Comparative testing report on the detection and quantification of soybean 40-3-2 and maize MON 88017 : preliminary reportComparative testing round: ILC-EURL-GMFF-CT-02/12

In the frame of Regulation (EC) No 882/2004, the European Union Reference Laboratory for Genetically Modified Food and Feed has the duty to organise comparative testing rounds and to ensure an appropriate follow-up of these activities. This preliminary report describes the outcome of the sixth comparative testing round ILC-EURL-GMFF-CT- 02/12. Participants were required to screen two test items denoted compound feedstuff levels 1 and 2, for the presence of maize events Maize MON 88017, MON 89034 and soybean events 356043, 40-3-2 and MON 89788. Any events detected were then to be quantified. The test items were produced in-house. The EU-RL GMFF managed the on-line registration and submission of results. A total of 160 laboratories were invited to participate in ILC-EURL-GMFF-CT-02/12. Participants could report the results of the exercise either in mass/mass % or in copy/copy %. Overall, in this sixth comparative testing round results expressed in m/m % were characterised by a higher percentage of z-scores lying within the working range of -2 to +2 with respect to results reported in cp/cp % (mean 94.37 % versus 78.76 %, estimated only on z-scores calculated on the basis of the robust means). However, The differences in the percentages of z-scores within the working range between m/m % and cp/cp % results for maize event MON88017 (mean 96.08 % versus 71.57 % for m/m % and cp/cp % respectively, estimated only on Z calculated on the basis of the robust means) appear more relevant than what can be seen for soybean event 40-3-2 (mean 92.66 % versus 85.95 % for m/m % and cp/cp % respectively, estimated only on z-scores calculated on the basis of the robust means). Moreover, a relevant difference between the assigned value and the robust mean for level 2 concentration of maize event MON 88017 was observed. Such a difference occurred because of a substantial under-estimation of the actual concentration level by 43 out of 51 laboratories (i.e. 84.31 %) reporting results in m/m % format