Review paper "EFSA Electronic Supplemental Material

Review paper “EFSA’s scientific activities and achievements on the risk assessment of genetically modified

organisms (GMOs) during its first decade of existence – looking back and ahead” by Devos et al.

Electronic Supplemental Material

GMO market approval applications

GM microorganisms

Approximately 39 GM microorganism applications have been assessed, or are under assessment by EFSA. Two of these applications are for feed materials derived from biomass of GM bacteria (Escherichia coli) submitted under the GM food and feed Regulation. The rest are for products such as enzymes, vitamins or amino acids, obtained by fermentation of GM microorganisms in which no material derived from the production strain is present at detectable levels, or for biodegradable polymers intended to be used in food contact packaging material which are produced by fermentation of GM microorganisms producing the monomers. (Note: these products are considered as made with GMOs by EU legislation, in contrast to products made from GMOs, in which the GMO or rest of its cells are still present in the product.) The vast majority of products made with GM microorganisms correspond to feed enzymes, vitamins or amino acids, but also food enzymes are currently under investigation in the frame of the Regulations (EC) No 1332/2008 (on food enzymes), 1333/2008 (on food additives), 1334/2008 (on food flavourings) and 1831/2003 (on feed additives). Production organisms encompass GM bacteria (E. coli but also Corynebacterium glutamicum, Bacillus subtilis and B. licheniformis), and GM fungi, including yeast (Trichoderma reesei, Aspergillus niger, A. oryzae, Penicillium funiculosum, Saccharomyces cerevisiae, Pichia pastoris). Although these products are not considered GM food and feed and therefore are out of the scope of the GM food and feed Regulation, they still have to undergo a risk assessment prior to authorisation in compliance with existing EU legislation specific for each type of product. In the course of the assessment of GM microorganism applications, EFSA takes into account the characteristics of the production of the GM microorganism and the potential effects that its genetic modification can exert on the safety of the final product (e.g., production of toxic compounds, presence of antibiotic resistance genes). Key aspects considered include the characterisation of the parental microorganism and the inserted sequence(s), the stability of the GM microorganism, and the possible presence of GM microorganism cells or recombinant DNA in the final product. For the assessment of feed applications, EFSA’s FEEDAP (Additives and Products or Substances used in Animal Feed) Panel established a working group on GM microorganisms. Food applications (enzymes and polymers) are risk assessed by the corresponding working groups of EFSA’s CEF (Food Contact Materials, Enzymes, Flavourings and Processing Aids) Panel which involve experts on GM microorganisms. The several working groups evaluate data on the molecular characterisation and environmental risk assessment of the production strains. From 2013, the number of applications for food and feed products made with GM microorganisms is expected to increase significantly owing to the entry into force of the EU legislation on food enzymes, which requires that all products already on the market undergo a safety assessment by EFSA.

2

Guidelines for the risk assessment and monitoring of GMOs Risk assessment of GM plants and derived food and feed

Molecular characterisation Data included in the genetic modification process part – apart from the technical details of the modification process itself – help to determine if the nature of the donor organism(s) or the nucleic acid sequence(s) may trigger a safety issue. In this respect, applicants are asked to provide a full description of the sequences intended to be inserted, as well as information on the history of their safe use. Information about the GM plant itself focuses on issues inherent to the particular molecular build-up of the plant. Applicants need to provide information on the structure of the insert(s) in the GM plant, including the complete nucleotide sequence of each insertion locus(i). Although so far all GM plants assessed by EFSA had a nuclear insertion, data should be given on the subcellular location of the insert(s) (e.g., nuclear, plastidic). To evaluate potential unintended effects of the insertion at the molecular level, applicants should provide a complete bioinformatic characterisation of the insertion site(s) and the insert(s). Such bioinformatic package consists of studies on the flanking regions of each insertion site with the aim of identifying interruptions of known genes, and analysis of all putative open reading frames present within the insert and spanning the junction sites to investigate if any of those putative potential protein show similarity to known toxins or allergens. Results of the bioinformatic analyses are strongly affected by the bioinformatic databases used. Therefore, EFSA advocates the use of up-to-date databases when performing such analyses. Depending on the results, further studies may be required (bioinformatic or other, such as expression studies) to complete the risk assessment. In addition to the above listed aspects of the molecular characterisation, the updated guidelines for the risk assessment of GM plants and derived food and feed (EFSA 2011a) consider aspects that are more related to the phenotypic properties of the GM plant. These include information on the expression of the inserted sequence and data on genetic/phenotypic stability. Expression is interpreted in a broad sense; apart from the most common case where new proteins are expressed, traits based on RNA interference (RNAi) technology are also considered in EFSA’s updated guidelines. Expression data should derive from plants grown under conditions representative of typical cultivation practices and from plant parts relevant to the scope of the GM plant application. On a case-by-case basis, complementary data may be required on developmental expression, tissue-specific expression or on effects of specific treatments, such as herbicides. In terms of genetic and phenotypic stability, applicants need to demonstrate that the trait (phenotype) and the integrity of the insert (genotype) are maintained over five generations or vegetative cycles. Molecular characterisation of GM plants containing a combination of transformation events (stacks) mainly focuses on possible interactions between the events. The risk assessment of stacks is therefore based on: (1) demonstrating that the integrity of the individual inserts is preserved in the stack as compared to the single events; and (2) comparing expression levels in the stack with those in the respective single events.

3

Comparative analysis Compositional data to support the comparative analysis of GM plants are generated in field trials carried out in several locations (EFSA 2011a). The minimum replication of field trials over sites and years required to ensure sufficient statistical power are specified in the updated guidelines for the risk assessment of GM plants and derived food and feed (EFSA 2011a); at least eight field sites, chosen to be representative of the range of likely receiving environments where the plant will be grown, must be used with at least four replicates per site in a complete randomised block design. If the GM plant contains a trait for herbicide tolerance, blocks sprayed with the target herbicide or with the conventional herbicide must be included, in addition to a near-isogenic non-GM counterpart sprayed with the conventional herbicide. To reliably estimate the natural variation of analytes present in the crop, at least six non-GM reference varieties must be included with at least three varieties being represented at each field site. The testing of equivalence is designed to focus exclusively on results from the study-specific non-GM reference varieties. Values derived from crop composition databases or literature values are not used in the equivalence testing. Toxicology The toxicological assessment begins with an extensive biochemical characterisation of the newly expressed proteins (EFSA 2011a). This assessment requires information on the molecular and biochemical properties of the newly expressed proteins, an up-to-date search for homology to proteins known to be toxic, information on the stability of the proteins under relevant processing and storage conditions for the food and feed derived from the GM plant, data concerning the resistance of the newly expressed proteins to proteolytic enzymes (e.g., pepsin), and repeated dose toxicity studies using laboratory animals, unless reliable information demonstrating the safety of the newly expressed proteins (including their mode of action) is available and it is demonstrated that the proteins are not structurally and functionally related to proteins known to adversely affect human or animal health. The extensive biochemical characterisation of the newly expressed proteins is then followed by an assessment of the toxic potency of new constituents other than proteins, which may arise as a consequence of the genetic modification. This assessment may require toxicological testing, unless there is a documented history of safe use and consumption as food and/or feed for the new constituents. Nutrition Food consumption data are necessary to complete the risk assessment of GM plants with changed nutritional or toxicological compounds. Depending on the data source and data collection methodology, the estimated intake for a given food product can differ, sometimes even significantly. To circumvent this limitation, EFSA has consolidated 27 national dietary surveys into the ‘comprehensive European food consumption’ database

(http://www.efsa.europa.eu/en/datexfoodcdb/datexfooddb.htm). Food categories that are of interest for the food and feed safety assessment of GM plants and derived food products were recently extracted. Both summary statistics of chronic and acute dietary consumption datasheets are now publically available at http://www.efsa.europa.eu/en/datexfooddb/datexfooddbspecificdata.htm for food products derived from maize, potato, rapeseed, rice, soybean and sugar beet.

4

Environmental risk assessment of GM plants

Risk assessment principles In line with a number of internationally agreed risk assessment principles, the updated guidelines require environmental risk assessments: (1) to use quantitative information where available; (2) to use a comparative approach whereby the level of risk is estimated through comparison with an appropriately selected comparator and its associated farm management and cropping practices; (3) to be case-specific; and (4) to be iterative by examining previous conclusions in the light of new information (EFSA 2010b). As outlined in Directive 2001/18/EC, applicants are required to conduct the environmental risk assessment in six successive steps, consisting of: (1) problem formulation as a critical first step; (2) hazard characterisation that examines potential hazards and the seriousness of potential harm; (3) exposure characterisation that considers levels and the likelihood of exposure and thus how likely it is that harm occurs; (4) integrative risk characterisation in which the magnitude and likelihood of harm are integrated to estimate the level of risk; (5) mitigation of the identified risks to reduce an identified risk to a level of no concern; and (6) evaluation of the overall risk based on proposed risk mitigation measures (EFSA 2010b). Problem formulation is given a central role in environmental risk assessment, as it enables a structured, logical approach to detecting potential risks and scientific uncertainties by summarising existing scientific knowledge and explicitly stating the assumptions and principles underlying the risk assessment. Problem formulation involves: the identification of characteristics of the GM plant capable of causing potential adverse effects (hazards) and pathways of exposure through which the GM plant may adversely affect human and animal health or the environment; the definition of assessment endpoints, which are explicit and unambiguous targets for protection extracted from legislation and public policy goals; and outlining specific hypotheses to guide the generation and evaluation of data in the subsequent successive risk assessment steps. This process also requires the development of a methodology – through a conceptual model and analysis plan – that will help to direct the risk characterisation and to produce information that will be relevant for regulatory decision-making (Raybould 2006; Wolt et al. 2010; Gray 2012). Information considered in problem formulation includes published scientific literature, expert opinions, research data and relevant data derived from molecular, compositional and agronomic/phenotypic analyses performed during GM plant development. Protection goals The six steps followed to characterise the overall environmental risk of a GMO is to be applied to each of the eight areas of risk set by Directive 2001/18/EC (see above). For each area of risk, applicants should specify which protection goals are applicable to their environmental risk assessment, and what assessment and measurement endpoints they use (EFSA 2010b). Protection goals are general concepts that are defined in broad terms by risk managers and regulators, but are often too vague to be scientifically useful for environmental risk assessment and regulatory decision-making (Evans et al. 2006; Raybould 2012). To be useful, it is important that these general and broadly formulated protection goals are translated into concise and concrete measurable endpoints and scientifically testable hypotheses.

5

The operationalisation of protection goals for the environmental risk assessment requires specifying: the unit of protection; the harmful effect; and both the spatial and the temporal scales relevant for the ecological entity and its attribute/function to be protected (Nienstedt et al. 2012; Sanvido et al. 2012; Garcia-Alonso and Raybould submitted). The harmful effect describes above what level a difference between the GM plant and its appropriately selected comparator may lead to harm and would be considered a disturbance in environmental quality. For the conservation of biodiversity, a relevant decrease in abundance of protected or valued species can be seen as a harmful change. Similarly, for ecosystem services, a relevant disturbance in ecological function can be seen as a harmful change. The spatial and temporal scales are the habitats in which, and the period during which, environmental quality should be preserved, respectively. Persistence (weediness) and invasiveness Some environmental concerns about GM plants relate to the potential persistence (weediness) or invasiveness of the plant itself, or of its sexually compatible relatives, as a result of vertical gene flow within either agricultural or other production systems, or semi-natural and natural habitats. EFSA proposed a staged framework which specifies data requirements of differing levels (EFSA 2010b). The purpose of the staged approach is to ensure that relevant information is supplied to test hypotheses formulated in problem formulation, and that information requirements remain proportionate to the potential risk. Ten questions broken down into four stages, outline the issues to be addressed to estimate the likelihood of occurrence of adverse effects in ruderal, semi-natural and natural environments. Whether information is required for all stages or only for specific stages will depend upon the trait(s), plant species, the intended use, receiving environments under consideration, and the conclusions drawn from lower stages. Information required for testing the hypotheses formulated in problem formulation can be species-, trait- or event-specific. Background information is always required at the outset, describing the biology of the parental species including reproductive biology, survival, dispersal and cultivation characteristics in different environments. In addition, sexual compatibility with other cultivated or wild plants occurring in the EU, and the biology and ecology of these relatives should also be considered. It is then considered whether the GM plant is similar to the untransformed counterpart with respect to the characteristics identified as being important for persistence and invasiveness in the environment. If data indicate changes in these characteristics, those changes would need to be assessed for their potential to alter the likelihood of persistence and invasiveness of the GM plant.

6

Non-target organisms Issues to which special attention was paid in the guidelines for the evaluation of possible adverse effects of the cultivation of GM plants on non-target organisms were: (1) the consideration of functional biodiversity as a relevant protection goal; (2) criteria for the selection of non-target organisms for risk assessment and, if required, for testing purposes (focal species); and (3) approaches to assess the consequences of intended and unintended changes in GM plants on non-target organisms (EFSA 2010a,b).

(1) The guidelines distinguish the protection of biodiversity (biodiversity conservation) and that of the ecological and anthropocentric functions provided by non-target organisms (termed hereafter as ecosystem services). Focus is put on species that are representative for valued ecosystem services in an agricultural context (e.g., natural enemies for pest regulation, bees for pollination) and species of conservation concern (e.g., rare and protected species, or species of aesthetic or cultural value). Valued ecosystem services to preserve in an agricultural context are pest regulation, pollination, decomposition of organic matter, soil nutrient cycling, soil structure, water regulation and purification, and cultural services (such as aesthetic value) (Moonen and Bàrberi 2008).

(2) For the selection of focal species for risk assessment, and if required, for testing purposes, a four-step approach combining the strengths of two existing species selection approaches – the ecological and ecotoxicological approach – is proposed (Andow and Hilbeck 2004; Romeis et al. 2008). In the first two steps, valued ecosystem services provided by the plant ecosystem are to be identified, along with the non-target organisms performing those services. In the next two steps, non-target organisms are to be prioritised based first on ecological criteria (such as species’

exposure to the GM plant, abundance, feeding habits, relevance to ecosystem functioning, species vulnerability, sensitivity to trait), and then on practical criteria (such as species’ availability, testability and conservation status). Whether new data should be generated and hence a specific species be tested depends on the outcomes of problem formulation. Useful information of sufficient quality may already exist in the scientific literature from previously conducted studies, and may be used to corroborate the formulated hypothesis under consideration. In these situations, no additional testing may be required to acquire the data. To facilitate the identification of most abundant and widespread arthropod species, as well as the contribution of taxa to particular valued ecosystem services, EFSA commissioned the compilation of a database on arthropods found in different crops in the EU in 2010. The fauna database provides a detailed overview of the composition of the arthropod fauna and the abundance of species found in maize, oilseed rape, potato, sugar/fodder beet, soybean, cotton and rice, and in different geographic regions across Europe. Species attributes and abundance data, retrieved from over 1000 publications, give ecological information for 3030 arthropod species and 14762 abundance records from 31 European countries (Meissle et al. 2012). To ensure that the fauna database remains a useful tool over time for EU risk assessors from the public and private sector, scientists and risk managers, it is planned to be updated by regularly including data derived from newly published scientific literature. In addition to updating the fauna database, its usefulness is planned to be increased further by

7

expanding the database to small grain cereals, and potentially field margin data (de Lange et al. 2012).

(3) For the assessment of potential adverse effects on non-target organisms caused by the intended genetic modification (e.g., the expression of the Bt-protein) in the GM plant, a tiered testing approach needs to be followed that progresses from highly controlled worst-case exposure lower-tier studies in the laboratory over greenhouse studies to more realistic but less controlled higher-tier studies in the field. Moving to a higher-tier is only considered relevant if adverse effects are detected, or if unacceptable scientific uncertainty remains. In some cases, higher-tier studies may be conducted at an initial stage when lower-tier studies are not possible or meaningful. The guidelines also require an assessment of potential adverse effects arising from unintended changes in the GM plant, which go beyond the primary objectives of the genetic modification. For the assessment of such effects, a weight-of-evidence approach is proposed that relies on in planta (event-specific) data. Such data are generated during the comparative analysis of the GM plant with its appropriately selected comparator at molecular, compositional and agronomic/phenotypic levels. In addition, the guidelines recommend the detection of unintended changes in the GM plant through comparisons of interactions of the GM plant and its comparator with non-target organisms. Field studies or laboratory studies with in planta material are considered suitable means to retrieve data on interactions of the GM plant and its comparator with non-target organisms. In these studies, one focal species of each relevant functional group needs to be tested. Should differences in interactions of the GM plant and its comparator with the non-target organisms be identified, their biological relevance is to be assessed further.

Farm management and cropping practices For the assessment of the potential adverse effects of specific cultivation, management and harvesting techniques, the use of scenario-analyses is recommended to assess under which situations the specific farm management and cropping practices accompanying the adoption of GM plants may lead to greater, similar or lower adverse environmental effects on farmland biodiversity than the current practices applied in conventional cropping systems, and to identify for which situations mitigation and monitoring may be required (EFSA 2012a). Scenario-analysis is considered a suitable forecasting tool, as it enables one to account for the scientific uncertainty associated with the assessment of complex, diverse and dynamic cropping systems. Consistent with EFSA (2010b), applicants should explore various adoption scenarios, and estimate their impact relative to baseline scenarios representing typical conventional cropping systems. In case of GM herbicide tolerant (HT) plants, one adoption scenario is a ‘substitution’ scenario, in which the conventional plant (with its specific management practices) is substituted by the GMHT plant with its specific herbicide management, but without any other changes in other management practices. Another scenario is a ‘worst-case’ one, which describes the effects on receiving environments of repeated,

large-scale, and intensive management using the GMHT plant and its adapted management practices, including the adoption of no- or reduced-tillage systems. This scenario considers the use of the GMHT plant that would lead to maximum environmental exposure (e.g., high adoption and continuous cultivation, maximum dose and number of herbicide applications). A third scenario is a ‘best-case’ scenario in which the GMHT plant is cultivated using an

8

approach which embeds its management into practices consistent with the goals of sustainable agriculture (as defined in, for example, Directive 2009/128/EC). Cross-cutting considerations The updated guidelines provide criteria for the identification of EU geographical regions where GM plants may be released and for the selection of appropriate techniques for assessing potential long-term effects of GM plants, and give specific requirements for the assessment of stacked events (EFSA 2010b). Special attention was paid to the experimental design and statistical approaches for analysis of ecological field studies, as poorly-replicated field studies may not be useful to resolve remaining scientific uncertainty due to their limited statistical power to detect biologically significant differences, especially if data are generated without a clearly framed, scientifically robust, hypothesis (Perry et al. 2009). It is therefore recommended that the nature and size of relevant biological changes or differences are defined by applicants before a field study is initiated. The size of such changes should be used to design the study in a way that it has sufficient statistical power to detect the anticipated effect if it truly occurred (Perry et al. 2009; EFSA 2010a,b). In field studies with non-target organisms, a power analysis is usually performed retrospectively instead of prospectively, as anticipating the abundance of non-target organisms in the field may not always be possible for all species (Albajes et al. 2012; Comas et al. 2013). However, the updated guidelines point to the feasibility to perform a prospective power analysis for those non-target organisms that are the most abundant in the field (EFSA 2010a,b). Because conventional cropping systems are not characterised as having a history of safe use, it is not recommended to use reference varieties to estimate natural variation in ecological field studies. Instead, limits of concern are defined ab initio, from the minimum degree of difference that might be expected to lead to environmental harm. These limits of concern represent the level of environmental change that constitutes harm. Limits of concern are directly related to the type of studies that are to be performed either in the laboratory or in the (semi-)field. For laboratory studies, limits of concern are conservative trigger values which, if exceeded, will indicate potential risks and the need for exposure assessments and determination of field scale effects. For field studies, the lower limit, which corresponds for example to a decrease in the abundance of a particular species in the presence of the GM plant relative to that for its comparator, will usually be defined by the threshold effect deemed to be of just sufficient magnitude to cause environmental harm. Post-market environmental monitoring of GM plants

The updated guidelines describe various sources of information that can deliver general surveillance data and make recommendations on how their use, design and analysis can be best optimised (EFSA 2011c). General surveillance data can originate from farmer questionnaires, existing monitoring/surveillance networks (e.g., plant health surveys, soil surveys, ecological and environmental observations), scientific literature, industry stewardship programs and alert issues (Sanvido et al. 2005, 2011).

- Farmer questionnaires form a useful part of general surveillance, as they enable the reporting of any observations of effects linked with GM plant cultivation on farm, which presents in practical terms the smallest unit where monitoring characteristics can be observed and where the influencing factors, especially farm management and

9

cropping practices, can be assumed to be comparable to those of conventional plants. Questionnaires, directed at farms or production systems where GM plants are grown and utilised, use first-hand observations and rely on farmers’ knowledge and

experience of their local agricultural environments, comparative crop performance and other factors that may influence events on their land (Schmidt et al. 2008; Wilhelm et al. 2010). To reduce interviewer bias, the updated guidelines recommend that impartial and standardised interviews are carried out by independent parties and that effective quality and auditing procedures are implemented.

- With farmer questionnaires, monitoring focuses mainly on the cultivation area of the GM plant and its surroundings, which is relevant to protection goals such as sustainable agriculture, soil function, or plant health. However, aspects of biodiversity are only addressed indirectly (such as the adoption of conservation/no-till practices, rotation regimes, biological control failures) and may not be sufficiently resolved. Therefore, additional sources of information should be considered in the frame of general surveillance (Sanvido et al. 2011). Directive 2001/18/EC and Council Decision 2002/811/EC propose to make use of established routine monitoring/surveillance networks. EU Member States have various networks in place – some of which have a long history of data collection – that may be helpful in the context of general surveillance of GM plants. The networks involved in routine monitoring/surveillance offer recognised expertise in a specific domain and have the tools to capture information on important environmental aspects over a large geographical area. However, existing monitoring/surveillance networks do not necessarily provide data of relevance to monitor the impacts of GM plant cultivation (Bühler 2006; Mönkemeyer et al. 2006; Graef et al. 2008; Sanvido et al. 2011; Smit et al. 2012). Therefore, the updated guidelines recommend the use of criteria (i.e., relevance in terms of environmental protection goals, monitoring subject, specificity and methodology, reporting and data accessibility) to identify and select existing monitoring/surveillance networks suitable for the purposes of monitoring. EFSA has been mandated by the European Commission to consider to which extent existing monitoring/surveillance networks can effectively contribute to general surveillance of GM plants, and to develop a set of criteria to assess their suitability. In response to this request, an external scientific report to be delivered in 2014 was commissioned. This report will review statistical methods and data requirements to support PMEM, and will give recommendations on the use of existing monitoring/surveillance networks to support general surveillance in Europe.

- An important part of monitoring involves the annual consideration, reviewing and reporting of findings reported in peer-reviewed and grey scientific literature. This information enables one to assess whether the initial risk assessment conclusions and risk management recommendations remain valid and applicable in the light of new scientific information. The updated guidelines request applicants to be specific about the search strategies followed to retrieve relevant publications, and subsequently, to specify the criteria used to assess the relevance of the identified publications.

- Industry stewardship programs include a wide range of activities during commercialisation, and aim to facilitate compliance with risk management conditions, to ensure that the products are used responsibly in a way that has similar or less environmental impacts compared to conventional crop cultivation, and to ensure the sustainable use of the technology. Stewardship activities also provide farmers and third parties with contact details and directions on where they can report any

10

unexpected findings related to the GM crop and its cultivation, and are therefore relevant to general surveillance.

Risk assessment of GM microorganisms and derived products intended for food and feed

use

Since the publication of the first guidelines for the risk assessment of GM microorganisms and their derived products intended for food and feed use in 2006 (EFSA 2006), EFSA has gained significant experience in assessing applications for products obtained with the use of, or derived from, GM microorganisms. This, together with the input received from different stakeholders and EU Member States, and the entry into force of legislation on food enzymes, motivated the need to update the guidelines. The updated guidelines provide a complete and comprehensive tool for the efficient risk assessment of GM microorganisms and their products for food and/or feed use (EFSA 2011b). Owing to the diversity of GM microorganisms and their products, the guidelines propose a categorisation, depending on the nature of the product and the extent of the scientific information required for its evaluation (see also Aguilera et al. 2013). Categories 1 and 2 correspond to products that do not contain GM cells, or recombinant genes. Products made with GM microorganisms (such as most enzymes and additives) fall mainly in one of these categories, depending on the purity. For these products the assessment of the product itself is not covered in the updated guidelines, as they are addressed in other available EFSA guidelines (see EFSA 2011b and references therein). Category 3 corresponds to products in which viable GM microorganisms are not present, but which still contain recombinant genes. Products belonging to Category 4 are those consisting or containing viable GM microorganisms, i.e., in which the production microorganism remains alive in the food or feed. For these products, the updated guidelines provide a detailed description on how to handle the risk assessment in relation to human and animal health, as well as the environmental risk assessment and the post-market monitoring. The guidelines provide detailed indications on how to allocate a product into the correct category, including the scientific evidence to be provided. The assessment starts with a full characterisation of the GM microorganism, based on information on the strain receiving the genetic modification including taxonomy, biology, safety aspects, and previous uses in food/feed production. The genetic modification has to be described in detail, and data must be provided on the structure, function, sequence and stability of the inserted DNA fragment. The possible effects of the genetic modification on the safety of the GM strain should also be assessed (e.g., pathogenicity, production of toxic compounds, antibiotic resistance). After characterising the GM microorganism, the production process should be fully documented, particularly how the production strain is inactivated or removed (if this is the case), and whether recombinant genes remain in the final product. This is essential to allocate the product into the correct category. The guidelines fully cover the assessment of products made from, containing, or consisting of GM microorganisms. Aspects considered include composition, toxicology, allergenicity, and nutrition. Like for GM plants, the comparative assessment is the basic strategy followed. Depending on the nature of the product, the comparator may be the microorganism before being genetically modified, or the food or feed produced with the traditional microorganism. The concept of “Qualified

Presumption of Safety” (QPS) can be used as a justification for the safety of the parental

organism (EFSA 2007; list of QPS organisms is updated yearly by EFSA’s BIOHAZ

11

(Biological Hazards) Panel) (EFSA 2012b). If the parental microorganism has a QPS status, the risk assessment can focus on the changes introduced (intended and unintended) during the development of the GM microorganism. Where no comparator can be identified for the GM microorganism and/or its product, a comprehensive safety assessment should be carried out. For the toxicological, allergenicity and nutritional assessment, the same basic approach as explained for GM plants is applied. The environmental risk assessment of the GM product is different depending on the category to which the product belongs (Aguilera et al. 2013). Provided that there are no detectable GM microorganisms and no recombinant genes in the final product, the potential environmental impact resulting from GM microorganism products under Categories 1 and 2 is considered negligible, and therefore no further data are required for such GM microorganism applications. However, other guidelines may be applicable to consider environment-related issues of the product per se, depending on its nature. For Category 3 products, the environmental risk assessment includes the potential environmental impact of the food or feed, and the possibility that the recombinant genes in it are horizontally transferred to indigenous microorganisms. The most extended scrutiny corresponds to products belonging to Category 4. For those products, in addition to the aspects considered for Category 3, the relationships between the GM microorganism and the biotic and abiotic factors in the environment (including human and animal gastrointestinal tracts) need to be assessed. This assessment should consider whether the GM microorganism survives and persists in the environment (including the human gastrointestinal tract), and eventually grows and/or mates with indigenous microorganisms. In addition, it is to be determined whether the GM microorganism can outcompete natural counterparts, and if its relationships with other organisms have been altered as a result of the genetic modification (for example, the development of potential pathogenicity or the loss of symbiotic interactions). Possible alterations of ecological processes such as decomposition or carbon and nitrogen cycles are also to be considered. Therefore, detailed information on the biology and ecology of the GM microorganism, including habitats, growth parameters, nutrient sources and requirements, production of secondary metabolites, antimicrobial resistance, reproduction, capacity of sporulation, VBNC state, or role in biogeochemical process, is essential for an appropriate evaluation. If the GM microorganism can actively transfer genetic material to other microorganisms, the assessment should also cover the potential effects of the expression of the introduced genes in the receiving organisms. If there are indications of adverse effects of the GM microorganism on indigenous microorganisms or in ecological processes, further data from specifically designed studies should be provided in order to evaluate the consequences of these effects. References

Aguilera J, Gomes AR, Olaru I (2013) Principles for the risk assessment of genetically modified microorganisms and their food products in the European Union. International J Food Microbiol, DOI:10.1016/j.ijfoodmicro.2013.03.013

Albajes R, Farinós GP, Pérez-Hedo M, de la Poza M, Lumbierres B, Ortego F, Pons X, Castañera P (2012) Post-market environmental monitoring of Bt maize in Spain: non-target effects of varieties derived from the event MON810 on predatory fauna. Span J Agric Res 10:977-985

Andow DA, Hilbeck A (2004) Science-based risk assessment for non target effects of transgenic crops. Bioscience 54:637-649

12

Bühler C (2006) Biodiversity monitoring in Switzerland: what can we learn for general surveillance of GM crops? J Consum Prot Food Safety 1:37-41

Comas J, Lumbierres B, Pons X, Albajes R (2013) Ex-ante determination of the capacity of field tests to detect effects of genetically modified corn on nontarget arthropods. J Econ Entomol 106:1659-1668

de Lange HJ, Lahr J, Brouwer JHD, Faber JH, 2012. Review of available evidence regarding the vulnerability of off-crop non-target arthropod communities in comparison to in-crop non-target arthropod communities. Supporting Publications 2012:EN-348 [53 pp.], EFSA, http://www.efsa.europa.eu/fr/supporting/doc/348e.pdf

EFSA (2006) Guidance document for the risk assessment of genetically modified microorganisms and their derived products intended for food and feed use by the Scientific Panel on Genetically Modified Organisms (GMO). EFSA J 374:1-115, http://www.efsa.europa.eu/en/efsajournal/doc/374.pdf

EFSA (2007) Opinion of the Scientific Committee on a request from EFSA on the introduction of a Qualified Presumption of Safety (QPS) approach for assessment of selected microorganisms referred to EFSA. EFSA J 587:1-16, http://www.efsa.europa.eu/en/efsajournal/doc/587.pdf

EFSA (2010a) Scientific Opinion on the assessment of potential impacts of GM plants on non-target organisms. EFSA J 8(11):1-72 [1877], http://www.efsa.europa.eu/en/efsajournal/doc/1877.pdf

EFSA (2010b) Guidance on the environmental risk assessment of GM plants. EFSA J 8(11):1-111 [1879], http://www.efsa.europa.eu/en/efsajournal/doc/1879.pdf

EFSA (2011a) Guidance for risk assessment of food and feed from GM plants. EFSA J 9(5):1-37 [2150], http://www.efsa.europa.eu/en/efsajournal/doc/2150.pdf

EFSA (2011b) Scientific Opinion on guidance on the risk assessment of genetically modified microorganisms and their products intended for food and feed use. EFSA J 9(6):1-54 [2193], http://www.efsa.europa.eu/en/efsajournal/doc/2193.pdf

EFSA (2011c) Guidance on the post-market environmental monitoring (PMEM) of genetically modified plants. EFSA J 9(8):1-40 [2316], http://www.efsa.europa.eu/en/efsajournal/doc/2316.pdf

EFSA (2012a) Scientific Opinion on an application (EFSA-GMO-NL-2005-24) for the placing on the market of the herbicide tolerant genetically modified soybean 40-3-2 for cultivation under Regulation (EC) No 1829/2003 from Monsanto. EFSA J 10(6):1-110 [2753], http://www.efsa.europa.eu/en/efsajournal/doc/2753.pdf

EFSA (2012b) Scientific Opinion on the maintenance of the list of QPS biological agents intentionally added to food and feed (2012 update). EFSA J 10(12):1-84 [3020], http://www.efsa.europa.eu/en/efsajournal/doc/3020.pdf

Evans J, Wood G, Miller A (2006) The risk assessment–policy gap: An example from the UK contaminated land regime. Environ Int 32:1066-1071

Garcia-Alanso M, Raybould A (submitted) Protection goals in environmental risk assessment: a practical approach. Transgenic Res

Graef F, De Schrijver A, Murray A (2008) GMO monitoring data coordination and harmonisation at EU level – outcomes of the European Commission Working Group on Guidance Notes supplementing Annex VII of Directive 2001/18/EC. J Consum Prot Food Safety 3:17-20

Gray A (2012) Problem formulation in environmental risk assessment for genetically modified crops: A practitioner’s approach. Collect Biosafety Rev 6:10-65

Meissle M, Álvarez-Alfageme F, Malone LA, Romeis J (2012) Establishing a database of bio-ecological information on non-target arthropod species to support the environmental risk

13

assessment of genetically modified crops in the EU. Supporting Publications 2012:EN-334 [170 pp], EFSA, http://www.efsa.europa.eu/en/supporting/doc/334e.pdf

Mönkemeyer W, Schmidt K, Beiβner L, Schiemann J, Wilhelm R (2006) A critical examination of the potentials of existing German network for GMO-monitoring. J Consum Prot Food Safety 1:67-71

Moonen AC, Bàrberi P (2008) Functional biodiversity: an agroecosystem approach. Agric Ecosyst Environ 127:7-21

Nienstedt KM, Brock TCM, van Wensum J, Montforts M, Hart A, Aagaard A, Alix A, Boesten J, Bopp SK, Brown C, Capri E, Forbes V, Köpp H, Liess M, Luttik R, Maltby L, Sousa JP, Streissl F, Hardy AR (2012) Development of a framework based on an ecosystem services approach for deriving specific protection goals for environmental risk assessment of pesticides. Sci Tot Environ 415:31-38

Perry JN, ter Braak CJF, Dixon PM, Duan JJ, Hails RS, Huesken A, Lavielle A, Marvier M, Scardi M, Schmidt K, Tothmeresz B, Schaarschmidt F, van der Voet H (2009) Statistical aspects of environmental risk assessment of GM plants for effects on non-target organisms. Environ Biosafety Res 8:65-78

Raybould A (2006) Problem formulation and hypothesis testing for environmental risk assessments of genetically modified crops. Environ Biosafety Res 5:119-125

Raybould A (2012) Can science justify regulatory decisions about the cultivation of transgenic crops? Transgenic Res 21:691-698

Romeis J, Bartsch D, Bigler F, Candolfi MP, Gielkens M, Hartley SE, Hellmich RL, Huesing JE, Jepson PC, Layton R, Quemada H, Raybould A, Rose RI, Schiemann J, Sears MK, Shelton AM, Sweet J, Vaituzis Z, Wolt JD (2008) Nontarget arthropod risk assessment of insect-resistant GM crops. Nature Biotechnol 26:203-208

Sanvido O, Widmer F, Winzeler M, Bigler F (2005) A conceptual framework for the design of environmental post-market monitoring of genetically modified plants. Environ Biosafety Res 4:13-27

Sanvido O, Romeis J, Bigler F (2011) Environmental change challenges decision-making during post-market environmental monitoring of transgenic crops. Transgenic Res 20:1191-2101

Sanvido O, Romeis J, Gathmann A, Gielkens M, Raybould A, Bigler F (2012) Evaluating environmental risks of genetically modified crops: ecological harm criteria for regulatory decision-making. Environ Sci Pol 15:82-91

Schmidt K, Wilhelm R, Schmidtke J, Beissner L, Mönkemeyer W, Böttinger P, Sweet J, Schiemann J (2008) Farm questionnaires for monitoring genetically modified crops: a case study using GM maize. Environ Biosafety Res 7:163-179

Smit E, Bakker PAHM, Bergmans H, Bloem J, Griffiths BS, Rutgers M, Sanvido O, Singh BK, van Veen H, Wilhelm R, Glandorf DCM (2012) General surveillance of the soil ecosystem: An approach to monitoring unexpected adverse effects of GMO’s. Ecol Ind 14:107-113

Wilhelm R, Sanvido O, Castanera P, Schmidt K, Schiemann J (2010) Monitoring the commercial cultivation of Bt maize in Europe – conclusions and recommendations for future monitoring practice. Environ Biosafety Res 8:219-225

Wolt JD, Keese P, Raybould A, Fitzpatrick JW, Burachik M, Gray A, Olin SS, Schiemann J, Sears M, Wu F (2010) Problem formulation in the environmental risk assessment for genetically modified plants. Transgenic Res 19:425-436