Sub Groups

• generally speaking, in relation to the reliability of testing, it was possible to be confident that at present an AP threshold of 1% for known GM events in crops and food was a practicable number. It had been agreed by the EU Scientific Committee on Plants that the detection limit of PCR testing should be considered as 0.1% because that was what could be reliably detected in practice in laboratories, even though the technology could in principle allow one to detect levels below 0.1%.
  
  
• the problem of quantitation only applied to approved varieties for which threshold levels were set. The concept of zero tolerance meant that that presence/absence was the only indicator for unauthorised events. There were quantifiable errors associated with testing to the 1% level in terms of standard deviation, and in terms of numbers of false positives and false negatives. These data were being compiled by CODEX and were sourced from all EU Member States
  
  
• unknown GM constructs in production were a major potential issue for the reliability of testing. Effective GM testing for particular events depended on having access to reliable information about those events. In order to be able to use or design primers that could be used to test for events. This information was gathered by CSL through its contacts, reviewing the literature, etc. CSL could not be sure it knew about all events, particularly as the number of new GM products being developed around the world accelerated. Some workshop participants thought that it would become simply impossible to be aware of all relevant material, especially given the rate of development of the technology in China and elsewhere, and global trade in agricultural produce. Moreover, the necessary information was often commercially confidential and so was difficult to obtain. It was unclear whether the Cartagena protocol alone would be an adequate means of overcoming this problem. The present and envisaged regime required testing for all known GM events (‘negative testing’) as opposed to positive testing (i.e. testing that a sufficient percentage of a crop was what it was meant to be, rather than seeking to detect other material). CSL thought that the information they had about GM events meant that the net was cast wide. It seemed to some workshop participants that a regime of positive testing in order to verify the non-presence of all unauthorised GM events and permitted levels of authorised events would become increasingly impracticable. The implications of requiring ‘negative’ testing needed further consideration, particularly under any commercialisation of GM crops in the UK.
  
  
• further research into methodologies, particularly direct detection and fast screening methods, ought to improve effectiveness of testing, as would adoption of harmonised protocols. The European Network of GMO laboratories (ENGL) was trying to deal with the latter, and to push forward an initiative to obtain sequence, primers and reference standards from the companies to support future testing.
  
  
• the effectiveness of testing depended in part on the proficiency of individual laboratories. Laboratories’ performance often improved with practice. It seemed reasonable to assume that in a dispute in Court about a case of the presence of GM, the performance of the laboratory would be taken into account in the credence given to a particular test result. One would also expect in a dispute that samples of the material in dispute would if possible be tested in different laboratories. Participation in proficiency testing was mandatory for accredited laboratories (EC 93/99). Results from proficiency testing were reported annually to the accrediting body. In the UK this was UKAS (United Kingdom Accreditation Service)
  
  
• importers and handlers worked at present to minimise AP in bulk grain shipments of crops where GM varieties were being commercially grown (principally soya) for the UK market. They did so by a range of quality assurance (QA) procedures and risk assessments of the production and transport of the soya. Industry worked to a - market-driven – threshold of 0.1% for food ingredients which suppliers had strong financial incentives to avoid breaching. In laboratory testing commissioned by industry, it was reported at the workshop, values were almost always lower than 0.1%, although in rare cases these were between 0.1 and 0.15%; and very occasionally a laboratory test yielded a value greater than 1%.
  
  
• the FSA baked goods study had shown unexpected levels of GM material. The wheat in the baked goods had been contaminated with small amounts of 99% GM soya, which contributed significantly to the overall GM content of the foods.
  
  
• The DEFRA GM Inspectorate (based at CSL) had no legal authority to require testing of a company’s product for GM unless they have reasonable suspicion. The Inspectorate would work with seed companies to ensure GM thresholds were met by means of assurances about production and through testing. The new EU seeds legislation required that seed companies took responsibility for testing to comply with GM thresholds, with official (Inspectorate) auditing.
  
  
• for seeds, demonstration of compliance with the ‘zero tolerance’ requirement for unapproved events could only be demonstrated in fact to the 0.1% sensitivity limit of the testing process. Some companies, however, reported tests done to a limit of 0.01% sensitivity. If when doing so a company in the UK found a level between 0.01% and 0.1%, then the company was required by the GM Inspectorate not to market the product. But it was unclear whether companies elsewhere in Europe did the same.
  
  
• there would be real practical difficulties in testing all crops for GM under widespread commercial growing conditions. For example, the window for testing, i.e. the period between seeding and harvest, for oil seed rape was about 2 weeks. There was not the laboratory capacity at present to test vast numbers of samples, as would seem to be required by forthcoming regulations if GM crops were commercialised. Morphological inspection of the crops for GM events prior to seeding would be in many cases impossible, and testing of vegetative material in the field prior to seeding impracticable. This issue needed to be considered in thinking about testing regimes and models of co-existence.
  
  
• a concern raised about Part B, i.e. experimental, releases of GM crops was that if absolute zero tolerance of cross-pollination of (by definition) unapproved events was demanded, this could drive experimental production overseas because of the impossibility of completely preventing rare cross-pollination occurrences.
  
  
• about 40% of oil seed rape produced last year in the UK was from farmer-saved seed. After a couple of generations of growing the saved-seed crop, the farmer would be unlikely to be able to have confidence that the crop would be below the required AP threshold.
  
  
• some workshop participants thought that the revised labelling and traceability and seed regulations were absurdly absolutist. Other participants noted that consumer pressure underlay them and it was the role of the AEBC to advise on what it was practicable for agriculture to deliver and the implications of what one could reliably test for and to what levels.
  
  
• GM testing of seed raised a different set of problems compared to food testing. An example of this was the identification in testing seeds of elements associated with the presence of GM constructs that are also known to be naturally present in soil borne organisms in the absence of transgenes. This had suggested the presence of transgenes when in fact none had been present. The marker genes had probably come from bacterial contamination of seed coats. This illustrates the need for a forensic capability in GM testing as well as more information on the nucleic acid sequence of constructs used.
  
  
• another difference was that the present quantification standard requiring a sample of 3000 seeds for testing, while relatively unproblematic for, say, oil seed rape, would be an expensive proposition if a potato seed crop had to be tested. (A sample of 3000 seeds was needed for testing for a 1% threshold in final products; for a 0.1% threshold, a much larger sample would be required.) There were international (ISTA) protocols for seed testing. Quantification standards needed further work, something on which ENGL was engaged.
  
  
• seed crop harvests were usually taken from individual farms to seed merchants and mixed there with other seed lots, at which point it would be too late to do GM testing that could pinpoint from which farm any AP had come. Testing for GM on the farm in addition to testing for quality of the crop, the results from which formed the basis for the value assigned by the seed merchant to the crop, aside from the practical issues of timing noted above, would seem likely to impose an additional burden on the farmer, though it was unclear how serious a burden.
  
  
• homogenised material was used in testing by PCR – not individual seeds or grains.
  
  
• whilst GM testing at CSL was done for both regulatory and commercial purposes, great carewas taken to ensure the complete separation of the processes. Results from the CSL commercial testing service were never reported to the GM Inspectorate. In the same way results from official testing were never divulged to commercial clients

• bulk shipments of grain typically involved 25,000 and up to 65,000 tonnes of grain in each ship. There was mixing of crops all the way up the supply chain. Each ship would contain produce from many farms. IP crops for human consumption typically came in much smaller shipments of around 4-5000 tonnes. A typical large soya crushing plant would process around 1,000,000 tonnes per annum. There was one such plant in the UK (Cargill’s at Liverpool).
  
  
• for each shipload, the crop would be either ‘GM’ or ‘non-GM’. Segregation of two types on board the ship was not attempted.
  
  
• UK supermarkets – and hence suppliers - were already working to a 0.1% threshold for food ingredients. In the rest of Europe 1% was more usual. UK supermarkets were also already working to source non-GM material in derivative products (e.g. oils). Again, this was not so elsewhere in Europe.
  
  
• at present it was usually relatively straightforward to source non-GM materials. The alleged problems about much of Brazilian soya being GM did not seem to be the general experience of companies, which sourced from northern Brazil.
  
  
• suppliers relied on risk assessments and quality assurance to ensure non-GM status, to industry (BRC/FDF IP) standards and in line with customer demand. Industry tried to minimise reliance on testing, because of problems of sampling (from these massive shipments) and of reliability of testing. Industry was confident that its procedures were robust and delivering to the standards required by customers.
  
  
• bulk commodity crops were sold on the futures and spot markets. The markets would adjust to meet demand. Prices generally would rise in the short term and stabilise thereafter. In 1999 there was a substantial premium for non-GM animal feed when demand suddenly increased; now the premium was marginal. If there was a sufficient level of macro demand from customers for a particular sort of product e.g. non-GM then suppliers will seek to source that, provided there was a clear specification to distinguish the product. Probably about two-thirds of non-GM soya imported into Europe was for the UK market. If demand for non-GM soya in the rest of Europe increased, prices for the commodity would go up initially, but would most likely reduce in time.
  
  
• the premiums for non-GM bulk produce were relatively modest at present. A lot of chickens raised in the UK were being fed non-GM animal feed due to retailer requirements. This cost the producers more than unsegregated feed but had not yet impacted on high street prices.
  
  
• for bulk crops, regionalised production to help ensure non-GM status might be a practical outcome of consumer wishes to preserve consumer choice. For example, one UK supplier got all its non-GM soya in Canada from a particular area in Ontario where most of the farmers were growing non-GM soya (although whether the main driver was to market non-GM status or because the GM varieties conferred negligible agronomic advantage in that locality was unclear).
  
  
• the cost of testing and other measures to ensure non-GM status of bulk crops was probably around $1-$1.50/tonne. The costs were spread through the supply chain up to but not including supermarkets. These costs had to be viewed against the severe financial penalties to a manufacturer of having to stop a production line or withdraw products because an AP threshold had been breached.
  
  
• from a commercial perspective, thresholds for purity of product could be considered as primarily a market-specification and so a contractual issue. It was generally agreed, that there could be differences between toleration of unapproved, experimental events and approved events that needed to be taken into account, but some workshop participants believed that it was a mistake to take a blanket rather than a case-by-case approach to this in regulation.
  
  
• AP if it occurred seemed for many staple crops more likely to come from physical mistakes in loading at the farm, failure to clean machinery, and other physical mixing (e.g. of corn and wheat) than through geneflow by means of cross-pollination.
  
  
• maintaining segregation involved some work in the supply chain as well as on the farm. For example, soya was very dusty and measures had to be taken accordingly in grain handling facilities between loads of GM and non-GM soya, including cleaning machinery and belts, running through and discarding an initial amount of the new material to be unloaded, etc. The grain-handling industry was confident that these procedures worked well at present and would continue to do so. It seemed to the industry more likely that measures to prevent AP would need firmer policing on the farm rather than elsewhere in the supply chain.
  
  
• industry thought that it would be possible, at a cost, to establish traceability for derived products such as oils, notwithstanding views that the proposals were illogical (no transgenes to test in the material) and practical difficulties of defining how far down the product chain one should go (products derived from the oils?) that many people in industry saw with imposing such a requirement.

• views varied as to the relevance of experience of seed certification to coexistence. There were remarkably few failures to achieve certification in conventional seed production. Was there, however, greater degrees of geneflow of recessives between crops than commonly supposed? Morphological analysis in conventional seed production could not detect this.
  
  
• it was clear that crops had to be considered case-by-case to determine what measures might be needed for coexistence.
  
  
• there was a question of whether there was enough empirical data to make sufficiently accurate predictions about levels of cross-pollination on the farm in commercial GM crop production. It was understood that some time ago NIAB had put forward proposals for projects to enhance the empirical data experiments lasting five years and costing around £1m per crop, but these proposals had not been accepted. Some limited empirical data about degrees of cross-pollination between adjacent fields for winter and spring oilseed rape and forage maize crops was expected to emerge from the farm-scale evaluations.
  
  
• there was some debate about the practical value of modelling studies. Taking these studies and available empirical data, while making clear that because there were so many variables it would be very difficult to predict levels of cross-pollination in the farm situation, Dr Ingram said that in respect of cross-pollination, it seemed likely that an AP threshold of 5% ought to be relatively easy to achieve, 1% would be difficult but probably achievable and 0.1% almost impossible.
  
  
• some workshop participants felt that in the light of the available data and experience, a phased introduction of commercial GM cultivation in the field, with careful monitoring, would be the most practicable way to assess and deal with co-existence.
  
  
• it was agreed that the key to considering possible coexistence was to look at the complete picture of agronomic practice. Cross-pollination had to be considered alongside other factors.