further discussion and clarification before the report is finalized.
• Transboundary movement and its potential implications are described. This, and the potential irreversibility are identified by the authors as the most challenging areas for assessment of engineered gene drives.
• The authors recognize and underline that many issues raised in discussions in the context of gene drives are not unique to that technology which indicates that the concerns are familiar and have been considered for existing technologies (page 5, line 36; page 30, lines 26-31). The pitfall of generic discussions of gene drives rather than considering specific cases is acknowledged as confounding understanding and arriving at recommendations (Page 31, lines 40-45).


Areas where the report could be improved on
• The technologies that are identified as most relevant experiences to inform RA are generally limited to mosquito genetic control. But with regard to ecological significance, there are other control technologies in wide use that could be informative e.g. biocontrol and conventional population suppression methods. This leaves the reader with the impression that informed decisions cannot be made without bespoke studies. Identifying other models of species-specific suppression as sources of ecological effects would be a useful effort to supplement the report.
• The report is equivocal on the issue of species-specificity of engineered gene drives. At times it claims this is strong, in others a weak risk-introducing concern. However, it does note examples in which a close relative of a pest might be beneficial but does not also notice that in some systems where closely related species occur, they are all potential disease vectors (Anopheles gambiae).
• The receiving environment is often assumed by the authors to be defined too conservatively. This assumes an application for release or RA that does not include the entire likely affected area.
• There is often mingling of discussions of information needed to conduct an RA and the adequacy of guidance. This makes the issues unclear at times and the authors could clarify better the difference between methodology and specific information needs to conduct RA, as well as the value of “guidance” in support of RA.
• The report provides little context for the uses of gene drives and the potential beneficial effects. Therefore, risk is considered against a backdrop of no harm in the event of inaction. However, the purpose of the research and development for potential gene drive applications is to reduce harm due to disease or invasive species which threaten biodiversity or human health. This will affect what is considered acceptable risk.


With kind regards to all the participants of the forum, and looking forward to the interesting and informative posts and discussion,

Ana

Dear Maria and colleagues on the online forum. Thank you for the opportunity for me to provide input into the study on engineered gene drives. As requested, I am offering the attached comments in order to provide information that

(1) Could complement the study, which may include: further development of concepts, explanatory comments, relevant resources, bibliographic references, among others;
(2) Could identify any information gaps or factual errors
(3) Is relevant to one or more of the aspects of annex I of decision CP-9/13

In addition to the attached comments, I wish to support the opinion of Werner Schenkel that we keep the current focus of our discussions on the scope of gene drives as we presently understand it. Finally, I note that the report does not provide us with a means to conclude that the criteria in annex I of decision CP-9/13 support the development of new guidance on gene drives. It has not been agreed upon how the criteria will be applied to drive that decision. Do all criteria have to be met? Only one? This is an issue that requires resolution.

Thanks again for this opportunity.

----Message posted on behalf of Moderator----
Dear Participants of the forum,

thank you very much for opening the discussion in a lively and comprehensive way. This is very productive! I hope that within the coming days more people will join.
Please keep in mind our focus of the discussions and provide comments that are as specific as possible – this is very important for the revision process of the document.

I have compiled, assisted by the Secretariat, a list of issues that have been raised:
- Scope of the definition of gene drives: some suggestions to use the term gene drive in consistency with its use in the literature, regulatory offices and other entities (as is the approach of the study), while others are of the view that the definition currently used in the study is too narrow;
- Suggestion to improve the literature review of the study / on the way the literature search was conducted;
- The fact that not all the conclusions and /or statements of the study may be equally applicable to all types of gene drives;
- Suggestion for the study to also consider information on how new and more stable gene drives are being developed;
- The need for more information on the effects of removing a species from an ecosystem;
- There have been suggestions on how to deal with the scope of the study and with comparators, as well as on the need to point out to the familiarity with existing insect vector/ pest control strategies;
- Comments about strong points and areas for improvement of the study; and
- Comment referring to the fact that it has not been agreed upon how the criteria from annex I will be applied.

We have still more than a week to go! I look forward to more discussions and want to thank you for taking time from your busy schedules to participate into this important discussion.

Please keep in mind, that for interventions at the last minute you will not receive any important feedback :)

All the best,

Marja from Helsinki, Finland

I again extend my gratitude to colleagues for their thoughts, insights and critical comments in this forum and appreciation to Marja for her ongoing care in moderating the forum.

I have reflected on other interventions and decision CP-9/13. The purpose of the commissioned study described in the Perseus Report was described in the decision as to facilitate “a process for the identification and prioritization of specific issues regarding risk assessment of living modified organisms...with a view to developing further guidance on risk assessment on the specific issues identified, taking into account annex I;”.

Furthermore, the terms of reference for the AHTEG say that it should “Consider the needs and priorities for further guidance and gaps in existing guidance identified by Parties in response to decision CP-VIII/12 with regard to specific topics of risk assessment and prepare an analysis.”

In my interventions I put forward that there are at least two gaps which the draft report on gene drives does not either identify or close in its present form.
1. Different aspects of a risk assessment depending on whether the vector has been constructed and tested in a laboratory vs those that are created in the environment, as was raised in the submission from Mauritana (110734 dated 2016-08-02) at least, and which I have also covered as part of my research (and references therein):

Heinemann, J.A. Should dsRNA treatments applied in outdoor environments be regulated? Environ Int 2019;132:104856
Heinemann, J.A.; Walker, S. Environmentally applied nucleic acids and proteins for purposes of engineering changes to genes and other genetic material. Biosafety Health 2019;in press

2. Gene drives in LMOs that do not use meiosis ever (i.e. prokaryotes) or sometimes (which includes species of fungi, plants and animals) and the organelles or inseparable symbionts in the cytoplasm of organisms (including viruses and plasmids).

Sexual reproduction is but an example of a mechanism by which drives can spread. Since the 1950s, gene drives dependent on meiosis have been called “meiotic drives” because they use sexual reproduction to spread.

“Instances are known both from studies of natural populations and laboratory experiments in which heterozygotes produce two kinds of gametes, not with the customary equality, but with unequal frequencies. Such meiotic behavior will profoundly affect gene frequencies in populations, and is referred to as meiotic drive.” L. Sandler and E. Novitski, "Meiotic Drive as an Evolutionary Force," The American Naturalist 91, no. 857 (Mar. - Apr., 1957): 105-110.

The use of gene drives in asexual organisms is also discussed in the current biotechnology literature, such as in de Lorenzo, V. Microbial Biotechnology (2017) 10( 5), 995– 998. “The issue involves not only genetic assembly of all biological activities necessary to manifest the desired activity, but also its propagation through the environmental microbiome through either horizontal gene transfer or some type of prokaryotic gene drive, e.g. with engineered phages.”

Werner Schenkel presented in his intervention [10114] a list of other reports that included coverage on meiotic drives. In those cases sexual reproduction is important for spread because of the biology of the organisms being discussed. But as I’ve illustrated, that is a feature of the biology of the LMO, not the distorting effect of a drive which still cause distortions in other organisms if designed appropriately for their reproductive or infectious cycles. Limiting discussion to drives on nuclear genetic elements in sexually reproducing organisms may be understandable in the context of other discussions because those may have been conducted under terms of reference that were more limited than the Convention on Biological Diversity. In any case, exclusion of gene drive mechanisms based on arbitrary definitions risks making the work of the AHTEG obsolete before it even begins.

Furthermore, the types of drives (also called selfish genetic elements in some literature) that do not require sexual reproduction to either propagate or distort conform completely to the criteria listed in Annex I. I won’t repeat them all here, but make special note of paragraph (c): “They pose challenges to existing risk assessment frameworks, guidance and methodologies, for example, if the issue at hand has been assessed with existing risk assessment frameworks but poses specific technical or methodological challenges that require further attention” because, as discussed above, existing reviews from NASEM, OGTR and others have been exclusive to meiotic drives. Thus, the two items I raise are particular issues that pose challenges to existing guidance because their specific technical or methodological challenges are not addressed in existing guidance.

Marvis Suárez Romero [10118], Rebecca Carballar-Lejarazu [10119] and Yann Devos note that this field is moving quickly and is broad. Yann Devos [10127] offers a compromise position in suggesting that the draft report focus on the most likely drives for practical application. That is a sensible suggestion, but unless we had a study that included surveys of drives other than meiotic drives we would not know which drives are most likely to be applied first. For example, a company in Madison Wisconsin USA called Conjugon which was founded 19 years ago (in 2001) was/is developing a plasmid-mediated drive to eliminate antibiotic resistance in bacteria. They just don’t use the term “gene drive” in their description.

If the draft report is finalized as a report only on meiotic drives, I think at the very least that qualification should be in the Executive Summary and accompany any actual or implied definition of gene drives. The limitations of the report because it applies only to meiotic drives should also be highlighted in the section on risk assessment.

Thank you again colleagues for this stimulating exchange of ideas.
Jack

Dear All

Let me introduce myself: my name is Christoph Then, I was nominated by ENSSER (European Network of Scientists for Social and Environmental Responsibility) to participate in the debate. My main affiliation is with Testbiotech (http://www.testbiotech.org).

First of all, I would like to say that Testbiotech and ENSSER have just released a number of studies from the RAGES project (http://www.testbiotech.org/en/content/research-project-rages) on the risk assessment of GE (genetically engineered) organisms in the EU. One of the reports is on the issue of gene drives (http://www.testbiotech.org/en/content/rages-subreport-new-genetic-engineering-technologies). This report concludes that, from the perspective of the precautionary principle and EU GMO regulation, releases of GE gene drive organsisms should not be allowed.

I would also like to introduce the report on GE gene drives published by ENSSER in 2019 (https://genedrives.ch/wp-content/uploads/2019/10/Gene-Drives-Book-WEB.pdf). It comes to similar conclusions as RAGES, but also includes ethical and social considerations.

Finally, I would like to mention the outcome of the GeneTip project (http://www.genetip.de/en/biotip-pilot-study/) in which Testbiotech participated and from which some results have already been published (Frieß JL, von Gleich A and Giese B, 2019. Gene drives as a new quality in GMO releases – a comparative technology characterization. PeerJ, 7, e6793. https://doi.org/10.7717/peerj.6793) – others are still awaiting publication. This project largely dealt with technology assessment and in its conclusions emphasised the role of the precautionary principle.

However, in regard to the report compiled by Perseus, the issue of the precautionary principle seems to have been poorly considered, with only random observations. This needs be corrected. Therefore, I suggest the group works on this specific gap in the report.

With kind regards

Christoph Then

Dear Marja, dear all,

My name is Boet Glandorf, I am a senior risk assessor with more than 20 year experience and am actively involved in discussions on the environmental risk assessment (ERA) of organisms with engineered gene drives, such as in the workshops that are mentioned in Table 2 of the report.

In my opinion the study by Perseus gives a good overview of the topics that are (or may be) applicable for organisms with engineered gene drives in general. However, I like to follow up on earlier reactions in this online forum that pinpoint to a number of issues that are essential for any environmental risk assessment, and even more so for organisms with engineered gene drives.
The following could be improved in the report:
- It could be stressed more that any ERA can only be performed on a case-by-case basis, and not in a general way.  It is clear that some of the concepts, such as a low threshold drive, will not be applicable to e.g.  yeasts or mammals with a gene drive, but may be applicable to flying insects with a gene drive.  And potential risks like effects on unmanaged ecosystems will not be applicable to gene drive mosquitoes that live in close proximity to humans, but may be applicable for other gene drive insects.
- The report could benefit by focusing on (or including) examples of organisms with engineered gene drives that may expected to be environmentally released in the coming 10 years, such as malaria mosquitoes or Aedes mosquitoes modified to suppress important human diseases.
- It could be stressed more that, in contrast to GM plants, there is little experience with the ERA of LM insects. A number of aspects that are mentioned in the report to be specific for insects with a gene drive, are actually only related to the fact that it concerns a LM insect that is able to spread and propagate and are not directly related to the gene drive itself.
- With respect to the baseline, it could be included that there are existing ‘natural’ gene drive systems in insects, such mosquitoes with Wolbachia, that are released in large quantities for vector control. In addition, reference could be made to the long experience that exists with releasing living insects for biocontrol.
- Another aspect worth mentioning in the report is that effects of gene drive insects for vector control should also be considered in the context of  current vector control methods, like spraying with chemicals.
- A last remark is on 2.2.3 where there is mentioning of a selection of stakeholders (in 2.2.3) that ‘had already communicated on the topic’. They were included with the expectation ‘to broaden the collection of views on the subject’. These stakeholders are listed in Table 3 of the report.  The report would benefit from indicating where the selection of stakeholders was based on and how this information is used in the report.

Kind regards,
Boet Glandorf, Institute of Public Health and the Environment, the Netherlands

Dear Colleagues,

My name is Nikolay Tzvetkov and I work at the Bulgarian Ministry of Environment and Water where my responsibilities include the regulation of GMOs and biotechnology in general.

First I would like to congratulate and thank Marja for moderating the forum.

Regarding the study presented by Perseus provides a pretty good overview of the field of gene drives and the related topic of the environmental risk assessment. A lot of suggestions for improvements have been made in previous contributions (# 10117, # 10119, # 10122, # 10127, # 10128, # 10129, #10134 and # 10137) and as I agree with many of them I see no sense repeating them.
I would like specifically to support the suggestion by Boet Glandorf (# 10137) to provide examples of organisms with engineered gene drives that may expected to be released into the environment in near and midterm future. It will be useful to supplement those examples with brief discussion of the type of gene drives that will be most useful to achieve the objectives of such a release. My personal opinion is that when we talk about vector control (either by population suppression or by modification of the vector potential) or removal of IAS, it will be beneficial to use highly efficient low threshold drives.
I think it is beneficial that the report focuses on gene drives that require sexual reproduction for transmission. Other systems are very interesting indeed, but focusing on them can make the discussion too broad.

Regarding the analysis of the organisms with engineered gene drives vs the criteria set in Annex I of decision CP-9/13 my view is that the topic essentially fulfils them all. The references provided (section 5.2) cover many relevant aspects of ERA of organisms containing engineered gene drives but cannot considered as guidance documents as such. In this respect it will be very interesting to see the results of the exercise conducted by EFSA that is mentioned by Dr Devos (#10127).

Regarding the ERA my own opinion is that the general methodological approach to ERA of LMOs/GMOs will be applicable in the case of living modified organisms containing engineered gene drives. But as the report shows there can be a lot of challenges related to the lack of data, more complex interactions of those organisms with the environment and potential to have irreversible effects that are transboundary. This makes an even stronger case for case-by-case ERA and we should be fully aware of the higher levels of uncertainty (both of the input information and of the conclusions reached).

So are the living modified organisms containing engineered gene drives good topic for development of guidance documents for ERA. In principle I would think yes, but practically it may not be possible to develop such general document due to the differences of the biology of the organism (e.g. mosquitoes vs mice vs plants), the specific gene drive system (low threshold, high threshold, self-limiting, etc.), the cargo carried, the potential receiving environment (agroecosystems vs natural ecosystems, small islands vs global populations, etc.) and so on.

I think it will be more useful if document is developed that outlines the general principles and considerations that should be taken into account when conducting ERA of living modified organisms containing engineered gene drives with the view of the release into the environment. Such document should ensure that ERA is carried out to the acceptable standards using the appropriate data and methodology. It also should ensure that all stakeholders and potentially affected parties and communities have been consulted.

Best Regards,
Nikolay

In relation to what was expressed by the colleague Hector Quemada, I think that an important point to consider by the regulatory authorities of our countries is related to the regulations that should be applied or not for the organisms obtained from new breeding techniques that apply modern biotechnology to generate genetic modifications. A good example is the Argentine biosafety legislation that already includes this topic and the risk assessments are carried out based on the case-by-case scientific analysis, an issue that has been pointed out by several colleagues in the forum and in the Perseus report itself.

Dear Marja, online-forum organisers and fellow participants,

My name is Keith Hayes, I am a senior research fellow at the Australian Commonwealth Science and Industrial Research Organisation (CSIRO). I lead a small team who specialise in environmental risk assessment (https://data61.csiro.au/en/Our-Research/Programs-and-Facilities/Analytics-and-decision-sciences/DEERA). My team and I have completed two risk assessments for the importation, contained use and controlled field release of genetically modified, male-sterile, mosquitoes in West Africa (https://targetmalaria.org/resources/). The male-sterile mosquitoes are the first stage in the development pathway of a self-sustaining genetic control technology intended to reduce the burden of malaria in the Africa by reducing transmission of the malaria parasite.

I would like to thank Marj and the forum organisers for the opportunity to provide the following information that: (i) could complement the study conducted by Smets and Rudelsheim; (ii) could identify any information gaps or factual errors; or (iii) is relevant to one or more of the aspects of annex I of decision CP-9/13.

With regard to the definition of gene drives I support the previous suggestions (for example #10129, #10122) to maintain the current focus of the study to sexually reproducing organisms, however, I don’t believe this precludes some discussion (perhaps in an Annex) of other types of selfish genetic elements, as suggested in submission #10117.

I also agree with submission #10137 (and also #10127) that the report could be improved by a closer examination of the risk associated with gene-drive modified mosquitoes (GDMM) because these are the most advanced examples to date and hence the most likely candidates for environmental release in the next decade. The risk assessments we have completed for the recent field-trials of male-sterile mosquitoes in West Africa highlight a number of specific issues, and risk analysis methods, that are directly relevant to gene drive modified mosquitoes.

Important issues omitted from the Perseus report (but prominent in the Topic 2 analysis) are gene by environment interactions, and the associated uncertainty when using laboratory-based observations to predict outcomes in the field. Our analysis provides examples of how to account for and quantify this uncertainty.

Ultimately, however, as in any science-based assessment of a new process or product, this uncertainty cannot be resolved until risk predictions are carefully compared to post-release outcomes (submission #10119 makes a similar point). In this context the report cites examples of post-release surveillance of Wolbachia infected Aedes aegypti mosquitoes. The report should, however, acknowledge that these mosquitoes are known to have dispersal distances that are significantly smaller than, for example, the dominant Anopheline malaria vectors in Africa. This has important implications for post-release monitoring strategies.

I also agree with submissions #10127, #10128 and #10137 that it is important to view the potential environmental impacts of gene drive modified organisms in light of the effects of the environmental impacts of current control techniques. In this context it may also be useful to highlight the GDMM risk-acceptance principal advocated by James et al (2018) to do “no more harm to human health than wild-type mosquitoes of the same genetic background and no more harm to the ecosystem than other conventional vector control interventions”.

I believe the report’s assertion that technologies for population replacement are likely to induce less ecological harm than population suppression technologies, because the former does entail the removal of a population, should be viewed with caution. This is a generalisation that ignores important case-by-case differences between target species, and the possibility of other risk pathways, such as horizontal gene transfer, whose likelihood may increase as the number (and duration) of genetic constructs in the environment increases.

I support the report’s conclusions that the potential for impacts in non-managed systems, with non-domesticated species, across potentially larger spatio-temporal scales (relative to LMO’s assessed to date), presents technical and methodological challenges. I would emphasise, however, that these challenges are not novel, nor insurmountable, and suitable risk assessment techniques are available to address them. The critical challenge in this context is developing the technical capability needed to address these challenges within relevant national and trans-national governance structures.

Dear all,
Thank you for a very informative discussion. I am an innovation researcher based at African Centre for Technology Studies (ACTS) in Kenya and very happy for an opportunity to contribute to this important and timely discussion.
I concur with others with regards to the Perseus report providing a good baseline of the topic under discussion. Good comments have been raised that may improve the report. I particularly take note below some points that might be important for economies under transition – particularly countries in Sub Saharan Africa.
A) General comments
1) The report may consider specific case studies of organisms that have a high likelihood of being released soon and where there is considerable information in the public domain pertaining to the same especially in the area of public health particularly malaria control. There are other areas of consideration but with limited information available like challenge on Invasive Alien Species.
2) As some participants have observed, there is baseline information that relates to application of of biocontrol agents to combat pests and diseases. This information if unpacked can inform requisite risk assessment and monitoring strategies in different contexts.
3) On risk assessment: I support the emphasis on a case-by-case approach but further support the recommendation by NASEM (2016) for consideration of ecological risk assessment approaches for gene drive applications.
4) Where suppression gene drives may be considered and in the event that there is a likelihood of extinction of target or host organism, the need to explore other control mechanisms is an idea that is alluded to in the report but which should be expounded (page 26: section 4.3.1).

B) Specific comments
5) The report may explore the following aspects that may enrich the current report.
a. Capacity building at different scales needed for emerging economies: Many countries may argue that the risk assessment approach for normal LMOs may suffice in the case of gene drive engineered LMOs. The report seems to suggest that with the latter a more rigorous and elaborate risk assessment will be necessary. I want to argue that this raises capacity challenge at different levels/scales (technical, regulatory, policy and institutional) considering the multi-disciplinary nature of the gene drive product deployment value chain. I cite for instance the different expertise beyond biologists that will be required to deliver an objective decision on a specific risk assessment.
b. Values with regard to biodiversity conservation:
The report may want to put emphasis on this aspect considering that values influence how different communities of practice assess new technological applications. With regards to gene drives and biodiversity, values might be critical in different respects e.g. in evaluating likelihood of species extinction; creation of proxies for extinct species or in evaluating ethical issues surrounding applications of gene drives more generally.
c. Public consultation: Gene drive applications will certainly require public consultation which might address the interests of IPLCs. In my view it is not just a “public” consultation that might be required, but a strategic “stakeholders” participation. This is because unlike normal LMOs, diverse communities of practice including the IPLCs have a stake in the gene drive application in question which is confounded further by the scarcity of information. Deliberate stakeholders engagement might be a more strategic way of enhancing legitimacy of the process as well building trust among the different groups. 
6) Resources on similar issues that may complement this report:
a. IUCN efforts especially - IUCN publication “Genetic frontiers for conservation: An assessment of synthetic biology and biodiversity conservation” report. This report is available on the IUCN Library portal. Section 2 – “Governance of synthetic biology and biodiversity conservation” pages 19-46 is particualry quite informative.
The report is work of a taskforce commissioned by IUCN after the World Conservation Congress 2016 to: examine the organisms, components and products resulting from synthetic biology techniques and the impacts of their production and use, which may be beneficial or detrimental to the conservation and sustainable use of biological diversity and associated social, economic, cultural and ethical considerations; and to assess the implications of Gene Drives and related techniques and their potential impacts on the conservation and sustainable use of biological diversity as well as equitable sharing of benefits arising from genetic resources.
b. There are other non-binding standards by industry and researchers (who are important stakeholders) providing recommendations towards responsible research. The current report may consider drawing on some relevant insights that can inform specific practice relevant for minimisng risk (e.g. Akbari et al. 2015; Emerson et al. 2017 and FINH, 2018). 
-Akbari, O. S., et.al. 2015. Safeguarding gene drive experiments in the laboratory. Science 349 (6251):927–929. https://doi.org/10.1126/science.aac7932
-Emerson, C., James, S., Littler, K. and Randazzo, F.F. 2017. Principles for gene drive research’. Science 358(6367):1135–1136. https:// doi.org/10.1126/science.aap9026
-Foundation for the National Institutes of Health (FNIH) 2018. Gene Drive Research Sponsors and Supporters Forum. Available at: https://fnih.org/what-we-do/current-lectures-awards-and-events/gene-drive-research-forum

Ann Kingiri

Thank you Marja for moderating this important discussion.

Further development of concepts, explanatory comments, relevant resources, bibliographic references, among others; The following are areas that needs further consideration: 1) there is considerable interest in applying gene drives to currently intractable invasive across a broad taxonomic range. While acknowledging that different authorities and research have already indicated the need for containment measures for preventing unintended releases of LMOs with an engineered gene drive. The applications of the technology on vector control or control of invasive species will require deliberate introduction in the environment. This entails that, control option yet to be trialled in the field is to deploy transgene-based ‘Gene Drives’: technologies which force the inheritance of a genetic construct through the gene pool of a wild population, suppressing it or replacing it with a less harmful form. However, not all species will make efficient or safe targets for these technologies. Additionally, the safety and efficacy of these systems will vary according to where they are deployed, the specific molecular design chosen, and how these factors interact with the ecology of the target pest. Given the transformative but also controversial nature of gene drives, it is imperative that their first field trials are able to successfully demonstrate that they can be used safely and efficiently.

An additional example as outlined by Campbell et al, 2015 considered the next generation of rodent eradications and the need of innovative technologies and tools to improve species specificity and increase their feasibility on islands’. The control of invasive species would be enhanced through the addition of novel, more effective and sustainable pest management methods. One control option yet to be trialled in the field is to deploy transgene-based ‘Gene Drives’: technologies which force the inheritance of a genetic construct through the gene pool of a wild population, suppressing it or replacing it with a less harmful form. However, not all species will make efficient or safe targets for these technologies.

In general, the success rate of gene drive technology is based on the ability of the host species that are highly dispersed, allow high homing rates and where the drive has a positive impact (or limited cost) on fitness. Therefore, understanding how life-history factors influence these key parameters could facilitate the screening of diverse plant species to identify their potential as targets for genetic management. Theoretical modelling, based on population demographic and genetic dynamics, provides a means of predicting the general influence of life-history traits when a drive is introduced into populations.

The other element that need to be broadened is the consideration of RNAi based insecticides will provide regulatory precedents

Identified any information gaps or factual errors;Information gaps guiding the stepwise approach on the area of low threshold systems. Various literature and the current study suggested that low threshold systems will require a comprehensive risk assessment much earlier in the staged introduction and depending on the case, specific framing of questions may be required and the data requirements may need to be adapted to reflect the type of organism, gene drive system, payload gene, environment of the release and the type of application.

In addition, Barrett, 2019 provided some additional information on how to gain insight into optimal ways to design and deploy synthetic drive systems. The study investigated the diversity of mechanisms underlying natural gene drives and their dynamics within plant populations and species. It also reviewed potential approaches for engineering gene drives and discussed their potential application to plant genomes.

Citation:
Barrett LG, Legros M, Kumaran N, Glassop D, Raghu S, Gardiner DM. 2019 Gene drives in plants: opportunities and challenges for weed control and engineered resilience. Proc. R. Soc. B 286:


Is relevant to one or more of the aspects of annex I of decision CP-9/13:Yes it is relevant to the criteria as outlined in annex I of decision CP-9/13
However, the report provides us with list of resources that the interviewees has recommended but it doesn’t provide substantive argument on whether or not  there is a need for the development of new guidance on gene drives

Dear all—

My thanks to our moderator Marja for her guidance and to all the participants for their helpful contributions so far. 

I am Bob Friedman, an ecologist by training and now a policy researcher with the J. Craig Venter Institute.  I have been closely following the development of gene drives since 2015 and have served as a member of the three Synthetic Biology AHTEGs.

I would also like to thank the authors of the Perseus report and agree with many other participants that the report provides a good overview of gene drives and useful information to help answer the questions posed in Annex 1 of Decision CP-9/13.  That said, I also have suggestions for improvement. 

First, I think the report’s authors missed an important opportunity by limiting their interviews to competent national authorities and risk assessors.  Important insights can be gained from: 1) developers of gene drives and 2) population modelers directly examining the dynamics of gene drive organisms (many working directly with developer teams).  Such interviews would provide additional “horizon-scanning” regarding new gene drive designs and the risk assessment capabilities and questions being explored by the current developer teams.  As an example, Rebeca Caballar [#10119] provides an update on research by one prominent team. I have attached a recent article in Issues in Science and Technology (the policy journal of the US National Academy of Sciences and Arizona State University) by Omar Akbari (a developer), John Marshall (a population modeler) and me that describes the breadth of designs under development (for example, geographically localized and time-limited designs), several of which are not yet to publication stage.

Second, as pointed out by several others (Yann Devos [#10127], Ana Atanassova [#10128], Boet Glandorf [#10137], and Ann Kingiri [#10147], important lessons can be learned from experiences with other biological control methods, as well as traditional chemical pest control.  The Perseus report does include some discussion of control using Wolbachia, but omits, e.g., both traditional and newer Sterile Insect Technique (SIT) designs.  Interviews with national authorities who have reviewed these types of products, as well as applications to use rodenticides to control invasive rodents on islands, would provide useful information about the capability and challenges of understanding ecosystem impacts as part of the risk assessment process.

A key component of Annex 1 of Decision CP-9/13 is a “stock-taking exercise” to useful resources  developed by other bodies.  A key oversight is the omission of the activities and procedures used by the World Health Organization’s (WHO) Vector Control Advisory Group (VCAG), which reviews a wide variety of new vector control methods, including mosquitoes containing engineered drives.  VCAG not only advises WHO, but also provides guidance to product developers on providing the evidence base needed to carry out the assessment.  A general description of VCAG’s activities is here: https://www.who.int/vector-control/vcag/en/ A list of products under evaluation here: https://apps.who.int/iris/bitstream/handle/10665/274451/WHO-CDS-VCAG-2018.03-eng.pdf?ua=1.  These include both suppression and replacement gene drives for Anopheles mosquitoes that carry the parasite that causes malaria; Wolbachia and SIT approaches to control Aedes mosquitoes that carry the parasite that causes Dengue; as well as a variety of chemical control methods. 

Finally, I would like to add my strong agreement with those participants who suggest that additional general guidance on risk assessment of gene drives is not a top priority.  Many have expressed (including the authors of the Perseus report) that risk assessment can only be performed on a case-by-case basis.  Keith Hayes [#10145] states that the risk assessment challenges posed by gene-drive containing organisms “are not novel, nor insurmountable, and suitable risk assessment techniques are available to address them. The critical challenge… is developing the technical capability… within relevant national and trans-national governance structures.”  Nikolay Tzvetkov [#10139] points out that general guidance, though in principle sounds useful, may not be practical or possible given the wide variety of organisms, gene-drive designs, and applications possible.   

Regards to all.  I look forward to continued interesting and informative discussions during the remaining time of this online forum.

Bob Friedman

Dear Marja. Thanks for opening the discussion on gene drive.
My name is Abdallah Ramadhani Mkindi from Tanzania.
We need also to consider ethical issues associated with gene drive risks because they become more complex when social amplification is taken into account. Social amplification of risk is a term of art used in the social sciences to cover a variety of social and psychological factors that influence people’s perception or judgment about the level, seriousness and degree of threat associated with a given risk or risky situation. It is a fact that social factors clearly influence people’s estimate of the nature and extent of risk, The publics who encounter gene drives through reading, media, campaigns, research or public health teams developing gene drives will develop estimates of the risk of gene drives that diverge from that of the expert community.

Also, capacity issues of carrying such assessment, especially in developing countries might be a challenge.

Dear all,
My name is Yongbo Liu, from Chinese Research Academy of Environmental Sciences.
The report document provides relatively rich information on engineered gene drives that have huge potential in the application of diseases and others.
I still focus on the ecological risk assessment of LMOs for biodiversity during environmental release that is very important to protect nature. And, scientists and policymakers have proposed and adopted a range of ecological risk assessment approaches and systems. In my view, thus, the document lacks of detailed information on processes, methods and technology of risk assessment.
Secondly, in general, local administration departments have relatively rational processes of risk assessment for LMOs. However, as we know, engineered gene drives are now limited to laboratory experiments or only in modelling, and no actual application events in nature, which results in kinds of risks and risk assessment processes that mentioned in the document were from the studies on LMOs. Thus, I suggest to conduct “risk warning mechanisms” to deal with the unforeseen emergencies and risks once the environment release unintentionally in the future.
Thanks.

Refs:
Liu and Stewart. 2019. An exposure pathway-based risk assessment system for GM plants. Plant Biotechnology Journal. 17, 1859–1861.
Wedell N, Price TAR, Lindholm AK. 2019 Gene drive: progress and prospects. Proc. R. Soc. B 286: 20192709.

Dear colleagues,
many, many interesting interventions since last Thursday. Very many thanks for that! Especially valuable are your concrete suggestions on missing information and issues that you suggest to add to the study. I will reflect these together with the Secretariat and see how they assist in analyzing the issue against the criteria in annex I. The consultants of the study are following the discussions, and they are also taking note of the comments for further analysis.

Please keep in mind the decision 9/13, what it says about the focus of the study, and the criteria in annex I. The study is supposed to inform the RARM AHTEG in the application of annex I.

While a discussion on RARM of organisms containing engineered gene drives in general is very interesting, and I see your enthusiasm on the subject itself, including the suggestion of information that would be useful for this purpose in the future, our discussion at the moment should be aimed to assist in providing information under the criteria listed in annex I.

I think that it is of utmost importance to follow the process we agreed upon at COP-MOP 9. The decision on whether new guidance is needed or not will be done in COP-MOP 10. And the studies and our discussions are needed for that very process.

Once more, many thanks for your interest, time and comments!

Still 5 more days to go!

Marja from rainy Helsinki (we should have snow!)

My thanks to all contributors for the many valuable insights.
In my view, two main tasks are ahead on the road to Kunming:According to Annex II (c) of CBD/CP/MOP/DEC/9/13 the ATHEG is requested to make recommendations on
(i) the need for guidance to be developed on risk assessment of living modified organisms containing engineered gene drives and living modified fish, and
(ii) any adjustments to annex I of decision CP-9/13.

COPMOP 9 also decided to extend the online forum on risk assessment and risk management in order to assist the ATHEG. The first task shall be accomplished according to Annex I, and supported by the data in the study.
Marja asked to share information that is relevant to one or more of the aspects of Annex I.
For this reason, I am trying to conclude on the criteria laid out in Annex I based on the information provided in the study.
Annex I (a) Priorities identified by Parties:
The specific issue of LMOs with engineered gene drives has been identified as a priority topic some by Parties. However, it should be considered what the specific needs of these Parties are in order to decide if a guidance is the best way forward to meet these needs.
Annex I (b) Scope and objective of the Cartagena Protocol:
It is out of question that organisms containing engineered gene drives are LMO according to the definition of the Cartagena Protocol and, therefore, fall within the scope of the Protocol.
Annex I (c) Challenges to existing risk assessment frameworks:
Specific challenges have been identified. However, I would agree with those interviewees reflected in the draft study who stressed that it should be differentiated between challenges to the risk assessment methodology and challenges relating to obtaining information required to inform the risk assessment. This is a crucial point, as it determines if and what kind of guidance would be helpful. If the challenge is in obtaining the data required, guidance will not help. If the challenge is to identify what kind of data are required, technical guidance might help if it is specific enough (in many cases capacity building and/or bi-/multilateral assistance might be more effective). A guidance on the risk assessment methodology will help only if the challenges are indeed identified in this area.
The study identifies four features which, in effect, distinguish organisms with engineered gene drives from LMOs assessed so far (p. 40, l. 42-47):
- the modified inheritance pattern (p. 40, l. 44)
I would agree on this being the identifying feature of gene drives.
- Targeting non-managed environments (p. 40, l. 45)
- Targeting non-domesticated species (p. 40, l. 46)
To these two features I can only agree in so far as the focus lies on the “targeting”. LMO without gene drive can spread into non-managed environments or be of non-domesticated species. For this reason, existing risk assessment frameworks are not tailored or limited to domesticated species and managed environments. If gaps with respect to these features would be identified, more LMOs than just those containing engineered gene drives would be affected.
- Threat of an irreversible impact at a scale exceeding the intended release (p. 40, l. 47)
This is in fact a potential feature that was assigned to all living organisms since the start of the release of LMO. There are, without doubt, significant differences with respect to the likelihood of such impacts. This likelihood depends on all features of a LMO, including if the inserted gene has drive potential or not. This problem is best covered by a case by case approach which is the basis of all sound risk assessments.
In summary, the study identified only one feature specific to LMO with gene drives differentiating them from other LMOs: the modified inheritance pattern.
Further on the study states that many contributors indicated that the principal framework for risk assessment can be used to assess LMO with gene drives, but questions will need to be tailored to what is distinctive on a case by case basis (p.31 l.17-23). Again four points were identified.
- some of the assessment principles such as the comparative approach may not be fit for purpose in this context; (p.31 l.24-25)
I find this consideration hard to follow. In a comparative approach I would agree with Hector Quemada „Whether a domesticated or wild species, the comparator would continue to include the non-modified organism“ and would extend this to whether a gene drive or no gene drive is included.
- the “stepwise” approach may not be applicable….(p.31 l-26ff)
This is probably the central problem, at least for some low threshold gene drives. The central idea of the stepwise approach is to obtain robust data to inform a sound risk assessment before full scale release. If such robust data could be obtained by other means it may still be possible to finalise a risk assessment. Depending on the specific case there may be different approaches to realise a stepwise approach on a temporal, special or molecular genetic level. It is advisable to gather information on this question in order to reach a common understanding how to proceed.
- Using robust models….(p.31 l.30-31)
Based on the available literature I would agree that robust models will be a valuable tool to inform a risk assessment of LMO with gene drive. Such models will improve the understanding of potential effects on a population and ecosystem level. Even more important, such models will improve the understanding of what kind and quality of data are required to feed a model to obtain robust results.
- Concepts like the “receiving environment” must be revisited… (p. 31 l. 32ff)
Again I have problems to follow this consideration. For a sound risk assessment of the release of LMO, knowledge on the receiving environment is essential. This is even more true for LMO containing a gene drive, as the functioning of the drive effect is directly depending on the receiving population. Not the concept of the “receiving environment” needs to be revisited, but rather the amount and type of data required on the receiving environment.
Summing up, I come to the conclusion that there are certainly challenges for a risk assessment of LMO containing gene drives. However, these challenges are mostly of a technical nature, regarding the type of data required to finalise a risk assessment. I do not see challenges that cannot be approached within the frameworks and methodologies available.

Annex I (d) The challenges in addressing the specific issue are clearly described.
With respect to this criterion the draft study (p.31 l.38ff) refers to points identified in the previous section. The consideration that the only way to clearly describe specific issues is in relation to specific cases is not really helpful. This would imply that no “specific issue” in the meaning of CBD/CP/MOP/DEC/9/13, i.e. group of LMO, would ever be able to satisfy this criterion as a clear description of the challenges would only be possible case by case.
The draft study finds (p.31 l.43-44) that “Most -if not all- of the considerations that were identified are not related to engineered gene drives per se“. As laid out above I tend to agree and would add that those challenges that are prominent for engineered gene drives are rather of a technical nature than challenges to the principal framework.
It is not expedient to develop guidance specifically on the risk assessment of LMO containing gene drives to tackle challenges that can also be found in LMO without gene drive, for example wild species or LMO in unmanaged ecosystems.

Werner

Dear All

Having taken note of several further comments posted on this interesting online forum, I think it is important to point out a specific weakness in the draft report on GE gene drive organisms.

As mentioned by several experts, there are some crucial aspects missing in the draft report, such as the perspective of the precautionary principle or, more specifically, the interactions of GE gene drive organisms with the environment and potential adverse effects. However, it is also important to be aware of somewhat misleading comparisons in the report, e.g. referring to methods of biocontrol such as the Wolbachia infection in mosquitos.

It should be understood that the risks emerging from the Wolbachia system can hardly be compared to those of GE gene drive organisms: genetically engineered gene drives typically require dozens of generations to establish the desired effect in the target populations. This requires the integration of additional gene constructs in the genome, which will then undergo further processes such as crossings and repeated cuttings of the DNA. These processes can cause a broad range of specific and unintended next generation effects. However, there is no such process or technology involved in the Wolbachia system or with techniques which are used in SIT.

Thus, if any comparison is made, the differences and their consequences for environmental risk assessment should be emphasized. Whatever the case, the specific risks associated with the application of GE gene drive organisms need to be identified and assessed properly. This is not achieved in the published draft report.

We all are aware that the next generations effects that go along with the introduction of GE gene drives can not be assessed the same way as risks of domesticated crop plants which are meant to be cultivated just for one year in the fields. Therefore, the group should deal in depth with these new challenges which in the end definitely will need new methodology and new criteria for risk assessment and risk management.

Thanks for your interest and comments

Christoph Then

I am Nada Babiker Hamza, a Biotechnology and Biosafety researcher and involved in the risk assessment of GMO and heading the National Technical Committee for Biosafety (Sudan).
Thanks to the forum moderator: Marianela Araya, and the Secretariat for hosting the forum. I am grateful to have the opportunity to participate and interact with colleagues from several countries on this new and important subject to our region.
I would like to Thank G. Smets & P. Rüdelsheim for providing the study on the state of knowledge on genedrive provided with literature and results of interviews.  The study is intended to inform the application of Annex I of decision CP-9/13 to LMOs containing engineered gene drives for the process for recommending and prioritizing specific issues of risk assessment for consideration under the Cartagena Protocol on Biosafety mentioned on paragraph 6 of that decision.
However, in their interventions, Dr. Jack Heinemann [#10117] and Ms. Rebecca Carballar-Lejarazu [#10119] pointed out that the study narrowed the definition of gene drives and their applications, which was reflected in the conclusions.
In order to reduce scientific uncertainties and understand concerns on the effects of gene drives (mentioned in 4.2 of the document), and with consideration to Article 1 of the Cartagena Protocol. I believe that there are regulatory challenges, and to assess the engineered organisms and products before they are introduced into the environment. A serious concern on how these effects will perform directly in the field and how to assess unintended and off target effects. As there have been some evidences that the emerging genetic engineering has unintended effects. Therefore, more R & D and capacity building are needed to tackle the technical and methodological challenges of risk monitoring, assessment and precautionary measures, considering case by case.
I agree with what was mentioned by James and Tountas (2018) “it should be emphasized that continued research is the only way to decrease the uncertainties that underlie the perception of risk”. Due to the possibility of adverse / irreversible impact on a species it can change entire species, populations at global level and consequently biodiversity and ecosystem services.
As a regulator, I agree with Devos et al. (2019) mentioned in the study, who stressed that regulators and governments have to clearly identify their environmental policy goals (protection priorities) as a prerequisite for the environmental risk assessment (ERA) to address them (problem formulation). From previous experiences, I think Laws and regulations should be done from Bottom up including public consultation, socioeconomic and ethical considerations.
From the specific issues of risk assessment for consideration by the mechanisms for transboundary considerations for applications of engineered gene drives that can span several countries as describes in Article 4 of the Cartagena Protocol.

Dear Marja, dear all,

This has been an interesting discussion and reflection so far and I would like to contribute with a few points for now, but will send in comments to the draft text as a separate doc later.

As others pointed out – e.g. Jack Heinemann and Yann Devos – it would be important to be clear on the scope as well as on definitions, even if we decide on using working definitions, as the NASEM report did.

I very much appreciate Jack reminding us that gene drives -seen as genetic mechanisms that distort the segregation of traits- are found widely in nature and are not limited to sexual reproduction. We need to be aware though, that the term “gene drives” is not a term that has been commonly used for most of these genetic mechanisms that distort segregation of traits or sequences, and it is only recently that they are being increasingly captured under the overarching category of gene drives. Of course this highlights the importance of us needing to be clear what exactly it is we are focusing on – and for that we need to look at the whole spectrum first, which unfortunately is not covered presently in the draft report, which I hope will be amended.

The suggestion to focus on “most likely cases of gene drive modified organisms moving to practical application” (Yann Davos [#10127]) could possibly be a good option for a later stage, for example when looking at developing very specific guidance for RA&RM – if so decided. First of all though a comprehensive overview of what is being considered and/or developed (engineered gene drives and LMOs with engineered gene drives) needs to be provided by the report, which will then need to be studied and discussed, including among other how to judge what is moving to practical application.

I find it is also important to not instantly think of case by case risk assessment, as that takes away from opening up to and understanding the bigger questions, challenges, and risks as well as the knowledge requirements that come or may come along with LMOs containing engineered gene drives. It is at this level that I find we need to deliberate and investigate. What are the knowledge requirements, questions and concerns that will need to be addressed, almost irrespective of which GDO (gene drive organisms) we are talking about.

Thus, if the scope is narrowed down too early it would not allow for the development of a broad-sweep catalogue of questions, concerns and inquiry. I would for example like to see a whole range of questions linked to the ecological role of an organisms, irrespective if it might be considered a key species or not. A set of questions on what the consequences might be if the organisms is modified (intentionally, or un-intentionally by for example off-target effects), replaced by gene drive organisms, or replaced by other organisms filling the niche, the target organism being locally eradicated and then returning, etc. In fact, I am talking about a detailed list of scenarios that (would) need to be investigated in order to allow for the possibility of carrying out a meaningful risk assessment. I will provide further comments on this separately.

In short, from my perspective, this report needs to be all-embracing.

In order to deliberate – whether on scope, on what guidance might be required, on what hazards might arise and what the risks are – a solid basis of information is required. I would expect and would want the (draft) report to be the place for such necessary information. If such information is not available because it does not exist, then this -also- should be noted.

In this context I want to point to three crucial aspects I find missing or lacking in the current draft report:

1) The report should include a detailed listing of all the LMOs with engineered gene drives currently being developed (including the gene drive systems employed), and species where feasibility is being explored with preliminary work. It is crucial to have such an overview so as to be able to understand the vastness of the field and to determine the challenges for risk assessment. Table 5 (page 22) is a helpful step in this direction, however much important information is not included. The table does not breakdown work into individual applications/projects and does not appear to include:
• Organisms where funding has been allocated for development of gene drives (including preliminary work)
• Organisms where preliminary work has been published (e.g. developing molecular genetics tools, genome sequencing, or identification of target genes) with the intention of developing gene drives.
For example funding has been allocated by DARPA for the development of gene drives in the mosquito species Culex quinquefasciatus and preliminary work has now been published describing molecular genetic tools in this species, see Anderson et al, PloS One (2019)  14 (11) e0224857.

As an example, please see tables 2a-c in Chapter 2 of our report, which includes additional species where funding has been allocated or preliminary research has been  undertaken (Steinbrecher et al. 2019, pp. 74-87, https://genedrives.ch/wp-content/uploads/2019/10/Gene-Drive-Kapitel-2-WEB.pdf). These tables are up to date to the end of March 2019. Further preliminary research has been published in the remainder of 2019, including in:
• Diamondback moth (Plutella xylostella) - Harvey-Samuel et al., Insect Mol Biol (2019) doi: 10.1111/imb.12628
• New World screw-worm fly (Cochliomyia hominivorax) and Australian sheep blow-fly (Lucilia cuprina) –Paulo et al., G3 (Bethesda) (2019), 9 (9) 3045-3055
• Tephritid fruit flies (Anastrepha sp., Bactrocera sp. and Ceratitis sp.) – Sim et al., Insect Mol Biol (2019) 28(2) 277-289
• Schistosoma haematobium -  Stroehlein et al., Gigascience (2019) 8(9) giz108
• Oriental fruit fly (Bactrocera dorsali) – Zhao et al., Pest Manag Sci (2019), 75(7) 1921-1932  (already published beginning of 2019)


2) A detailed overview of gene drive mechanisms, especially those that are being used, developed or considered for creating LMOs with engineered gene drives.  This would be to understand the landscape of what current possibilities are and where developments might be heading. This would also help for one aspect of hazard identification as well as with the identification of lack of knowledge and/or data, and the challenges that are being posed. Whilst Annex 3 gives further information for some of them, it is not sufficient to understand the limitations, shortcomings or hazards associated with them, nor their applicability, strength or predictability.

Separately it would be useful to elucidate their origin and how they function within current populations and ecosystems. It was suggested in the forum to use knowledge of such natural gene drive systems as baseline information for risk assessment engineered gene drives. I would like to caution against this approach, as not only will different species deal differently with the same type of gene drive system (leave alone with different gene drive systems), but we have extremely little knowledge about the co-evolutionary forces and actions that have shaped the development of such gene drives and the equilibrium found with natural gene drives. I agree we should learn all there is to know about natural gene drives, but not take them as baselines.

3) The listing of the considerations on risk assessment (as given in the 5 bullet-points page 3-5) are insufficient. For example the issue of evolution (and not just evolution of resistance to HEG gene drives) is not given any mention. Nor is the explicit recognition of unknown unknowns and how to deal with them. I will not elaborate here but come back to this later in a separate input.

Sorry for this rather lengthy intervention. I promise to be shorter next time!

With kind regards, Ricarda

I was part of the Biosafety Committee of Costa Rica, in charge of the risk assessment done by the National Academy of Science. We used the Problem Formulation approach and several guidelines including the one developed by the Secretariat. In regards to the decision 9.13 (CBD/CP/MOP/DEC/9/13), and the study provided by the Secretariat, and specifically, to the question of Do we need more guidance for gene drives?, We would like to share the following.
We already have Risk assessment guidance and what is needed is training, and experience sharing
We would like to note that there is enough guidance to conduct a risk assessment of LMOS containing gene drives, as described next,
General Risk assessment guidance
EFSA Guidance  http://www.efsa.europa.eu/en/applications/gmo/regulationsandguidance
EPA Guidance https://www.epa.gov/sites/production/files/2014-11/documents/eco_risk_assessment1998.pdf
UNEP Guidance UNEP/CBD/BS/COP-MOP/8/8/Add.1
UNEP/CBD (2012). Guidance on Risk Assessment of Living Modified Organisms. Available at: https://bch.cbd.int/protocol/guidance_risk_assessment/.

Specific gene drive guidance
WHO (2014) “The Guidance Framework for Testing Genetically Modified Mosquitoes.” WHO. Document Production Services (DUP), Geneva, Switzerland.

OGTR (Office of the Gene Technology Regulator) (2019). Guidance for IBCs: Regulatory requirements for contained research with GMOs containing engineered gene drives. Available at: http://www.ogtr.gov.au/internet/ogtr/publishing.nsf/Content/ibc-1.

ZKBS (Zentrale Kommission für die Biologische Sicherheit) (2016). Position statement of the ZKBS on the classification of genetic engineering operations for the production and use of higher organisms using recombinant gene drive systems. Available at: https://bch.cbd.int/database/record.shtml?documentid=110745.

Westra, J., van der Vlugt, C. J. B., Roesink, C. H., Hogervorst, P. A. M., and Glandorf, D. C. M. (2016). Gene Drives Policy Report. RIVM Letter report 2016-0023 (National Institute for Public Health and the Environment, Netherlands). Available at:  https://www.rivm.nl/bibliotheek/rapporten/2016-0023.pdf.


We are aware that a Risk assessment requires experience. We learned, at the Biosafety Committee, from regulators around the world that the methodology is flexible enough to make a case by case evaluation. Meaning that you can even evaluate different stressors to the Environment. We are also aware that the analysis could result in information missing. The missing information can be critical or not, depending on how important it is to characterize the risk or the exposition, and sometimes it depends on risk management practices as well.

In conclusion, we would like to stress out that LMO with gene drives can be subjected to the guides that already exist. Consequently, instead of new guidance, what we need is training and to learn from the experiences of other regulators.

Dear Moderators and colleagues, thank you very much for this discussion on a critical subject. My name is Dr Eva Sirinathsinghji, a biosafety researcher representing Third World Network.

I raise some issues and comments on the draft report:

1. Distinguishing features of GDOs raise fundamental challenges for RA:

While the report touches on the distinguishing features of meiotic gene drive organisms (GDOs), there are critical gaps in how they raise fundamental challenges for risk assessment.

a) Inclusion of genetic engineering machinery in GDOs: As raised in the draft report (which cited Simon et al., 2018), the inclusion of the genetic engineering machinery into GDOs means that essentially, the technology transfers the ‘laboratory to the field’. In order to spread, the genetic engineering process occurs at every generation for perpetuity. While off-target and unintended effects at the molecular level may potentially be assessed with current GMOs before an environmental release, this cannot be conducted on a GDO as there is no ‘final’ product, but instead the release of a GM tool that will continue to function in wild populations with potential next generation effects. This is further complicated by the fact that off-target effects are for example 1. unpredictable; 2. dependent on genetic background (impossible to control in genetically diverse, wild populations); 3. affected by cellular, environmental conditions. While the report acknowledges that robust modelling will be required to support risk assessment, it fails to recognise that such knowledge is arguably unattainable for a system that continues to perform genetic engineering once released. How such complex genetic engineering processes occurring continuously over time and space will affect issues such as outcrossing potential, molecular aspects e.g. genome stability, transgene stability/efficacy, or wider ecological impacts e.g. ecosystem function, species interactions, will not have been considered to the same extent with current GMOs. Unexpected interactions with changing environmental conditions and genetic backgrounds makes predictive modelling very difficult if not impossible to reassure against risk.  

b) Spread and persistence: The issues with GDO spread and persistence raise critical RA challenges regarding the phased-approach as well as contained-use experiments.  The report fails to acknowledge that any deployment of a gene drive as part of a ‘phased-approach’ is effectively an open release, even if performed on island locations. As stated by the Ad Hoc Technical Expert Group on Synthetic Biology: ‘Islands are not ecologically fully contained environments and should not be regarded as fulfilling the conditions in the definition of contained use as per Article 3 of the Cartagena Protocol unless it is so demonstrated’ (AHTEG, 2017: Para 51 (b)).  Indeed, island locations have been acknowledged by developers to not be fully confined (James et al., 2018), and would need to be accessible to researchers conducting the experiments (Harvey- Samuel et al., 2019), thus providing a route of escape. Despite this, islands such as the Ssese Islands, Uganda have been proposed for future GDO releases. Localised or self-limiting strategies are also largely theoretical, and are potentially capable of evolving into global drives or have other unintended effects e.g. mutation in gRNA that is designed to target only a locally fixed allele, or recombination of a split-drive that results in global gene drive conversion. R&D lags behind that of global or low-threshold/self-limiting gene drives, and cannot currently be relied upon to control the spread of GDOs.

I would thus suggest that the report be corrected to clarify that isolated locations such as islands, or biological containment do not ‘constitute a form of biological or molecular confinement’.

Another fundamental issue with GDO mosquitoes even if tested under a phased approach, is that they primarily feed off humans who will also not be confined to the trial site. This raises numerous issues regarding both containment and health risks. Any trial will require that people are exposed to mosquitoes (wild potentially disease-carrying and/or GDO mosquitoes). This is in contrast to GM crops where people are more easily protected from exposure to confined field trials. According to the World Medical Association’s Declaration of Helsinki which is based on the Nuremberg Code and outlines the internationally agreed ethical principles for medical research involving human subjects: “While the primary purpose of medical research is to generate new knowledge, this goal can never take precedence over the rights and interests of individual research subjects.” (Article 8). It also states such research “may only be conducted if the importance of the objective outweighs the risks and burdens to the research subjects” (Article 16).

With regards to contained use, Article 6 of the Cartagena Protocol preserves the right of Parties to risk assess subject LMOs to risk assessment prior to decision on import, and set standards for contained use within its jurisdiction. Numerous characteristics of global GDOs can be thought to resemble those of pathogens, such as ability to spread, the spread of lethal traits (in population suppression applications), and difficulty (even impossibility) of removing them from the environment once released. Indeed, modelling has shown that the release of as few as two individual GDOs could be sufficient to lead to establishment in the environment (Noble et al., 2017). Considering the accidental releases of pathogens out of high security labs in recent years, gaps in the draft report on this issue should be addressed.

The report also suggests that resistance development in a GDO would ‘limit potential impact’.  This assumption fails to take into consideration the unpredictable outcomes that may arise with resistance development e.g. generation of a mutation that may increase fitness and thus frequency in the population. Potential of resistance development to limit spread and establishment of a GDO following accidental escape from contained use or environmental releases cannot be regarded as a safety mechanism.


2. Precautionary Principle:

Unfortunately, the application of the precautionary principle has not been consistently applied throughout the document. While the draft report acknowledges the need for a precautionary approach, as raised in the recent report by ENSSER (previously referenced by Dr Christoph Then [#10135]), it is being largely addressed in relation to technical aspects of GDO assessment only e.g. phased-approach to testing, but not with regards to the wider ethical, political and philosophical dimensions that can be applied prior to the research phase. I thus echo comments raised by Abdullah Mkindi [#10142] that ethical issues should be thoroughly taken into account. As also raised in other posts, implying that the R&D ‘stepwise’ approach starts with experiments, fails to apply the principle to obtaining full, prior, informed consent from affected communities including indigenous people and local communities, before commencing the experimental stages. Echoing Heinemann’s submitted comments, only by obtaining such consent can issues of scientific uncertainty be addressed. As recently covered by the Ad Hoc Technical Expert Group on Synthetic Biology (AHTEG 2019, Para: 33): ‘the state of knowledge on potential impacts of current and near future applications of synthetic biology should consider that, for indigenous peoples and local communities, those applications that may impact their traditional knowledge, innovation, practices, livelihoods and use of land, resources and water should seek their free, prior and informed consent, and the assessment of those applications is usually undertaken in a participatory manner involving the whole community.’  The report further acknowledges the need to look beyond one scientific perspective to be able to address all potential impacts (AHTEG 2019, Para: 11): ‘It was noted that indigenous peoples and local communities could have different perspectives, different ways of perceiving potential impacts and be impacted differently by synthetic biology applications in relation to the objectives of the Convention, since, for indigenous peoples and local communities, natural elements are living entities’.

3. Comparisons to Wolbachia:

Finally, I would also like to echo comments made by Christoph Then [#10159] on the comparison of Wolbachia or SIT to GDOs. There is limited comparison of such techniques to GDOs and the differences should be emphasised in the draft report.  


Many thanks
Eva

Dear colleagues,
thank you for your interesting interventions and special thanks to Marja for pointing our focus back to annex I of CP-9/13 and the tasks at hand.
Having focused on the third bullet point Marja introduced [#10114] I would like to give some input on the second bullet point “- could identify any information gaps or factual errors” in this contribution.
I think the study presents a very good overview of the specific issue of gene drives, especially with respect to annex I. I would like to highlight just a few points I deem worth noticing:
- The possibly broad diversity of potential gene drive mechanisms is very well reflected in the study. It is important to keep this in mind in order to be able to identify and focus on the common characteristics potentially setting engineered gene drive organisms apart from other LMO.
- At several points the importance of differentiating between challenges to the risk assessment methodology and challenges relating to obtaining information required to inform the risk assessment is highlighted.
- At several points it is clearly stated that some challenges identified are not specific to LMOs with engineered gene drives but might also be found in other LMO. If these challenges could only be met by supplementary guidance for LMOs with engineered gene drives, it would consequently imply that other LMO facing the same challenges are not covered. At many points the differentiation is rather between non-domesticated versus domesticated LMO than LMO with or without gene drive. If this was the case, guidance on LMO with gene drive will not meet the requirements.
I would also like to second some points already identified by colleagues where the draft study might be improved:
- I agree with Yann [#10127] that the sources of evidence used to substantiate statements should be stated more clearly.
- While Annex I of the study gives a reasonably good presentation of methodology of the literature search performed, I am missing a clear presentation of the results obtained. I agree with Yann [#10127], that the section on literature search might be expanded.
- I agree with Hector [#10129], that the statement presented in 4.1.2 should not be left undisputed.
Engineered gene drives can clearly be seen a finished product. One of the known properties of this product is a changed segregation of the introduced construct. These LMO can be selected during R&D, tested and presented as a well-characterized product. The offspring will obtain the same construct in the same location as the parent. This is what defines an event in most regulations. Authorised events are passed on to offspring and this process is assessed during risk assessment and generally included in the authorisation. The only difference with gene drive constructs is that they are passed on to more than 50% of the offspring. Therefore, describing gene drives as performing genetic modification over and over again does not properly reflect what is actually happening.

Dear Marja, Dear All,
My name is Marion Dolezel and I work as a biosafety expert at the Environment Agency Austria, being involved with risk assessment and monitoring of LMOs for more than 15 years.

I would like to thank Marja for moderating and guiding this forum and take the opportunity to complement some points that have been raised by others as well as add some ideas to this valuable and lively discussion.

1) Considerations addressed in the report refer to the resistance development as a means for confinement of GDOs (page 4, line 28-29), a point that has also been taken up by intervention #10166. I would like to stress that the durability of a specific GD application is a major factor relevant for the success of the specific GD application (e.g. the suppression of a specific human-pathogen vector population). Resistance to the GD mechanism in the target population will compromise the original aim of the control strategy in the short or long term. This loss of efficacy can entail considerable risks to human health and the environment, e.g. if conventional control strategies have ceased in the meantime. Therefore a key question in this regard is how long a gene drive has to remain functional in order to achieve the goal of the intervention and if this can be reliably predicted before the release into the environment takes place. Any predictions of resistance development to gene drives are currently based on theoretical modelling or laboratory experiments considering molecular resistance, but excluding gene-environment interactions as well as the potential for behavioural resistance to the drive. Any control strategy which is based on resistance development as a confinement method is subject to high uncertainty and unpredictability, if not completely left to chance.

2) Risk assessment of GDOs must be based on a comprehensive empirical data basis regarding the biological and ecological specificities of the respective organism in its natural habitat and must use an appropriate spatial and temporal reference framework in order to assess potential adverse effects of the individual application. Up to now there is no empirical data or practical knowledge regarding potential effects from releases of GM animals in EU environments (see also intervention #10137), specifically those taxa which are currently relevant for GD approaches. As proposed by others (see e.g. #10137, #10147, #10127), the use of experience from the release of insects for biocontrol purposes might not be a panacea, as these are often non-native organisms released into “unfamiliar” environments with a lack of long-term evaluations regarding their effects on non-target organisms or ecosystems. I agree that there are some general lessons that can be learnt from other biological control methods (see #10149, Louda et al. 2003), however, when addressing the long-term and evolutionary consequences of GDOs their unique molecular and biological features must be taken into account.

3) I would also like to take up the note on the importance of the problem formulation process and the identification of protection goals (see #10127) as the starting point for the ERA process. In this context I would like to point out that decisions on risks can only be made if clear decision criteria for the acceptability of the specific risks for specific protection goals are available before GDOs are applied in the environment (see also intervention #10145).  Such criteria will have to be formulated by different stakeholders and not only the scientific community which has the advantage of integrating societal values and perceptions in the decision making process as to whether risks of a specific pest control strategy are acceptable or not.

4) In environmental risk assessment (ERA), long-term effects of GDOs on whole populations or ecosystems including potential evolutionary changes in target and non-target populations are difficult to assess. However, the assessment of long-term effects of a specific LMO is a requirement in the ERA in the EU. Often, modelling is recommended to address long-term risks. For gene drive organisms, theoretical modelling exercises address the spread and functionality of specific gene drive constructs (e.g. Marshall & Hay 2012, Akbari et al. 2013, Noble et al. 2017, 2018) while potential ecological implications have been discussed on a theoretical level only (David et al. 2013, Simon et al. 2018). Hence, appropriate risk assessment methodologies still need to be developed to address also long-term ecological effects of GDOs.

best wishes,
Marion

Thank you Marja and the Secretariat for hosting this interesting discussion. I would also like to thank the writers of the Perseus report for the comprehensive compilation of data. The report is a good basis for further discussion.

The current coverage of gene drives in the study is adequate (and other types of gene drives are presented in the Annex 3). However, the report should consistently use the term “engineered gene drive” or “synthetic gene drive” to indicate that that the scope is limited to these.

In my opinion, organisms with engineered gene drives fulfill most, if not all the criteria in the Annex I of the decision 9/13. I think that organisms with engineered gene drives challenge current environmental risk assessment and risk management of LMOs and have the potential to cause adverse effects on biodiversity (lack of comparators, challenges in stepwise approach, irreversibility, effects on genetic diversity, etc.; see e.g. Simon et al. 2018.). I do agree with Dr. Dolezel (10168) that appropriate risk assessment methodologies need to be developed to address long-term ecological effects of GDOs.

Katileena Lohtander-Buckbee, senior adviser, Finnish Environment Institute

Dear All,

My name is Piet van der Meer, my scientific educational background is in ecology and microbe-plant interactions (N2fixation), and I have more than 30 years of experience in risk assessment in the context if regulatory decision-making on GMOs.

My thanks for Marja for moderating this debate and for her call to focus the debate on the tasks given.

In response Marja’s request to share information that could complement the draft study, that could help identify information gaps or factual errors, and/or that is relevant to one or more of the aspects of annex I of decision CP-9/13:
1) I agree with Tzvetkov [#10139], Kingiri [#10147] and others that the draft study provides a good basis for the discussion by presenting an overview of the field of gene drives and related risk assessment questions. As Schenkel [#10155] and others stated: an important conclusion that study draws from the interviews is that environmental risk assessment of organisms with gene drives can be conducted, on a case by case basis, through the existing methodology for ERA (see Annex III CPB).
2) That said, I also agree with Glandorf, Devos and others that the report can be improved on a number of points. For brevity, I refer to interventions in contributions (# 10117, # 10119, # 10122, # 10127, # 10128, # 10129, #10134, # 10137 and #10139.

Following up on some of the points raised:
a) As regards the scope of the study: Heinemann and others have correctly noted that gene drives are found widely in nature and with very different mechanisms. Yet, while it would be good to reflect that in the beginning of the study, I agree with Glandorf and others that for practical reasons it would be good to focus the report on the meiotic gene drives that are currently under development.
b) I agree with Devos [#10127] and others that in framing risk assessment of organisms with gene drives we can draw on familiarity/experience with existing insect vector/pest control strategies, and with organisms with gene drives that are found in nature. The observation in [#10159] that such comparison is misleading seems to ignore how certain data can provide relevant insights even if the underlying mechanisms are not the same. The extent to which other experiences can provide relevant insights will depend on the case at hand. As Dolezel [#10168] rightly noted, while some lessons can be learnt from other biological control methods, the use of experience from the release of insects for biocontrol purposes is not a panacea.
c) As regards the observations with reference to the precautionary approach: the precautionary approach contained in Principle 15 of the Rio Declaration (as reaffirmed in the Preamble and article 1 of the Cartagena Protocol) has a role in the context of decision making, and should not be confused with risk assessment.

3) In the context of decision making, there is statement in the study that would benefit from some further clarification. In the paragraph in lines 24 – 26, the study notes “some applications are close to being released in trials”. It will be helpful to clarify that in the context of the Cartagena Protocol on Biosafety, under which this study and on-line forum take place, organisms with gene drives fall under the definition of LMOs and that the AIA procedure requires authorisation for releases that fall under that procedure. In fact, also under most, if not all, domestic regulatory frameworks for biosafety is it prohibited to release LMOs with gene drives without prior authorisation.

Looking forward to the remainder of the debate,

Piet

Dear All,
Thanks again to the moderator and participants for this important discussion.

Taking note of some of the comments on the forum, I would like to highlight issues with the report referring to eradication of invasive species.

I would suggest that there are errors within the draft report, with a failure to recognise that there are also risks to biodiversity associated with removing invasive species. Any suggestions that removing invasive species, by default, would be beneficial to ecosystems and biodiversity ignores previous experiences where in extreme cases, removal of an invasive species has resulted in serious ecosystem degradation that required expensive reparatory actions (see review by Kopf et al., 2017). Invasive species can have unexpected function roles in food webs, provide habitat or support ecological processes. Direct and indirect effects on the whole system should therefore also be carefully taken into account. Such effects go beyond those raised in the report of spread to native habitats, or lack of complete restoration of ecosystems. The above cited review emphasises that large-scale efforts for removal of invasive species could hasten biodiversity loss if risks are ignored, particularly with control methods such as gene drives where effects could be irreversible, diminishing capacity to deal with unintended effects.

The statement that 'removal of invasive species may be welcomed' (Section 4.6.1) in the report should be corrected to take into consideration such risks.

Many thanks
Eva

Dear all,

We thank the Secretariat for this opportunity to share Honduras inputs. We would like to frame our participation in the Context of decision 9.13 (CBD/CP/MOP/DEC/9/13) and the documents provided by the Secretariat, specifically to find out if we need more guidance.
6. Decides to establish a process for the identification and prioritization of specific issues regarding risk assessment of living modified organisms for consideration by the Conference of the Parties serving as the meeting of the Parties to the Cartagena Protocol with a view to developing further guidance on risk assessment on the specific issues identified, taking into account annex I;
12. Requests the Subsidiary Body on Scientific, Technical and Technological Advice to make a recommendation as to whether additional guidance materials on risk assessment are needed for (a) living modified organisms containing engineered gene drives, and (b) living modified fish for consideration by the Conference of the Parties serving as the meeting of the Parties to the Cartagena Protocol at its tenth meeting.
1. First we would like to support Dr Schenkel’s post in general, and specially in regard to the information needed instead of new guidance; and that case specific issues that can be handle by risk management. We would also support the post of Dr. Hector Quemada Western Mich. Univ to keep the technical science-based support of the topic using a comparative approach.
2. Second, given the documents and concepts we would like to remark that we need training and experience sharing instead of new guidance, and we will respond the three questions given by the Secretariat,
(1) Could complement the study, which may include: further development of concepts, explanatory comments, relevant resources, bibliographic references, among others; (2) Could identify any information gaps or factual errors
(3) Is relevant to one or more of the aspects of annex I of decision CP-9/13
As follows,
A. Risk assessment principles are the same.
The Report of Perseus states that risk assessment methodology is largely based on problem formulation methodology, providing a structured and systematic approach for addressing risk assessment, subsequent risk management and provides a tool for risk communication (Smets & Rüdelsheim 2019).
Environmental Risk Assessment principles have been used since the 80s and have been shown to be sufficiently robust to provide the appropriate information for regulatory decision making, in order to ensure an adequate level of environmental protection. The origins of the principles are stated in documents like the “Red Book” (National Research Council, 1983), and have been implemented to make objective science-based decisions to collect, analyze and communicate information (National Research Council, 1996) not just for GMOs but for different stressors such as chemicals, land change, disease, invasive species and climate change (Raybould, 2007).  The steps are always the same (step 1) problem formulation, a critical first step, including hazard identification; (step 2) hazard characterisation that examines potential hazards and their magnitude; (step 3) exposure characterisation that estimates levels and likelihood of exposure; and (step 4) integrative risk characterisation in which the magnitude of consequences and the likelihood of occurrence are integrated (EFSA, 2010; https://www.epa.gov/risk/ecological-risk-assessment).
B. We have training available, but countries note that experience sharing is required instead of new Guides
So far the training provided internationally and online show that the principles are well known for example in http://bch.cbd.int/cpb_art15/training/module3.shtml, ILSI https://ilsirf.org/elearning-courses/pfera/ , ICGEB https://showcase-icgeb.elearning.it
In the report the responses of the experienced countries versus the non-experienced countries denotes the need for experience sharing. That experience sharing instead of new guidance was also stated by parties like Honduras, Costa Rica, Canada, Japan, Australia, New Zealand and non-parties like USA in the intersession period 2017-2018  https://bch.cbd.int/protocol/cpb_art15_submissions/ . The need of training and experience sharing in how to perform a risk assessment instead of new guidance, and the application of Risk assessment principles to address a case by case analysis was also stated by Ethiopia, Nigeria, Australia, South Africa, Japan  and non-parties like USA in the 2019 submission of information https://bch.cbd.int/onlineconferences/submissions.shtml
C. The challenge is not the risk assessment methodology.
The report also states that interviewees differentiated between challenges to the risk assessment methodology and challenges relating to obtaining information required to inform the risk assessment. Most interviewees anticipated that it will be possible to use existing risk assessment methodology for evaluating LMOs with engineered gene drives. The report actually shares two examples on how to integrate it: NASEM (2016), and Kuzma (2019).
In regard to obtaining information, it is important to note that Identifying missing information is actually part of the Problem formulation process in order to have the risk assessment done. Sometimes it is required to have that information to dilucidated the exposition, or better characterize the hazard, but in both cases, the missing information is case-specific and is already part of the risk assessment methodology.
D. The issues are speculative given that we do not have real cases.
All interviewees pointed out that no case of an actual release of an LMO with engineered gene drives has been assessed, leaving the definition of issues speculative.
In this specific case, when the countries are looking to develop “guidance”, the fact of having limited information makes the process unreliable. Part of the hazard and exposure characterization requires gathering relevant information such as Ecotox studies, previous risk assessments conducted in other countries.
E. We already have Risk assessment guidance
We would like to remark what is noted in point 5.2 of the study, there is already enough guidance to conducts LMO risk assessments. That is key to the Annex I of decision CP-9/13

In conclusion, we would like to remark that we need training and experience sharing instead of new guidance. The result of the risk assessment can be the identification of missing information, but that is case-specific.

Best Regards
Roger

Bibliography.

EFSA Panel on Genetically Modified Organisms (GMO).(2010).Scientific Opinion on the assessment of potential impacts of genetically modified plants on non-target organisms. EFSA Journal 2010; 8(11): 1877. [72 pp.] doi:10.2903/j.efsa.2010.1877. Available online: http://www.efsa.europa.eu/efsajournal.htm

Kuzma, J. (2019) Procedurally Robust Risk Assessment Framework for Novel Genetically Engineered Organisms and Gene Drives. Regulation & Governance. https://doi.org/10.1111/rego.12245.

National Research Council.(1983). Risk Assessment in the Federal Government: Managing the Process. National Academies Press, Washing- ton, DC

National Research Council. (1996). Understanding Risk: Informing Decisions in a Democratic Society. National Academies Press, Washington, DC

NASEM (National Academies of Sciences Engineering and Medicine). (2016). Gene Drives on the Horizon: Advancing Science, Navigating Uncertainty, and Aligning Research with Public Values. The National Academies Press, Washington (DC), 215pp.

Raybould, A. (2006). Problem formulation and hypothesis testing for environmental risk assessments of genetically modified crops. Environmental Biosafety Research, 5(3), 119-125.

Smets G & Rüdelsheim P. (2019). Perseus Report, Study on Risk Assessment Application of annex I of decision CP 9/13 to living modified organisms containing engineered gene drives. Secretariat of the Convention on Biological Diversity

US EPA. (1998). Guidelines for ecological risk assessment. U.S. Environmental Protection Agency, Risk Assessment Forum, Washington, DC. Available at https://www.epa.gov/sites/production/files/2014-11/documents/eco_risk_assessment1998.pdf

Wolt, J. D., Keese, P., Raybould, A., Fitzpatrick, J. W., Burachik, M., Gray, A., ... & Wu, F. (2010). Problem formulation in the environmental risk assessment for genetically modified plants. Transgenic research, 19(3), 425-436.

Dear colleagues. My name is Eva Bermúdez from the Executive Secretary of the Intersecretarial Commission of Biosafety of GMO (Cibiogem), México. I am focused mainly in monitoring and detection of GMOs.

First of all, I would like to thank the moderator, and all the participants for how productive and active this forum has been. Now, as requested by the organizers, in order to share information that could complement the draft study, I would like to emphasize some points that could be improved in the document:

a) I  agree with Yann Devos [#10127] in the need to clearly distinct between the different cases of gene drive modified organisms; especially in order to propose considerations for risk assessment. Could it be helpful to create a table to put on the different gene drive systems and their particularities and the statements that apply to each one. 
b) It’s necessary to go deeper in study cases already available about gene drives experiences and the difficulties presented. In p.3, l.25 it’s mentioned that “applications are close to being released in trials” so it would be desirable a more detailed explanation about this trials and the risks assessment applied in them. Also, I agree with Ricarda Steinbrecher´s comment [#10162] about that natural gene drives could bring information but not should be taken as baselines for risk assessment.
c) The precautionary approach, is not properly mentioned nowhere in the draft. I think it´s important and necessary to complement the document including the application of the concept as stated in the Cartagena Protocol; under the fact that the organisms modified by gene drives have adverse effects on the environment and human health. More discussion is needed about mechanisms of risk assessment. I would like to thank Christoph Then [#10135] for sharing the report on GE gene drives.
d) About the effects in biodiversity, I also agree with previous comments about the convenience of consider ecological risk assessment approaches. Also, when suppression gene drives are mentioned as a way to eradicate invasive species, or as an option to stop vector-borne diseases, the ecological role of the target and the effects should be more discussed, considering, i. e. the possibility that other species occupy the ecological niche left by the extinct organism or the adaptation of the virus to a new vector with greater infectious capacity.  In this cases, before the implementation of a gene drive strategy, all the other control mechanisms should be tested and compared in order to find a plausible solution with the less risk possible. As mention by others before [#10128], [#10147], [#10153], [#10162], [#10166], this idea needs to be more discussed in the draft. Finally, the approach of restore ecosystems with gene drives engineering organisms to eliminate invasive ones is very limited because the removal of the invaders does not assure the reestablishment of the endangered species.
e) There is no mention to the protection of the biocultural heritage in the considerations on risk assessment identified in the draft. This is of a great importance, mainly in countries with a huge amount of biodiversity and cultural practices related to it. To improve this, the active participation of IPLCs is crucial in all the stages of the risk evaluation and it should be more than the free, prior and informed consent. Appropriate mechanisms should be developed with the participation of the communities and according to their cultural values.

Best regards 
Eva B.

Dear participants,

My name is María Elena Mondragón, I collaborate with the Executive Secretary of the Intersecretarial Commission of Biosafety of Genetically Modified Organisms (Cibiogem), Mexico. My responsibilities are focused on the socioeconomic considerations of GMOs.

I agree with Christoph Then [#10135] that the Perseus report considers in a poor way the Precautionary Principle. This is worrisome because the precautionary approach is very important to address threats to biological loss or reduction. In that sense, section 4.1.5 about “Considerations for risk assessment” of the report presents an inaccuracy by stating that in “In line with the precautionary approach, scientific uncertainty must be reduced in order to advance through R&D”. Although, the reduction of uncertainty its import for R&D activities, this not the principal objective of the Precautionary Principle. In my opinion, it is important to emphasize the relationship between precautionary approach and the conservation of biodiversity, and not with R&D.

Since my point of view, precautionary approach could be considered in an integrated way with risk assessment, because the context of decision making include policymakers, as well as scientist positions and depends on the scientific uncertainty degrees which is identified in the risk assessment. Considering the precautionary approach in the risk assessment does not necessarily imply a confusion as mentions Pieter van der Meer [#10170].

I also agree with Ann Kingiri [#10147] about the importance of evaluate ethical aspects of gene drives applications. From an environmental ethics perspective, it is possible to consider the value of nature and the conservation of species. Also, within the ethical aspects, is important take into account the right to public participation, especially of indigenous and local communities, who are carriers of traditional knowledge associated with nature conservation. These points are within the scope of the Cartagena Protocol and, therefore, in the scope of Annex I of decision CP-9/13. 

In this regard, I fully agree with Nada Hamza [# 10161] that regulations must be carried out from the bottom up, including ethical and socio-economic aspects. These aspects should not be secondary to the decision-making processes and should accompany the risk assessment, since the risks of technology can also be social and cultural. The Perseus report lacks of these issues.

In conclusion, I consider is a priority to improve the report with a multidisciplinary approach in which ethical aspects must be considered within risk assessment. Capacity-building activities must consider these aspects in order to address the biodiversity conservation. 

Finally, I would like to thank to Christoph Then [#10135] for sharing documents on the ethical aspects of gene drives applications. I am sure they will contribute significantly to improving the Perseus report.


Best wishes.
María E.

Dear colleagues,

As mentioned previously, I would like to share some more thoughts on the draft and the discussions.

One aspect that has not had the attention in the report that it requires is evolution, as well as co-evolution. The only mention of ‘evolution’ in the actual text of the report so far is on page 27, line 1, refering to the “potential evolution of resistance” in the context of HEG-based gene drives, where the CRISPR/Cas target site mutates. Reference is also made to Bull’s work, for example in the context of resistance development where the GDO would develop a “method of specifically inhibiting the drive endonuclease” (p27, L8), or to his 2019 publication “Gene-drive-mediated extinction is thwarted by population structure and evolution of sib mating” (p 34, when listing articles on modelling). Otherwise there is regrettably no mention or deliberation on evolution.

However, I regard it as crucial that Evolution and/or Co-evolution as such should have their own bullet point under the considerations on risk assessment, as well as feature elsewhere. Here a few reasons why:

(1) Most engineered gene drives are using or are based on so-called “selfish genetic elements” (SGE). SGE’s are part to a constant counter-adaptation and co-evolution process within a genome – which has bearing on GDO risk assessment.

In our recently published report ’Gene Drives’ for example we state: Rather than casting the selfish genetic elements (SGEs) in a negative light, and focusing on the as¬pect of ‘selfish’, it is precisely this co-evolution and co-adaptation that is becoming a focus of research. Instead, some of the elements or mechanisms, for example the over-replication ability of transposable elements, are increasingly regarded as vital components for genome evolution and even speciation (Biemont 2010). John H. Werren importantly noted: “The story that is emerging increasingly supports a central role of SGEs [selfish genetic elements] in shaping structure and function of genomes and in playing an important role in such fundamental bio¬logical processes as gene regulation, development, evolution of genetic novelty, and evolution of new species.” (Werren 2011, 10863). In fact, the study of these elements and the processes involved are now contributing to an emerging re-think of what a genome is and how it interacts with its environment. (Werren 2011; Lindholm et al. 2016).

    Consequently, when talking about engineered gene drives, which are all based on and are exploit¬ing the mechanism of these SGEs, there is a level that we cannot comprehend at this point in time. SGEs are vital evolutionary players, deeply embed¬ded in a long evolved and complex regulatory struc¬ture. What therefore does it mean to take SGEs out of their own context, reshape and alter them, and place them back into this interactive system? What might the consequences be at that specific level?

    This is an important discussion that needs to take place now. It would be wrong to inadvert¬ently assume that SGEs are only a “tool” that can be readily adapted and utilised for the purpose of modifying organisms and whole populations in the wild. They are much more than that and it could be most unwise to disregard this.  (Steinbrecher & Wells, 2019; p.35/36, Chapter 1 in: Gene Drives, CCS/ENSSER/FGS. https://genedrives.ch/wp-content/uploads/2019/10/Gene-Drive-Kapitel-1-WEB.pdf)

(2) It is important to study and to know whether the use of engineered gene drives will impact or speed up evolutionary processes within the target organism.

What for example will happen with an active CRISPR/Cas genome editing molecule present in the genome of an organism and where that organism has developed a resistance to the CRISPR/Cas gene drive due to sequence modification of the target site? Whilst the gene drive as such is no longer active, the CRISPR/Cas gene construct and its product are. CRISPR/Cas molecules will be produced for example during phases of sexual reproduction and may well inflict the occasional off-target cut, thus instigating off-target mutations. Would that therefore translate into LM organisms containing active CRISPR/Cas genome editing molecules having a higher mutation rate than their non-engineered counter-parts? What would that mean for an organisms that is known to be a carrier of diseases, or an alien invasive species, or a crop pest? Would it result in a speedier adaptation rate to stresses, pesticides or other adverse challenges?

There are many aspects that when seen through the lens of evolution are highly relevant to risks and risk assessment of GDOs. In fact some specific gene drive systems and mechanisms may be particularly vulnerable to this  – and these issues need to be addressed in the draft report.

This notion is in line with contribution [#10168], stating “…..however, when addressing the long-term and evolutionary consequences of GDOs their unique molecular and biological features must be taken into account.”

Given the complexity of the issue(s) at hand the different kinds of knowledge required I very much agree with Marvis Suárez Romero [#10118], stating that “properly trained multidisciplinary groups are needed” for the risk assessment of GDOs, which is also evidenced in [#10120].

I will come back on points regarding the precautionary principle and modelling relevant for risk assessment in a later submission.

With kind regards, Ricarda

Dear forum participants,

My name is Delphine Beeckman, and I’m for the first time participating in the CBD online forum on Risk Assessment and Risk Management. As a biosafety officer, focusing on the use of regulated and non-regulated biological materials in containment (e.g. labs, greenhouses, animal facilities), I am following up on developments in risk assessment for different types of biological materials, including gene drive organisms (GDO).

I would like to thank Marja for keeping us on track in the discussions, and also thank the authors of the Perseus draft study for providing a solid basis to discuss the topic of gene drives.

Having followed the multiple and often very detailed submissions to the forum, I would like to highlight the following aspects:
1. Many interventions discussed the scope of the draft study, and whether to broaden or not the currently applied definition of GDO. I would like to express my support to those that have argued to not open up the definition of GDO beyond the meiotic gene drives (e.g. Werner Schenkel [ #10122], Yann Devos [#10127], Hector Quemada [#10129] and Piet van der Meer  [#10170]), and this to allow focused and meaningful discussions, rather than getting lost in lengthy conceptual considerations with no concrete endpoint.
2. As regards to possible gaps or improvements for the report, I largely concur with the submissions from Yann [#10127], Ana [#10128], Hector [#10129], Boet [#10137] and Werner [#10167] and will not repeat these here for reasons of brevity.
3. Finally, on the question whether there is need for additional guidance materials on risk assessment for GDO, I agree with many participants stating that there is already considerable guidance available (nicely summarized by Giovanni Monge [#10164]!) which is also flexible enough to allow for a case-by-case application, but rather that there is a need for sharing of experience and good practices amongst risk assessors (e.g. Ana [#10128), Keith [#10145], Ann [#10147], Werner [#10155], Roger [#10174] etc.), both from a regulator’s and an applicant’s point of view.

Looking forward to the last days of discussions on this forum!

Kind regards,
Delphine

Dear Colleagues,
I would like to make some final reflections on the issues we have been discussing. First of all, I personally find most of the contributions quite useful and interesting. I am pretty sure that they will contribute significantly to the further improvement of the final report.
Regarding the analysis of the organisms with engineered gene drives vs the criteria set in Annex I of decision CP-9/13 I continue to believe that the topic essentially fulfils them all.
My own opinion is that what is really important is that the ERA of living modified organisms containing engineered gene drives before their release into the environment is up to the required standards and that all potentially affected parties and stakeholders have been consulted and have participated as appropriate.
The ERA principles are sound and applicable in the case living modified organisms containing engineered gene drives. But that is a general conclusion. Applying sound methodology, when you don’t have the necessary information inputs or even worse when you don’t even know what those inputs are, will not result in sound risk assessment. The problem is that for the living modified organisms containing engineered gene drives both the known unknowns and the unknown unknowns are many (although they well be reduced with the experience in the coming years). So it is important from the beginning to outline the general principles and considerations that should be taken into account when conducting ERA of those organisms. I think that is type of document that is most urgent to be consider.
Such document will help to reassure the public and get its acceptance of the technologies if they prove to be useful to solve human problems.
The potential benefits of the gene drive technologies are too many and significant to just ignore or dismiss them due to the lack of field data. So I think it is very important to design experiments (releases) of living modified organisms containing engineered gene drives that will provide necessary data without (or with minimal risk) jeopardising the biological diversity. Let me give a purely hypothetical example. It will be interesting to release on a small scale into the environment LM mosquitoes with gene drive (to be used in the future) that carry instead of the proper cargo, gene without any known effects on the organism (or the environment) but useful for monitoring, e.g. GFP. Such release will have the potential to provide plenty of real data about the spread of the gene drive in time and space, development of the resistance, interactions within and between species, etc. This the type of boldness combined with the necessary precaution that I believe we need in order to move the technology forward in a safe and sound way.
Best Regards,
Nikolay

Dear colleagues thsank you for the lively and inspiring discussion.

I wanted to add few thoughts on the precautionary principle.
The application of the precautionary principle is a compromise between a too late acting prevention of hazards on the one hand and an, constitutionally problematic, arbitrary regulation on the other hand. Prevailing purpose of the precautionary principle is to allow regulatory intervention at a point of time where a concrete harm has not yet occurred or is not yet of a very high likelihood to occur. Nonetheless regulatory bodies may only invoke the precautionary principle under certain premise, at least according to European jurisdiction. Accordingly, a preventative measure cannot be justified by purely hypothetical arguments based on assumptions lacking scientific verification.
Instead, a case by case risk assessment based on sound data and recent international research is an indispensable prerequisite to identify scientifically based indications towards risks that warrant preventative regulatory measures justified by the precautionary principle.
Werner

Posted on behalf of Mr. Christian Damgaard, Denmark:
*******************************************
Several participants in the discussion have advocated the use of the precautionary principle. I disagree and suggest we instead consider the scientific risk assessment approach on specified risk scenarios (case-by-case) where Risk = Probability * Effect.

I find that the precautionary principle, which comes in different flavors (soft, intermediary and hard interpretations of the principle), works poorly in concrete cases. In practice the consequence of the precautionary principle often is that the probability of a risk scenario is occurring a priori is set to one, and the principle therefore enforces too conservative risk recommendations.

Mr. Christian Damgaard, Denmark

Dear all,

My name is Luciana Ambrozevicius and I´ve being working in the biotech field since 2000, including GMO risk assessment and monitoring.

Thank you for you for the authors of the study and for this interesting discussion  Although I concur with others with regards to the Perseus report providing a good baseline with the conclusion that ERA of organisms with gene drives can be conducted, on a case by case basis, through the existing methodology, I also agree with #10149 regarding the fact that the report’s authors missed an important opportunity to gain information from interviews with developers of gene drives and population modelers to explore their approaches regarding new gene drive designs and the risk assessment capabilities. Having this information could help some of the issues presented as challenge for the existing RA framework: the lack of applicability of a stepwise approach and the use of robust models for gene drive organisms, according with the study.

I also agree with #10155 that it should be differentiated between challenges to the RA methodology and challenges relating to obtaining information required to inform the RA and, if the challenge is in obtaining the data required, guidance will not help. I would also like to mention the importance of a deep knowledge of the biology of the specie to fully consider its potential interaction with the environment and in this case I would like to refer to the OECD Consensus Document on the Biology of Mosquito Aedes aegypti http://www.oecd.org/officialdocuments/ publicdisplaydocumentpdf/?cote=ENV/JM/MONO(2018)23&docLanguage=En  as an important tool to be considered as a source of information for a RA and the case-by-case nature of any RA to be conducted.

I consider gene drive application a priority for many developing countries, not related with biodiversity but related with public health. In this case any decision will always have to consider the maximum acceptable risk in face of the benefits for the population and the other control methods available. But instead of use the scarce resources for new guidance, I believe that risk assessors training and technology transfer are much more important. I also believe that the precautionary approach should not be used as an argument to avoid the development of important technologies that can save lives in countries devastated by tropical diseases.

Thank you.

Best regards,
Luciana

Ministry of Agriculture - Brasil

My name is Todd Kuiken and I am a senior research scholar at the Genetic Engineering & Society Center at North Carolina State University.  Trained as an environmental scientist, I have been working at the intersection of environmental policy and emerging technologies for the past 12 years; focusing on synthetic biology and other emerging genetic technologies. First let me thank our moderator, Marja for leading and probing our discussions so far and the authors of the Perseus report for tackling such complex issues in a short period of time.

1. In relation to expanding the definition of gene drives, I support those that have argued not to open up the definition beyond meiotic gene drives (Werner Schenkel [ #10122], Yann Devos [#10127], Hector Quemada [#10129] Piet van der Meer  [#10170], Delphine Beeckman [#10178]). This can be complimented by the applications listed in the more recent work of Dr. Ricarda Steinbrecher mentioned in comment [#10162]. For other types of gene drives suggested by Mr. Jack Heinemann [# 10117 10134], among others, these additional gene drive systems could be monitored within the horizon scanning exercises being proposed by the 2019 AHTEG on synthetic biology which is designed to provide early identification of applications that may require adapting current RA methodologies. 

2. I support Boet Glandorf [#10137] and Dr. Werner Schenkel [#10155] comments in full.

3. I support Dr. Keith Hayes [#10145] and Ms. Ann Kingiri [#10147] in full, and would add in relation to capacity building that the 2019 Synthetic Biology AHTEG (which also evaluated gene drives under the auspices of synthetic biology) “recognized the need for capacity-building for developing countries and indigenous peoples and local communities and other actors to enable them to engage in the assessment of actual and potential impacts of synthetic biology. It also noted the need for the participation of youth and indigenous peoples and local communities in activities related to synthetic biology carried out under the Convention. It considered that appropriate communication and engagement with communities was important for building the necessary understanding for informed consideration. The AHTEG also considered that, although the need for capacity development had been repeatedly noted, more support was needed in that regard. The need for technology transfer was also noted.”

In addition to developing technical capacity, infrastructure capacity will also be needed. Research facilities (i.e. mesocosms and other field trial facilities), similar to the Experimental Lakes Area (https://www.iisd.org/ela/) may need to be established, taking into account specific issues related to engineered gene drives, in order to collect the data required for RAs.

4. In relation to FPIC, I’d like to submit a recent article focusing on gene drives to compliment the Perseus report (attached). In this paper, we argue for an articulation of FPIC that attends to issues of transparency, iterative community-scale consent, and shared power through co-development among Indigenous peoples, local communities, researchers and technology developers. In realizing a comprehensive FPIC process, researchers and developers have an opportunity to incorporate enhanced participation and social guidance mechanisms into the design, development and implementation of engineered gene drive applications. George DR, Kuiken T, Delborne JA. 2019 Articulating ‘free, prior and informed consent’ (FPIC) for engineered gene drives. Proc. R. Soc. B 286: 20191484. http://dx.doi.org/10.1098/rspb.2019.1484

5. I would like to draw our attention to a potential issue for RA related to Chapter 5, section 5.1.5 “Specific issues for engineered gene drives”; lines 3-7: “The specific issues concerning living modified organisms that: Have the potential to cause adverse effects on biodiversity, in particular those that are serious or irreversible, taking into account the urgent need to protect specific aspects of biodiversity, such as an endemic/rare species or a unique habitat or ecosystem, taking into account risks to human health and the value of biological diversity to indigenous peoples and local communities;”

There are proposed applications of gene drives (i.e. coral resilience and others identified in the IUCN publication “Genetic frontiers for conservation”) that address “the urgent need to protect specific aspects of biodiversity, such as an endemic/rare species or a unique habitat or ecosystem” while at the same time may “cause adverse effects on biodiversity”. This dichotomy, in my opinion, is the dilemma around utilizing engineered gene drives for environmental protection/restoration. And poses a challenge for RAs unless addressed in the problem formulation stage by identifying protection goals and establishing risk tolerances prior to such an assessment being conducted (as mentioned in # 10168, #10145, #10127). 

6. Finally I’d like to support Dr. Ricarda Steinbrecher’s comment #10177 in identifying evolutionary changes as an important aspect of evaluating engineered gene drives.  While I believe some of these issues can be evaluated through modeling, other aspects will need to be monitored for, which I do not believe our current RA frameworks are designed to capture.  This long-term aspect of monitoring will need to be addressed.

Kindest regards,

Todd Kuiken

Dear Marja, dear All,

Thank you Marja for moderating this interesting discussion.
I would like to congratulate the Secretariat and the authors J. Smets and P. Rudelshteim for the very useful draft Report on GD.

I am of the opinion that the Report has to reflecting the main status of existent knowledge on GDO at this time. In this context I fully support Jack Heinmann in his view that the report has to be complemented with the definition of GD non-sexual reproduction organisms as there are available knowledge and scientific publications.

I thank Ricarda Steinbrecher and others suggested a very important point to consider long-term consequences and evolutionary aspects of GDO as a special bullet point in the Report. I find the gained arguments to modeling possible effects of resistance in populations and natural ecosystems are very credible and have to be highlighted in the Report. I would like also to strongly support Marvis Suarez Romero for his suggestion to further discuss these issues within a multidisciplinary group of experts to have more understanding and clarity.

Additionally I would suggest to have in the Report a larger description and modeling and available synthesis of information on possible effects over wild species populations and ecosystems, including the receiving environment but also in a larger scale.

I would not support the opinion of those participants to the forum who consider that it is sufficient to provide risk assessment of GD on case-by-case method just to be available to exchange these results between the researchers. In my view the topic of GD has to be in attention of regulators and decision makers, even there are limited applications in present.

Thanks,

Angela

Angela Lozan, Biodiversity Office, Ministry of Agriculture and Environment, Moldova

Dear Marja, dear colleagues,
I would like to address a few points that are linked to the ability/ inability to perform risk assessments with regards to GDOs which I find not sufficiently covered in the current draft report:
The bullet point list of considerations for risk assessment on pages 3 to 5 is a helpful overview of some major areas of uncertainty. At the same time any analysis should be made with an awareness that understanding of systems and their interrelations is incomplete. It should therefore be explicitly recognised that there is potential for outcomes and hazards that have not been anticipated, so called ‘unknown unknowns’. The effects of deploying gene drive are not easily predictable at the molecular level, nor at the level of the whole organism. Effects on a whole species, the ecosystem, or longer term evolutionary trends are still more difficult to predict. Given that all of these highly complex systems are dynamically interrelated, it follows that gene drives could have consequences that are largely or completely unforeseen – either in their nature or in their magnitude. 
It is therefore necessary to consciously and continuously follow and apply the precautionary principle (see interventions #10135, #10155 and #10166). The report partially acknowledges this (p. 25), and while it might be that further research could reduce uncertainties to a level that may allow for a reliable risk assessment, it should be made equally clear that in some, or perhaps many, cases this may not be possible and plans for deployment would then need to be halted. The implications of the precautionary principle in this regard should be made more explicit, as well as the application of the precautionary principle in general.
A related point concerns the use of models for risk assessment purposes;  the issues involved in using models to address uncertainty should be made clearer (see also intervention #10168). The summary of the report includes the proposal that ‘robust models to predict long-term and ecosystem effects’ will be required to inform risk assessment (p.6). Much later a partial caveat is included (p49. Annex 4) stating ‘… models are as precise as the designer is able to mimic natural situations’, but does not go on to clearly state the difficulties involved.
We recently statet in our report: “To obtain field data to support the actual ability to do proper RA of gene drives will thus be difficult, and in some cases might even prove impossible. Modellig effects are seen as an attractive alternative to extensive field testing. However, most modelling approaches for gene drives have so far been performed in order to evaluate efficacy and spread of the desired genetic modification e.g. (Unckless et al. 2015), not in order to anticipate risks. Modelling of ecological effects caused by gene drives which would be useful for ERAs has yet to be developed. In comparison to the modelling of efficacy for purpose, approaches to simulate ecosystem effects are far more complex.” (page 127, Gene Drives, Chapter 2 https://genedrives.ch/report/)
I want to highlight the difference of modelling for efficacy and for ERA, with the latter not as yet available, and much work needing to be done to test such modelling systems and approaches and understand the uncertainties, accuracies and inaccuracies involved. Such modelling requires its own uncertainty assessment, which also still will need to be developed and tested. Given that modelling is an area of expertise in itself, not all readers may be aware of the issues, so the manifold difficulties in developing ‘robust models’ should be made explicit early on.
The complex spectrum of foreseeable possible outcomes from the deployment of any gene drive organism should also be explored more explicitly (see Figure 3, p100 of our report). The effects of a suppression drive for example could range from complete eradication of a species or of a population, through to patchy or partial suppression, at times followed by population rebounds, through to little or no effect. Similarly the collapse of a drive (e.g. due to resistance build-up or evolutionary responses) could result in a complex mosaic of modified and non-modified organisms, with the modified organisms potentially acting differently to the non-modified, and possibly also different to each other. The spectrum of scenarios would have to give rise to a parallel spectrum of data sets, hazard identification and risk assessments.
Awareness of all of these complexities needs to be held in mind to properly frame discussions on risk assessment, including challenges, need for further methodology, steps, and guidance, and should be covered explicitly in the report, including its summary.
With kind regards, Ricarda

Hello.

My name if Jim Louter and I work for the Government of Canada and have been conducting regulatory risk assessments of genetically modified (and naturally occurring) organisms for 27 years.  During that period, there have been no requests for assessment of a genetically engineered gene-drive modified organism. 

Thanks go to both the moderator of this forum (especially for her nudge that comments made towards the end of the two week period would not likely be responded to) and to the authors of the Perseus report for the fine report they have put together in a very short time and with a very limited scope.  The advantage of being late to the game in the on-line discussion is that most points of view have already been expressed and so I will simply say that I support statements made in interventions 10137, 10149, 10170 and 10182.

I have a few additional observations and apologies if these have already been made.  I am not sure I fully agree with the use of 'suppression' and 'replacement' as terms to describe two types of GE gene drives. It would seem to me that the example of a 'replacement' drive (block pathogen [or parasite] development) (page 18 of the report) would work equally well for 'suppression'; I would prefer terms that expressed the temporality of the drive (i.e. self limiting or not) and clarity that the range of traits that could be incorporated are not limited to those that impart lethality or sterility. 

Lastly, as a regulator I observe that often times, under national legislation, research is exempted from oversight subject to certain conditions.  It might be time to consider additional oversight such that such research is subject to appropriate notification such as suggested by the Netherlands and that calls, by some, for 'moratoria' on research not be supported. Appropriate risk research must be undertaken in order for uncertainties to be reduced.

thank you

Jim Louter, Environment and Climate Change Canada

Posted on behalf of Dr. Carolina Alduvin, Honduras:
****************************************
Dear Participants:
My name is Carolina Alduvín and I also belong to the National Committee on Biotechnology and Biosafety in Honduras.
Adding to the above stated by Dr. Orellana and regarding the GDO discussion, seems to me that no matter which guidance or procedure we follow ERA, we will never know what would be the effects on natural wild environments until we try their release; with the available methods and considerations we would only anticipate possibilities, the likelihood of each one of them is also hard to estimate. The unintended effects, barely unpredictable and we have to take risks in order to learn, even knowing some of them could be irreversible.  We must think about the purposes taken in count when the GDO modifications were made or are being made and evaluate what could we loose or what could we will end paying in terms of human lives or human cripling disesases which could be avoided or at least put under control, releasing the GDOs to the environment. For those modifications reverting and loosing the intended effect, there is the alternative to reintroduce it with possible corrections according to observed threaths and non intended effects.
Dr. Carolina Alduvin, Honduras

As Marja and Werner [#10155] have recalled for us, the forum serves the AHTEG to assist the Parties in deciding if guidance is needed on the topic of engineered gene drives. Werner has helpfully listed four issues from the report and evaluated whether any of them fall outside of existing guidance. I would like to continue with that approach because a number of issues have been raised that are not covered by the study report. This intervention won’t be exhaustive, but indicative.

Annex 1(c and e) Eva Singhji [#10166] reminds us that “While off-target and unintended effects at the molecular level may potentially be assessed with current GMOs before an environmental release, this cannot be conducted on a GDO as there is no ‘final’ product, but instead the release of a GM tool that will continue to function in wild populations with potential next generation effects.” This is the key that is not covered in any guidance or experience to date. It is the ability, using topical agents, drives, or topical agents to introduce drives, to generate GMOs that are not able to be assessed prior to release.

Further, as Angela points out [#10184] how the drive mechanism itself evolves, not just how drive dampening reactions to it might emerge, after release from the laboratory is an issue novel to GDOs and their descendants. For example, drives based on Cas9 may evolve to recognise other RNAs as guides over time, a possibility anticipated by the finding that existing Cas9 relatives are more promiscuous in their tastes for guides, and also they can have targets other than dsDNA (Ma et Molecular Cell v60:398-407, 2015).

The issue of comparators is for me still outstanding. This issue is fundamental to the risk assessment methodology. First, the assumption is that engineered gene drives will remain contained by breeding. This is only true when, at best, they are in organisms that have strict and known interbreeding populations. Even meiotic drives can escape this assumption even between sexual species that do not mix their genes through sex (Houck, M. A., J. B. Clark, K. R. Peterson and M. G. Kidwell. 1991. Possible horizontal transfer of Drosophila genes by the mite Proctolaelaps regalis. Science 253:1125–1129). Gene transfers that arise this way and establish a drive in these new species would have no laboratory preview and no opportunity to attempt to model, construct a comparator, or use alternative means to evaluate. Echoing Marion Dolezel [#10168], Katileena Lohtander-Buckbee [#10169], Christoph Then [#10159] and others, this to me is a challenge to all existing methodology.

In summary, for reasons that I’ve argued and referenced through my various interventions I believe that there are reasons to recommend the development of new guidance both because of gaps and inadequacies of existing risk assessment methodologies and because of challenges to conducting a risk assessment. My opinion on this does not change even if the scope is arbitrarily defined as applying only to meiotic drives.

Developing this guidance will complement the call from countries, e.g. [#10152], [#10147] and [#10118], for a greater capacity to apply risk assessment and risk management. They ultimately may be the definitive testing grounds of the assertions that existing methodologies and biological examples will provide a reliable basis for estimating the likelihood and scale of hazard and the effectiveness of mitigation strategies, which among others Ricarda [#10185] has emphasised are untested at least beyond the laboratory.

This has been a wonderful and constructive forum for the presentation of ideas.
Jack

Dear All,

my name is Ernst A. Wimmer and my resaerch group is mostly working on biotechnological improvements to the Sterile Insect Technique. However, we have also carried out some experiments to the resistance development in homing endonuclease-based gene drives. Currently, I am also serving as a hearing expert to the EFSA Gene Drive working group to evaluate existing EFSA ERA documents for covering gene drive organisms.

While I find the current discussion very interesting and fruitful, I will restrict myself at the moment to identifying scientific gaps in the perseus report as requested by the moderator.

The Perseus Report is good starting point and well drafted. However, it has some gaps in describing the clear differences of the diverse approaches to gene drive. Most of the discussion is currently based on overreplication genes drives deploying engineered CRISPR/Cas9-based homing endonuclases. However, in the future interference-based gene drives such as MEDEA drives might become much more important. In this category, the recently described Cleave and Rescue Drive (Oberhofer et al., 2019) provides a system that is much less prone to resistance evolution than the homing-based drives. But also for the homing-based drives, approaches have been developed to overcome resistance development at least for suppression drives (Kyrou et al. 2018).
Moreover, the developments into spatially (split drives, Li et al, 2020; split Killer-Rescue Drives, Webster et al., 2019) or temporally restriced drives (daisy chain drives, Noble et al., 2019) are not really covered. I believe it would be very important to completely cover all these approaches and provide a clear distinction of the divers strategies. The case-by-case evaluation will become quite different regarding the different approaches. Thus the diversity of strategies should be recognized in the report. This will also make clear that it will be very hard to make general statements to the ERA of gene drive organisms as their characteristics will be complex but also quite different dependent on the exact approach.

I hope the soon availabe EFSA GMO Panel draft scientific opinion on “the evaluation of existing EFSA guidelines for their adequacy for the molecular characterisation and environmental risk assessment of genetically modified insects with synthetically engineered gene drives”, which will soon be launched for public consultation, will also be infromative in this respect and might help also the Perseus Report.

Best regards,
Ernst Wimmer

We wish to thank Marja for moderating this forum, as well as the Secretariat for their ongoing steady support throughout this process.

We also wish to thank the authors for their work on the Perseus report, and find that it provides a good overview of gene drives, particularly those that are the focus of most of the current research and application, and thus likely candidates to fulfil criterion (e)(iv) of Annex I. We note that the segregation distortion in asexually reproducing organisms (as discussed in several posts in this thread) is a misnomer, because in the absence of sexual reproduction, there is no allele/trait segregation, and thus a gene drive cannot function. This is discussed in detail in the references in [#10122], and broadening the scope of the report will lessen its relevance to the criteria of Annex I.

We agree with the report, and comments made in many posts [#10119, 10137, 10139, 10141, 10145, 10147, 10161, 10174], that case-by-case analysis will be necessary for any gene drive application, due to the wide range of variables that will be in play in any problem formulation with a gene drive organism (GDO). Thus, we agree with the comment on p6, lines 25-26, in relation to criterion (d) of Annex I, “…the only way to clearly describe specific issues is in relation to specific cases.” Our reasons for this are further elaborated in the following two paragraphs.

We also agree with the report in stating that most of the environmental risk assessment (ERA) considerations of GDOs are not unique to GDOs, and that existing ERA methodologies can be applied to them. We therefore agree with the comments made in several posts, but particularly [#10174], that the principles of environmental risk assessment apply equally well to the assessment of LMOs, hazardous substances, epidemiology, invasive alien species, etc., and will also apply to the assessment of gene drive organisms. Thus, training and experience sharing in the application of the principles and methodologies of ERA will likely be more beneficial as guidance to risk assessors for all LMOs under the Cartagena Protocol.

Finally, we are all aware that research on gene drives is a rapidly moving area of research, and that methods for controlling the spread and mode of action are being developed and modelled, and the thinking for a “best practice” gene drive is rapidly changing. Thus, the GDOs that may eventually be risk assessed for environmental release may not greatly resemble those that are currently research and discussion subjects.

Thanks to all for the interesting discussion!

With regard to the risks associated with Horizontal Gene Transfer (HGT) #10191 writes: "Gene transfers that arise this way and establish a drive in these new species would have no laboratory preview and no opportunity to attempt to model, construct a comparator, or use alternative means to evaluate". I would like to emphasize that it is possible to model the risks associated with HGT, and we have done so on numerous occasions in past and current risk assessments, using techniques (such as fault tree analysis) that were specifically designed to assess the risks associated with high consequence, low likelihood events in complex systems. It is entirely possible to make probabilistic predictions of the likelihood of such events. The challenge in this context is not how to make these risk predictions (there is no gap in existing risk assessment methodologies) but rather how to obtain observations to (in)validate these predictions.

In this context I think it is also worth drawing attention to one of the main conclusions of the 2016 NASEM study "Genetically Engineered Crops: Experiences and Prospects": specifically:
"Overall, the committee found no evidence of cause-and-effect relationships between GE crops
and environmental problems However, the complex nature of assessing long-term environmental changes often made it difficult to reach definitive conclusions" (p. 100, NASEM 2016). To me this indicates that an important challenge (even in relatively simpler agro-ecosystems) is the design and implementation of post-release monitoring strategies that are capable of generating the evidence necessary to confidently test risk predictions.

Dear participants,

thank you, once more very much for your important and informative comments!

You have shared information by providing citations and publications. Moreover, you have made many suggestions in order to fill information gaps and/or correct errors, and to provide explanations and more details on concepts and ideas. All this is very helpful and will assist the consultants in generating the next version of the study.

Again, you have with enthusiasm discussed issues that are somewhat outside the main purpose of our exercise, which is to review the study taking into account the mandate from COP-MOP. Having said that, I of course do recognise the importance of some of these broader issues raised and discussed. But as a moderator of this important discussion, I want once more to point out the following: the main objective of the study was not to be a desktop study on engineered gene drives, nor to collect baseline information for an eventual risk assessment, but to inform the application of annex I to this topic.

A few words about the precautionary approach. The precautionary approach is contained in Principle 15 of the Rio Declaration on Environment and Development and further included into the Preamble and Article 1 of the Cartagena Protocol. It defines the objective of the Protocol and all Parties have reaffirmed it. The possible challenges in implementing the approach in decision making is not a topic for our present discussions.

Still some time left for your last interventions!

All the best, Marja

Dear Marja, dear All,

My name is Samson Simon, I work for the German Federal Agency for Nature Conservation (BfN) as a biosafety expert.

Here are some general and specific points I noted about the draft study on gene drives.

1. When evaluating the impact of synthetic gene drives on biodiversity it is important to have the correct reference system. The CBD has the clear term to conserve biological diversity in its entirety, which by definition is very broad including inter alia “diversity within species, between species and of ecosystems”. Synthetic gene drives may have the capacity to negatively impact all of those categories. Therefore a potential impact on all of those categories should be recognized for synthetic gene drives in the report. In general, the draft study reflects a too narrow view on biodiversity that is characterized by terms like “keystone species” (#10162; p. 4 l. 17), “valued species” (p. 4 l. 21) or “ecosystem services” (p. 30 l. 31).

2. One key element in the risk assessment of gene drive are the data needs on potentially affected species and ecosystems, and the study should elaborate more on the current knowledge and methodological gaps. The biggest gaps exist in ecological knowledge about those affected species, their role in nature and the ecosystems that might be affected. These knowledge gaps are exacerbated by the finding that ecological modelling is currently an underdeveloped scientific field. The draft study recognizes a need for robust modelling (p 31 .30-31). It should be clearly stated in the study that knowledge gaps and lack of ecological modelling tools probably hinder the execution of a sound (ecological) risk assessment currently, but also in the near future.

3. It is important to further work out that synthetic gene drives create fundamentally new LMO when compared to other LMO, as also recognized in the discussion (e.g. #10166, #10169). The general concept of engineered gene drives offer so many novel challenges that a technology assessment is justified for engineered gene drives (c.f. Simon et al. 2018). Widening the view from sound risk assessment to considering socio-economic considerations is laid down in article 26 of the Cartagena protocol and should be discussed by the study at hand.

4. The draft study and in this forum it has been argued for a strict case by case examination and against general characteristics of gene drives (e.g. global vs. local drives; suppression vs. modification drives). At the end every case is different and a case by case evaluation is crucial, but nonetheless categorizing gene drives will not be as straightforward in practice as it is attempted in theory. E.g. natural gene drives and importantly engineered gene drives can have states where wt and suppression drive alleles stably coexist in populations (Champer et al. 2019; Price et al. 2019). Vice versa, a recent publication shows experimental laboratory data for a population collapse using an efficient modification drive due to knowledge gaps on the target gene (Pham et al. 2019). Those findings for engineered drives blur the borders between suppression (the trangene will be lost) and modification drives (population numbers are stable) for risk assessment aspects. This should be taken into account when assessing e.g. the strict separation of suppression and modification drive for risk assessment by the authors of the study as an argument for sole case by case examination. It shows the general power of gene drives to potentially alter whole populations or ecosystems, which is an intrinsic property shared by all types of gene drives.

5. It is also important to emphazise that a guidance on general challenges to the risk assessment of synthetic gene drives is an important and helpful tool to risk assessors. I do see that there are multiple challenges ahead. I have already discussed some intrinsic properties of gene drives and knowledge gaps in ecological data and evaluation. Other important general problems are the lack of suitable comparators and and appropriate monitoring concepts as also discussed here before (e.g. #10183; #10195; #10169; #10191).

With best regards,
Samson

Literature:
Champer et al. 2019 DOI: 10.1101/769810
Pham et al. 2019 DOI: 10.1371/journal.pgen.1008440
Price et al. 2019 DOI: 10.1098/rspb.2019.2267
Simon et al. 2018 DOI: 10.15252/embr.201845760

Dear Colleagues,
As the precautionary principle has been mentioned a few times, I would like to point out that its application is not a legal requirement under CBD as it is part of the preamble. That’s not the case in some other legislations, such as EU where it is stipulated in Article 191.2 of Treaty on the Functioning of the European Union
On the application of that Principle has been a number discussions, but a detailed analysis is available (in English, French and German) on the website of the European Parliament https://www.europarl.europa.eu/thinktank/en/document.html?reference=EPRS_IDA(2015)573876 .
Best Regards,
Nikolay

My name is Anastasia Pagida and I work in the Biotechnology Unit of DG SANTE of the European Commission. I would like to thank you all for this active and rich exchange of views and information and also thank Marjia for moderating the discussions.
- Considering the use of the precautionary principle, you may find useful elements in the Commission Communication (https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52000DC0001&from=EN). I agree with Pieter (#10169) that it should not be confused with scientific assessment. The precautionary principle is essentially used by decision makers in the management of the risk. In order to apply the  precautionary principle decision makers need to consider an, as much as possible,  complete scientific assessment identifying at each stage the  scientific uncertainties. In this respect Perseus report should shed light  on whether guidance is needed for the evaluation of scientific uncertainties inherent to the risk assessments of organisms with engineered gene drives (GDOs). As many participants have indicated GDOs may create unprecedented challenges for risk managers, (requirement of consulting the public, relevant stakeholders and IPLCs, considering social values and ethical issues, identifying protection goals, setting the level of acceptable risk). However I do not think these challenges should be discussed in Perseus report under Decision CP-9/13, which is meant to address only  the risk assessment of GDOs.
- On the gene drives  in the context of organisms or genetic elements that reproduce asexually, I would agree to have relevant information complied in a separate Annex as proposed by Keith (#10145), but the discussion would benefit by focusing on applications that are more likely in the near future, as suggested by Yann (#10127),  Boet (#10137) and other participants.
- I believe that GDOs fulfil all criteria of Annex I of the decision 9/31 (see also posts 10168 and 10169, 10179). General principles and methods of environmental risk assessment apply, but the challenges highlighted in Perseus report and in the on line discussions might call for the development of specific methodologies, models and datasets.

My name is Camilla Beech and I am independent regulatory consultant with over 20 years experience in GMO’s and risk assessment, and recently for projects that involve gene-drives.

Thanks for  a fruitful and useful discussion on  synthetic gene drives and to Marja for moderating and reminding us that the scope is to inform the application of annex I to this topic [# 10197]. Congratulations to the report authors as well in producing the report for this discussion.

I would like to agree with #10193 regarding the requirement of a sexual  reproduction being a pre-requisite for gene -drives as without sexual reproduction a gene drive cannot function and therefore the extension to asexual organisms is unlikely to be relevant.

However the main point of my submission is to highlight to the forum that there are existing assessment frameworks that could be used for the examination of gene drives – and that these assessment approaches should be considered  for their utility and that “ de novo’ guidance for gene drives is unlikely to be necessary.

These approaches look at both biodiversity, socio-economic and health impacts ( both positive and negative ).  There is  already  established guidance and methods for these type of assessments such as from the  International Finance Corporation ( part of the World Bank).

e.g https://www.ifc.org/wps/wcm/connect/9608497e-56e8-4074-bab6-45c61a36a4ad/ESIA.pdf?MOD=AJPERES&CVID=jkCYZ3G  and https://www.ifc.org/wps/wcm/connect/Topics_Ext_Content/IFC_External_Corporate_Site/Sustainability-At-IFC/Policies-Standards/Performance-Standards
Thank you

Dear Marja and other experts,

I am Xubin Pan, working in the Chinese Academy of Inspection and Quarantine.

The Perseus Report is really a good start. I have three suggestions: 1) This document can collect more relevant reports from governments, international organizations and so on; 2) All these documents can be listed at the BCH websites, so everybody can read it by the list and then make a reliability assessment; 3) The authors can continue the interviews to have more participants.

Best,

Xubin

My name is Khara Grieger, and I am currently a Senior Research Scholar in the Genetic Engineering and Society Center at North Carolina State University. My main research areas relate to environmental and human health risk assessments for emerging technologies and advanced materials.

I would like to express full support for the concept of performing separate uncertainty assessments that complement risk assessments in order to identify key areas of uncertainty that go beyond statistical uncertainty (similar to comment #10185). This could be identifying the 'location,' 'level,' and 'nature' of uncertainty (following Walker et al. 2003) or using other broader conceptualizations to identify and describe aleatory and epistemic uncertainties in the risk assessments. Uncertainty assessments will be important to conduct in conjunction with the risk assessments, particularly if a case-by-case approach is needed for separate gene drive applications (similar to comments #10119, 10114, 10127, 10128, 10137 and others). Furthermore, conducting separate uncertainty assessments that complement risk assessments for gene drive applications will be important to accurately and transparently communicate the uncertainties within the risk assessment processes, including communication to decision-makers (related to comment #10139).

For reference: Walker W.E., Harremoes P., Rotmans J., Van der Slius J.P., van Asselt M.B.A., Janssen P., Krayer von Krauss M.P. 2013. Defining Uncertainty: A Conceptual Basis for Uncertainty Management in Model-Based Decision Support, Integrated Assessment, 4(1), 5-17.

First, I want to thank the CBD Secretariat for the opportunity to participate in the forum, and Marja for her role in chairing it.  This is my first opportunity to contribute, but I have been following the discussions and greatly appreciate the contributions that have been made so far.  Exchanging views always has value, even if it reveals that people may have different ideas about the subject.

We have been asked to provide comment specifically on the report and on its relevance to applying annex I of CP 9/13 to consideration of LMOs containing engineered gene drives.  While I’m very aware of Marja’s entreaty that we should participate early and often in order to give and receive feedback (and very sorry I have not been able to post prior to this), I am happy to take advantage of the fact that most of what I would like to say has been said (and probably said better) by previous posters so I will try to be brief.

Regarding the report itself, I think it is a very good attempt to summarize the general state of knowledge regarding gene drive organisms.  It is imperfect, as any such document will always be, but I think it is very helpful for facilitating discussion.  I would agree with Werner Schenkel [#10122] and others who suggest that history and practicality both indicate meiotic gene drives is the appropriate subject of the report and our discussions.  I would like to especially echo the comments of Yann Devos [#1027] as providing a series of useful recommendations if any further development is to be done on the report. Especially the idea that highlighting our relevant experiences that come from other activities and associated sources of information would be very valuable, and this is also highlighted by Ann Kingiri [#10147] in relation to experience with biological control.  In my opinion, putting effort into that exercise would be much more valuable than pursuing a new guidance document under the model previously followed in these forums and the AHTEGs.  I’d also like to strongly agree with Ana Atanassova [10128] and Keith Hayes [10195] who both point out that many of the issues raised in the document are not unique to consideration of gene drive organisms and that tools are available, together with some considerable experience, that allow us to explore these issues in the context of risk assessment.
I don’t think the document has any fatal flaws, although I would strongly agree with Keith Hayes [10145] that it is not possible to widely generalize about whether population suppression or population replacement strategies will have more ecological impact.  There are simply too many case specific considerations that will affect the outcome.  Which highlights the conclusion of Nicolay Tzvetkov [10139] and others with whom I would also agree that attempting to put together general guidance on gene drive organisms would not be productive and instead, as suggested by Roger Orellana [10174], efforts that encourage the sharing of knowledge and experience among researchers and regulators would be a more productive use of time and resources.  Tim Strabala [10193] provides a succinct conclusion that I would also strongly support.

Several contributors have suggested that gene drive organisms pose fundamentally different challenges to risk assessment than other LMOs with which we have experience, but I have not heard any considerations that have not been brought up many times before.  Horizontal gene transfer is not new, nor unique to gene drives.  The idea of long term or evolutionary impacts is equally relevant to any organisms that will be present in the environment over a long period of time.  The suggestion has been made that gene drives that introduce genetic engineering machinery into the environment, which may persist over generations, represents a unique circumstance. But this is undercut by the fact that all of the machinery in question has been derived from natural sources, and that (as many have pointed out) selfish genetic elements are ubiquitous in the environment.  Many naturally occurring selfish genetic elements are active and effecting changes to populations as we speak.  For one interesting example see Cash et al (full citation below), which describes the movement of Medea elements through insect populations in the United States over the last 20-30 years.  These elements arrive in genomes, spread through populations, and certainly impact the genetic diversity without affording the surrounding ecology millenia to “co-evolve” or adapt to their presence.  Certainly we could gain some insight by studying these genetic elements, and this suggests that the introduction of an engineered gene drive into a population would not be as unique or unprecedented as is sometimes claimed.

Finally, I would just like to highlight an observation that, in my view, continues to challenge discussions here and in the associated AHTEGs.  While it is expected (and in my view desirable) that experts have differing viewpoints, it remains problematic that there is no shared notion of what “guidance” is and what it is for among contributors.  The digression into the proper role of the precautionary approach in the report highlights this, and I would agree with Piet van der Meer [#10170] that it is misplaced here.  Guidance documents are not tools to dictate to Parties what their national biodiversity protection goals should be, nor how they will administer their risk assessments, nor how those risk assessments will be used to inform their national decisions.  Those are the sovereign rights of the Parties themselves to determine.  For this reason I continue to believe that providing access to information resources and forums for sharing experiences relevant to risk assessment would provide a much more valuable contribution to the needs of Parties than attempting to produce a guidance document in line with previous efforts in these forums.

Thanks again to the CBD for providing the opportunity to participate in the discussions, and to Marja for chairing them. 

Cash, Sarah & Lorenzen, Marce & Gould, Fred. (2019). The distribution and spread of naturally occurring Medea selfish genetic elements in the United States. Ecology and Evolution. 9. 10.1002/ece3.5876.

Dear Marja, dear colleagues.

My name is Christoph Lüthi, I work as a biosafety expert at the Swiss Federal Office for the Environment.

First, I would like to thank the authors for carefully preparing this report on a challenging topic. Because of the novel features of living modified organisms containing engineered gene drives and their currently discussed applications, it is crucial to proceed with caution. I think we all agree that the technology has the potential to cause “serious or irreversible damage” for the environment, as stated in principle 15 of the Rio Declaration, justifying the application of the precautionary principle. However, the specific risks can only be evaluated on a case-by-case scenario. There are several areas that can provide valuable information. E.g., there are relevant similarities to vector and pest control strategies or naturally occurring gene drives (as argued, e.g., in post [#10127]). To advance with gene drive organisms, it is crucial to seek an agreement on their regulation on all levels, notably regarding problem formulation, protection goals, appropriate risk assessment, risk management, monitoring and IPLC. However, as of today, I question the availability of sufficient empirical data regarding the potential impact of genetically engineered non-domesticated species in non-managed, highly complex environments across larger spatio-temporal scales. These issues need to be addressed prior to releasing GDO into the environment and the discussion needs to extend outside the scientific community.

Dear Marja,

Thank you for moderating this very interesting discussion.

My name is Aleksej Tarasjev, I am evolutionary biologist in Institute for Biological Research of University of Belgrade, Serbia. I am chair of Serbian Expert Committee on Biosafety and I am involved in various international activities related to biosafety (UNEP, FAO, SCBD).

It is extremely important to discuss various scientific aspects of gene drives, but from countries perspective Risk Assessment is almost always part of regulatory process. Therefore, for us is important to analyze those issues within broader concepts of Risk Analysis and Regulatory systems.

From Serbian perspective, it is important not only whether risks will be analyzed in context of Cartagena Protocol approaches or in broader context of Convention but as EU Candidate. Country, also how European Union will regulate them in case of gene drives. After the European Court of Justice decision on mutagenesis, in Serbia we already had applications involving New Breeding Techniques (CRISPER in particular). And for approvals to be issued we had to analyze risks as if applications are falling under the scope of our GMO legislation.

Therefore, I believe we will need lot of discussions and experience sharing regarding conflicting regulations from different regulatory systems in countries that are parties to CBD and CPB regarding various gene drive technologies, too. I think it falls under COP MOP agenda.

Regarding Risk Communication (as apart of Risk Analysis) gene drives can be discussed in broader sense, since most crucial point in those risks is their potential impact on environment and biodiversity and particular technique used is less important. I think they are also good example for making this point in public discussions regarding biosafety in general advocating case by case approach.

Sorry for posting my comments this late, I was following other contributions. I hope that we will continue our deliberations on those important issues on this platform soon. (maybe after EFSA report?).  Perseaus report and discussions are very good starting points.

Sincerely yours
Aleksej

Dear Colleagues,
The study document did touch some important areas as to risk assessment of LMOs containing engineered gene drives. The study falls within the scope of the Cartagena Protocol on Biosafety and annex 1 of decision CP 9/13 can also be applied.

Nigeria has prioritized LMOs containing engineered gene drives and has therefore amended its National Biosafety Management Act to include measures to ensure safety in the application of gene drive. The country is about to commence the process of putting in place procedure for risk assessing LMOs containing engineered gene drives following its Risk Assessment Framework that does not provide a systematic approach for addressing risk assessment of such LMOs.

The gene drive technology can lead to complete suppression or replacement of a particular species; modified competency in the transmission of other unpredicted diseases, and loss of acquired immunity in the case of gene drive application in tackling vector organisms, hence, requires special focus on risk assessment and mitigation measures.

While addressing the considerations of risk assessment of LMOs containing engineered gene drive, it would be necessary to consider alongside preventive and mitigation measures for the areas of considerations. The study document should have included information on the following:

1. Appropriate biosecurity and containment measures given the importance of a Confined Field Trial in decision for subsequent introductions.
2. Preventive measures for vector recovery in the case of gene drive application in human disease
3. Measures against gene drive resistance development
4. Measures to combat modified competency in the transmission of other diseases by gene drive vector organisms


Aligwekwe Blessing

Assistant Chief Scientific Officer

National Biosafety Management Agency, Nigeria

Dear Marja, thank you very much for moderating this Forum. Many thanks to the Secretariat for your valuable help, authors of the draft study J. Smets and P. Rüdelsheim for their not easy task, and to all the participants for the lively and very interesting discussions. Appologies for my late intervention.
Regarding to the question (c) I consider that this study is relevant to the aspects of annex I of decision CP-9/13. I think study is very good start and cover many important aspects. Regarding overall consideration to the report. I`m fully support Jack Heinemann [#10117] regarding narrowing definition of gene drives. It is very important. I`m of opinion that there is a need to take care about definition right now in the light of different heredity mechanisms, especially regarding non chromosomal inheritance (chloroplasts, mitochondria) which is totally differs from nuclear gene inheritance.
I also support Ricarda Steinbrecher post [#10177] that we need to pay special attention to the evolutionary aspects and think that consulting with experts in this field very important and definitely contribute this study. Also I support her previous post [10162] that the suggestion to focus on “most likely cases of gene drive modified organisms moving to practical application” could be addressed on the later stage and “if the scope is narrowed down too early it would not allow for the development of a broad-sweep catalogue of questions, concerns and inquiry”. I think it is important to have in this study as much comprehensive picture of developments as possible. My arguing here that we never know what unintentionally could be released into the environment, and also need to think on the possible effects and measures of monitoring or prevention. To that end I hope that it will be continuation of the study.

Thanks authors once again,

Best regards,

Galina

Dear All,

I was very delighted to read the post by Andrew Roberts (#10207) and fully endorse it. Even though we speak of synthetic or engineered gene drives, they are designed variations of naturally occuring genetic elements. They are still biological systems with all the disadvantages of variability and potential instability over long periods of time. Therefore, there are fair comparators, such as natural gene drives that do constantly occur, but also pest mangement applications such as classical SIT, Wolbachia drives or even chemical intervention. There is no such thing as a stable "natural" genome of a species or population. There is constant change! Selfish elements such as transposons move around within a species and between species.
In pest management applications, there is no reason to single out genetically engineered approaches to be particularly more problematic than other approaches. They need to be stringently evaluated, but should be compared to other approaches and not be disregarded just because they are based on genetic engineering. The continous use of insecticide-treated bednets has led to the selection of certain behaviours and the genome of the current Anopheles mosquitoes populations has dramatically changed in such areas. Whenever selection pressure is applied, genomes will alter. This is biology! There is no constant and stable genome, definitely not in the wild! Such ideas belong to romantic naturalists but are not scientific. This does not mean, we should not protect our environment. We defintely need to do that, and we should use the best ways possible.
Best regards,
Ernst Wimmer

Dear Participants,

The forum will close within less than two hours and I want to share with you some reflections after my posting from this morning. It is difficult for me to provide a full summary of all your comments at this point, but I will work with the Secretariat in compiling your very useful contributions into a report.

Some of the issues that have been discussed include, among others:

- The scope of the study and suggestions on how to address this issue (i.e. clearly stating in the study its scope and possible limitations that this may imply);
- The importance to consider experience with other processes to inform the risk assessment, while at the same time noting that the issue of engineered gene drives is quite unique and not all the experience from other processes will apply in full;
- The importance for capacity building, training, sharing experience and technology transfer as a key aspect for risk assessment of organisms containing engineered gene drives;
- The suggestions to use existing guidance and risk assessment methodologies to tackle the issue, supplemented with more information;
- On the other hand, suggestions for additional guidance and/ or possible outlines of the general principles and considerations that should be taken into account when conducting risk assessment of organisms containing engineered gene drives;
- Divergent views on the suitability of existing guidelines and methodologies to be used for the risk assessment of organisms containing engineered gene drives;
- The relevance of modeling as a tool that can provide important information, as well as the limitations of modeling for risk assessment;
- Challenges related to the need to acquire a large amount of data on many aspects to perform a solid science-based risk assessment, and the fact that not all the data needed may be available;
- The challenges associated with the stepwise approach and finding a suitable comparator; and
- Comments raising the importance of taking into account a wide range of stakeholders, including IPLC in the process.

This is not a comprehensive list, and many additional issues were discussed.

I thank you very much for your active participation and for the great sharing of information, which has without doubt provided important considerations not only for the work of the AHTEG, but also for the upcoming discussions on risk assessment under the Protocol.

All the best and thank you very much!

Marja

Dear participants to the forum,

In these last hours of the forum, I would like to follow up on my earlier post [#10128] and address question 3 about the aspects of annex I of decision CP-9/13 defining a process for the identification and prioritization of specific issues regarding risk assessment of LMOs organisms with a view to developing further guidance on risk assessment on the specific issues identified, taking into account annex I; annex I clarifies that the process should include a structured analysis to evaluate whether the specific issues fulfil 4 criteria:
(a) Are identified by Parties as priorities, taking into account the challenges to risk assessment
The report and our submissions to the on-line forum identify the issue of gene drives as a challenge, at least to some Parties, in the context of availability of relevant knowledge and information to support proper risk assessment. My take home message is that the challenges reported are not related to risk assessment methodology per se (this being a structured process with well-defined steps) but to the relevant information that is needed to support this process. Similarly, point (c) of the annex is asking whether engineered gene drives pose challenges to existing risk assessment frameworks, guidance and methodologies. The report pinpoints that interviewees “differentiated between challenges to the risk assessment methodology and challenges relating to obtaining information required to inform the risk assessment”. In my view this point is further reinforced by the postings on our forum. The report also mentions 4 perceived challenges to concepts in RA: the applicability of the comparative and step-wise approaches and the need for predictive modelling as well as the concept of the “receiving environment”.
I fully support the views and analysis of Dr. Werner Schenkel [#10155] that these perceived challenges are mostly of a technical nature but can still be approached within the frameworks and methodologies available, and the views that  the environmental risk assessment (ERA) considerations for organisms with engineered gene drives are not unique to them, e.g.: [#10164], [#10174], [#10182], [#10187], [#10193], [#10207], [#10203].
I would like to add that for a successful completion of this analysis, clarity and alignment is needed on what is understood under “risk assessment methodology”, “risk assessment framework” and “risk assessment guidance” and what purpose each of these serves.

On point (c) of the criteria - challenges in addressing the specific issue are clearly described, I share the views of many that these need to be done on a case by case basis and focused on realistically foreseeable applications (and not on hypothetical ones mined from patent applications or research funding proposals).

Lastly point (d) of annex I is asking whether the specific issues concern LMOs that
(i) Have the potential to cause adverse effects on biodiversity, in particular those that are serious or irreversible, taking into account the urgent need to protect specific aspects of biodiversity, such as an endemic/rare species or a unique habitat or ecosystem, taking into account risks to human health and the value of biological diversity to indigenous peoples and local communities;

While the theoretical eradication of a species would be clearly an irreversible effect, whether it is also a serious and adverse effect on biodiversity needs to be established on a case by case basis for each application, taking into account available alternatives and the intended aim of the application. One post [#10182] reminded us that we should consider the ‘maximum acceptable risk’ in light of benefits of gene drive and other current methods. Similarly, [#10145] pointed out to the risk-acceptance principle advocated by James et al (2018) for gene-drive modified mosquitoes to “do no more harm to human health than wild-type mosquitoes of the same genetic background and no more harm to the ecosystem than other conventional vector control interventions”. I think that these are very relevant considerations for risk assessment performed against the backdrop of public health hazards (e.g.: morbidity and mortality).

The last criterion under point (d) relates to applications that are “already, or are likely to be, commercialized or in use somewhere in the world”.
If there are any gene drives that are ‘close to being released in trials’ this should be described, as recommended in [#10137] for example. If they exist, they would provide the specific examples being requested. If they do not, it might dispel some concern about the urgency for guidance.  Furthermore, specific case studies could be added by requesting such information from developers with sufficient detail to consider the adequacy of existing guidance and RA methodology. This will also address the criticism of not reaching out to developers that were raised by Bob Friedman [#10149] and by Luciana Ambrozevicius [#10182].

With kind regards,
Ana

Dear participants to the forum,

I would like to share with you a list of references (non-exhaustive) arranged according to topic as far as possible (Technology of gene drives and LMOs; Theoretical perspectives and reviews; Gene drive guidance; Gene drive modeling; Gene drive ecology; Risk assessment; Social, legal and ethical considerations; Miscellaneous)

This list contains the references from the Perseus report, alongside relevant references shared in this forum and additional information that supports the point that there is already considerable volume of information on the topic of gene drives and their evaluation.

The rapidly developing field of gene drive research as noted in [#10119] and [#10193] also begs the question when and what applications should be examined in detail in the context of risk assessment challenges, thus attention should be focused on realistically foreseeable applications and these need to be established in consultation with the researchers actively involved in this field.
I would also like to support Hector Quemada [#10129] that reminded us that there are no “laws of segregation” and that describing Mendelian segregation as the norm, and other allele distribution patterns as “distorted” or “aberrant” is mistaken and scientifically flawed. Similar point was made by Ernst Wimmer [#10216] regarding the use of terms like stable “natural genome”.

And to finish on a lighter note, in the context of research and development of engineered gene drives aiming to achieve control of malaria mosquitoes, this Disney animation  from 1943 points to some of the approaches for malaria control of the time https://www.youtube.com/watch?v=I4lYtSEkiLc  (also available via Nature:  https://www.npr.org/sections/goatsandsoda/2018/08/19/638948839/video-the-7-dwarfs-whistle-while-they-work-to-fight-malaria). It also stresses that when we discuss specific applications, such as these aimed at controlling malaria, it is important to consider not only biodiversity issues but human health and wellbeing and public health issues.

Thank you all for the fruitful exchanges,
Kind regards,
Ana

I would like to thank all the participants for their contributions and to Marja for being the moderator, it has been a learning curve for me. As I mentioned in my earlier contribution, being a scientist developing gene drive systems applied to human health it is important to me to learn about the uncertainties the regulators may arise in this kind of discussion. In my earlier comment, I gave my criticism for the CP 9/13 comment. This new contribution is more focused on further discuss some concepts that I have been reading and that to me brings some confusion.

1) evolution and co-evolution [#10177]: I think the way these two concepts were approached is confusing because to assess evolution and co-evolution at the species level and moreover at the ecosystem level will probably take hundreds or thousands of years to have enough evidence that a positive o negative selection pressure was introduced by an engineered gene drive. The comparison made with the evolution of natural selfish-elements and the possible impact of engineered gene drives in evolution cannot be done. Evolution of natural selfish-elements depend on their co-existence with their suppressor and they move amongst different loci over time. These characteristics are modified from the engineered drives because we design them to be target-specific and an in silico analysis is done to predict possible off-targets on no desired populations.
If we want to include evolutionary aspects it will be good to consider the "gene's-eye view" which approaches evolution by natural selection as a process involving two separate entities: replicators (usually the genes) and the vehicles (usually organisms) (Arvid Agren, 2018, PLoS Genetics). Moreover, I think there is a confusion using the term evolution applied to "evolution of resistance to the drive". If resistance is achieved the drive will be extinguished from the population because it is necessary at the molecular level to have available wild-type alleles to transmit the drive. If full resistance is achieved that means that the drive cannot be transmitted to further generations.

2) It is evident that there is a gap in communication between regulators and scientists. I think as a scientist actually working to develop real-world products, it is important we start establishing a more formal dialogue because once a concern is raised, we can work to develop some experimental data to help evaluate the concern. Roger Orellana [#10174] perfectly described this as "efforts that encourage the sharing of knowledge and experience among researchers and regulators". I do not see any scientist (developer) being interviewed to generate documents such as the CP 9/13, it will be nice if the authors can extend an invitation to the scientific community.

3) I strongly agree with Carolina Alduvin [#10189], she pointed out that a risk-benefit analysis might be important to consider. Which will be the cost for the populations at risk (for example, human health) if this technology is not deployed? probably, this will be translated into thousands of deaths.

4) I will take the opportunity to again encourage the case-by-case evaluation as pointed recently by Ernst Winner [#10192], Tim Strabala [#10193] and others.

5) Finally, Keith Hayes [#10195] gave a very comprehensive answer to the Horizontal Gene Transfer (HGT). I would like to add that besides the modeling we can also perform introgression and introduction experiments that can provide empirical data to address this concern. I also support his comment that post-release monitoring strategies can generate the evidence necessary to test the risk predictions.

Thanks again to all the participants for their valuable contributions, it has been an extremely constructive discussion.

Rebeca Carballar, PhD
University of California Irvine
Project Lead Specialist

Hello to the whole group, thanks to Marja and the CBD Secretariat for opening the topic to discussion, thanks to this opportunity I will contribute my opinion on the topic under discussion. My name is Claudia Fabián Navarrete, I am from Mexico and I work in the Ministry of Environment and Natural Resources, on the subject of biosecurity.
Although “genetic drivers” are not being considered as a LMO in some countries, these drivers are being generated to act in specific sites for specific purposes, therefore it is necessary to generate a monitoring scheme to know whether or not there could be any Potential risk to biological diversity, especially to countries such as Mexico and other megadiverse countries or that are centers of origin and genetic diversity, where the conservation and protection of wild species is essential.
In Mexico there is a legal framework that emanates from the Cartagena Protocol and that within that legal framework there is the specific section for risk assessment and management, which was created with general specifications to cover all those LMOs that could be released to the environment, it is also true that genetic engineering has made great strides. In that sense, in Mexico it has been considered whether these “new genetic drivers” enter or not within the framework of the Law of Biosafety of Genetically Modified Organisms (LBOGM), we consider that with the existing legal framework it is sufficient for these “new genetic drivers ”comply with all applicable legal provisions regarding risk assessment and analysis.
It should not be forgotten that this Mexican legal framework encompasses the precautionary principle, following up on the "case by case" and "step by step" basis; It is essential to generate ex post monitoring mechanisms, containment mechanisms and monitoring mechanisms, such as a specific marker and to be able to detect in the field when it is illegally released or there is an escape from this type of organisms that contain “genetic drivers”. Without these traceability mechanisms there will be no control over the LMO that contains “boosters”.
It is a long road yet, therefore, training on the subject should be continued and follow-up on organisms that contain “genetic drivers” in accordance with articles 15 and 16 of the Cartagena Protocol.
It is also important to generate schemes for the cases of importation and exportation of the same, thinking about those countries that are not part, thinking about the investigations and how to contain those organisms that contain “genetic drivers”, among other things.
The definition as it stands today is sufficient, because it is general and specific.
Regards, Claudia.

Dear Marja, dear participants,

My name is Catherine Golstein. I am senior scientific and European affairs officer at HCB (Haut Conseil des biotechnologies / High Council for Biotechnology), a French body in charge of advising the French Government on issues relating to biotechnologies, including LMOs.

I would first like to thank the CBD secretariat for providing the international community with this exceptional opportunity to collectively reflect on gene drive risk assessment. Many thanks to the authors of the commissioned study and to Marja’s enthusiastic and tactful moderation of the forum.

As member of the AHTEG, I look forward to carefully considering the many interesting contributions that have been posted on the forum with a view to preparing the next phase of the work.

A few points:


Scope of the mandate:
It is difficult to anticipate all possible types of synthetic gene drives. Those considered in the commissioned study already encompass a diversity of techniques, objectives and fields of applications, whose heterogeneity sometimes leads to statements that may not be applicable to all cases. At this stage, I would strongly suggest to keep to the scope of GD as has been defined and used so far (see #10122) in sexually reproducing organisms. That said, however, it may be interesting to learn more about potential “GD” in asexual reproducing organisms (mentioned in #10117), which may be addressed in a future work mandate under the Protocol of Cartagena.


HCB’s opinion:
HCB’s first opinion addressing gene drive (HCB, 2017) is mentioned among the documents reviewed as part of the information on stock-taking exercise related to existing guidance (5.2). I would like to thank the authors of the study for considering our work, and clarify that, contrary to what is indicated, the opinion is available in English, at http://www.hautconseildesbiotechnologies.fr/en/avis/avis-relatif-a-lutilisation-moustiques-gm-dans-cadre-lutte-antivectorielle. I reckon HCB’s website is not the most user-friendly. For easier consultation, I post the pdf at the end of this message.

HCB’s opinion is restricted to the context of mosquito vector control. Of relevance to this work, it includes:
- risk assessment of the GD-enabled mosquitoes described by Hammond et al. (2016) and Gantz et al. (2015), the first approach seeking population elimination, and the second, population modification – to make the mosquitoes incapable of pathogen transmission, conducted in comparison with other existing and emerging vector control techniques
- a reflection over risk assessment guidance of LM mosquitoes (including GD-enabled mosquitoes), seeking to answer the following questions, in substance: (1) what do the current applicable legal frameworks and corresponding guidance cover regarding risks associated with LM mosquitoes? (2) Are they sufficient to assess the risks associated with LM mosquitoes, including GD-enabled mosquitoes?

I would like to highlight two key points relative to this work:
- the importance of specific-case analyses rather than broad statements,
- the informative value of comparative analysis, and in particular the comparison with alternative techniques employed for comparable ends.

To identify benefits and limitations (including risks) of applying mosquito control methods using LM mosquitoes (including GD-enabled mosquitoes), we carried out a cross-cutting assessment of:
- different conventional methods (using specific examples of chemical, biological and environmental methods)
- different emerging methods, such as the sterile insect technique using irradiated mosquitoes (classical SIT), techniques using Wolbachia-infected mosquitoes for population reduction (Incompatible Insect Technique) or population modification (Pathogen Interference propagation), Oxitec RIDL approach (considering the example of Ae. aegypti OX513A line, developed by Phuc et al., 2007) as well as CRISPR-Cas9 gene-drive based methods for population reduction (considering the specific working example developed by Hammond et al., 2016) or population modification (considering the specific working example developed by Gantz et al., 2015).

Questions addressed in a comparative manner included:
1) ability to achieve specific goals : population reduction (or suppression) versus population modification (or alteration), at different scales and degrees (e.g. local elimination vs species eradication)
2) efficacy, durability and sustainability (including questions of different types of resistance evolution and genetic drift, inherent limitations of efficacy in space and time, constraints related to initial field conditions such as population density, etc.)
3) technical constraints (including requirements for mosquito mass rearing conditions, separation of males and females, etc.)
4) risks to the environment and public health (including broad questions such as the degree of specificity of the different techniques and associated consequences on public health and non target organisms, potential risks associated with resistance evolution and genetic drift, and more specific questions for different sets of techniques -- for gene drive: the additional questions of CRISPR-Cas collateral (off-target) mutagenesis activity and their case-specific consequences, of GD-cassette potential horizontal or vertical transfer and its consequences, of species eradication, etc.).

Case-specific examples of thorough risk assessments of different types of gene drives were instrumental in addressing the questions of corresponding risk assessment guidance for LM mosquitoes. To enlarge the scope to the scale proposed by the current exercise under the Cartagena Protocol, other case-specific examples may be useful to consider.


Evolution of resistance:
GD-enabled LMOs face the problem of resistance evolution in the target populations. GD techniques are being improved, however, to minimize the probability of resistance evolution, making it possible for the drive to spread as theoretically expected from the technique. In contrast, other GD techniques are being engineered to control the spread of the techniques on purpose.

When discussing resistance evolution in the document (including the executive summary), the following issues should be clearly distinguished:
- Probability of resistance evolution
- Impact on gene drive efficacy
- Impact in terms of risks
- Potential use as a confinement strategy


Terminology of gene drive seeking “population replacement”:
We strongly recommend using the terminology of “population modification” rather than “population replacement”.

In such strategies, gene drive allows the spread of a gene within a pre-existing population. It does not replace a pre-existing population with a new population, as the terminology “replacement” would suggest.

This is all the more confusing as, in the gene drive strategy seeking population suppression, one of the risks that must be assessed concerns the potential unintended replacement of the target species by another species that may be more harmful to health or the environment (e.g. another vector species).


Off-target:
We recommend to avoid the confusion between “off-target mutations” and “off-target effects”. It would be clearer to distinguish:
- Off-target (or collateral) mutagenesis activity of the CRISPR-Cas system, that may lead to off-target mutations
- Impact of off-target mutations in terms of risks


Effect on biodiversity :
In addressing effects on biodiversity, it is of interest to keep in mind that a theoretical LMO provided with a resistance-proof GD technique designed for population elimination may reach its objective of eliminating a (for example) mosquito population in a delimited area, or even eradicating that particular mosquito species and potential inter-fertile relatives. In itself, this would constitute a reduction in biodiversity since a species (and potential inter-fertile species) would be locally eliminated or globally eradicated. But this was the very objective sought by the gene drive, here in order to control disease, as an alternative approach to using other techniques bearing their own risks.

Subsequent risks on the ecosystem following species elimination would depend on the local ecological role of these species, and the possibility that other species may fulfil it. This has to be assessed case by case and region by region (which should be better emphasized in the document) considering sympatric mosquito species. Of note, there are hundreds of sympatric mosquito species in the tropics. Gene drive will affect one of them (and potential inter-fertile species).

Given that an ecological niche has been vacated, another type of risks following species elimination (that could be better emphasized in the study) concerns unintended replacement of the eliminated population. Replacement by a mosquito population with a vector competence of greater concern than that of the eliminated population would constitute an adverse effect on human health.


Gantz, V.M., Jasinskiene, N., Tatarenkova, O., Fazekas, A., Macias, V.M., Bier, E., and James, A.A. (2015). Highly efficient Cas9-mediated gene drive for population modification of the malaria vector mosquito Anopheles stephensi. Proc Natl Acad Sci USA 112, E6736-E6743.

Hammond, A., Galizi, R., Kyrou, K., Simoni, A., Siniscalchi, C., Katsanos, D., Gribble, M., Baker, D., Marois, E., Russell, S., et al. (2016). A CRISPR-Cas9 gene drive system targeting female reproduction in the malaria mosquito vector Anopheles gambiae. Nat Biotechnol 34, 78-83.

HCB Scientific Committee (2017). Scientific Opinion of the High Council for Biotechnology concerning use of genetically modified mosquitoes for vector control in response to the referral of 12 October 2015 (Ref. HCB-2017.06.07). (Paris, HCB), 142 pp. Available online: http://www.hautconseildesbiotechnologies.fr.


Kind regards,

Catherine

Thanks to Marja for steering us in our discussions, and to everyone for their thoughtful contributions. (as an aside, Marja opined about snow in Helsinki – I wish we had some in Australia right now!).
My name is Peter Thygesen, and I have worked in the Australian Office of the Gene Technology Regulator as a risk assessor and regulator for about 20 years. I have been following this discussion closely but am only now joining with some comments.
The Perseus report provides a good basis for discussion, and as with all such reports it could be improved, and I agree with many of the suggestions for expansion / inclusion, especially on providing a more elaborated picture of the state of development of engineered meiotic gene drives. Whether that is practical or useful in the timeframe of the scheduled activities in the lead up to the next COP/MOP is another matter.
I agree with the suggestion that the report, and these deliberations in general, would benefit from input from those working on the development of gene drives. The field is developing fast and from following the literature on gene drives it seems clear that many (perhaps all?) developers are in fact very aware of the concerns over the deployment of such organisms and that significant attention is being paid to design features of drives with a view to minimising risks and/or providing greater control over the operation of a drive.
Input from such experts might shed light on the various questions raised by Jack Heinemann, Ricarda Steinbrecher and others about horizontal gene transfer, evolution and ‘genetic engineering machinery in the environment'. My feeling is that these issues are not the main issues that need to be considered in the ERA of a GDO, and as others have noted these issues are not unique to GDOs.
I strongly support the view expressed by many contributors that a focus on engineered meiotic gene drives is the relevant and appropriate one, as these are the ones being actively researched / developed and proposed for various applications/uses in the environment. It would be appropriate for the report to acknowledge that there are other selfish genetic elements and that some of these might also the subject of investigation for potential applications, but this should not be the focus.
I also strongly support the points made by many contributors that the approach to and conduct of risk assessments for LMOs with engineered gene drives can be informed by the experience of risk assessments for other environmental interventions, and that the main issues for consideration are not unique to GDOs. I welcome Luciana Ambrozevicius’ reminding us of the OECD principle of biology of the parent organism, the nature and effects of the introduced trait(s), the receiving environment and the intended use, and their interactions, and that this applies to GDOs as for other LMOs. I think it is important to make a clear distinction between the ‘drive’ mechanisms per se and the cargo that the drive would propagate through the target population. Much of the popular debate about GDOs has focussed on gene drives that would be aimed at removing or reducing a population and the related ‘extinction threat’, however other phenotypes may be conferred depending on the cargo.
A number of contributors have spoken about protection goals, and I would make an observation that is not unique to GDOs or LMOs. Clarity about protection goals is very important for risk assessors, however these are very often only expressed in general terms, and protection goals are based on societal values. Recent papers by Alan Raybould have note this issue in relation to LMOs, however this issue also arises in other environmental assessments. For many of the ‘use scenarios’ being proposed for gene drives to date the GDO is intended to address a pre-existing risk to either human health or the environment, and where there is some societal imperative or desire to address that risk, by some means. This is somewhat in contradistinction to our experience with ‘product’ LMO such as GM crops, where the regulatory decision is about whether that product should enter into the market. Thus, the decision making in relation to GDOs that are intended to redress human health or environmental risks may be informed by assessments and decision making for other environmental interventions (biological control of weeds/pests, control of invasive alien species) where there is comparative assessment of the current situation and controls vs other possible interventions. It would often be the case that mitigation or amelioration of an existing risk or harm would be characterised as ‘benefits’, however I would suggest that sticking to risk language would be better, and if the assessment and decision making framework accommodates it to focus on ‘delta risk’ – what is the quotient and nature of risk/harm from any proposed interventions. This would avoid sliding into the monetary and political debates that have tended to characterise discussion of benefits for commercial product LMOs, such as GM crops.
I have been following the gene drive debate so I have enjoyed following the discussion, and certainly there are a number of scientific points and references I will follow up on.
Best wishes
Peter

Dear Marja, dear all.

Apologies, Marja, as somehow we all seem to fail to address what you are asking us for. I am actually not quite sure any more what it would be that would be real helpful for informing the application of annex I. 

Thank you for giving us one more chance for a last intervention, I will give it a try.

For me the crucial points are the entire new quality of LMOs, that GDOs bring, in particular the CRISPR-based gene drives. I heard other scientists describe this new quality as “placing the lab into the organisms” or “taking the lab into the field”. As different organisms of the same species will have different genetic backgrounds and compositions, and as HEG-based gene drives undertake the modification in the wild with potentially numerous outcomes, there is a completely different category of LMO that needs to be dealt with, and there is nothing that is anywhere near this new quality that we could draw on. Previous LMOs were lab made and largely uniform and could be characterised as such. HEG-based gene drives are miles apart from that.

The term “familiarity” brought up in [#10127], picked up by our moderator in her first summary and later in [#10170] again, is not suitable in this context, as “drawing on familiarity” with insect vectors / pest control strategies “when framing the risk assessment of gene drives” gives the wrong notion of familiarity. There is no familiarity. SIT is very different. And so is the use of Wolbachia – as already commented on by others above. It is important that the report makes clear distinctions.

Another point that requires urgent clarification is the problematic use of the term “reversibility”. The changes introduced into wild populations by the deployment of gene drives are by large not reversible. The population can not be returned to what it was before. The notion for “reversible” to mean “the ability to replace an existing gene drive system with another one”, as currently suggested and defined in the draft report (page 19, table 4 and line 22), is counter to any common understanding of reversibility. This needs urgent attention. If the question is whether a gene drive might be stopable, for example by sending a second gene drive after it – well, there are theoretical or hypothetical concepts looking at such “gene drive catchers” or “immunisers”. There are however no proofs of concept nor is would such second line gene drives be able to “repair” and undo what the first gene drive did. Rather, they would insert further genetic changes into organisms of the natural/wild population. We have written to this point in chapter one of our 2019 report (mentioned previously). This item, I believe, is important for informing the application of annex I.

Another area of misunderstandings in need of clarification is the spectrum of gene drives from replacement (eg underdominance), modification (eg via payload HEG-drives), suppression and eradication. A point to the use of terms was also made in [#10199].

Finally, and this has a crucial importance despite being outside the common practice of risk assessment – namely the dimension of ethics. The draft report is neither adressing ethics, nor social issues nor governance issues, which are of major concern. Equally, the concerns and rights of and the implications for indigenous peoples are not reflected sufficently in this report and vastly under-represented. May I for the first three points (ethics, social issues and governance) refer once more to our CSS/ENSSER/FGS report, namely to chapters 3, 4 and 5, which are focussing especially on these issues (https://genedrives.ch/wp-content/uploads/2019/10/Gene-Drives-Book-WEB.pdf).

I am looking forward to the outcome of our explorations and discussions. It has been a real pleasure having been part to it.

With best regards,

Ricarda