CORDIS - Forschungsergebnisse der EU

Multiplex detection of (un)authorized GMOs in food and feed

Final Report Summary - GMULTI (Multiplex detection of (un)authorized GMOs in food and feed)

Project objectives:
The main goal of the project was to apply and further develop a multiplex DNA detection system for identification of (un)authorised genetically modified organisms (GMOs) in European food and feed chains. The padlock-probe (PLP), ligation-based multiplex detection method provides a method that can potentially meet all the demands of (un)authorised GM detection. The project's objectives were to design PLPs for several DNA targets and test them for the most important parameters: specificity, sensitivity and the relation between performance and number of targets. Further aim was to build a model to identify the most informative elements, constructs or events that could be tested in case of an unknown sample, prior to any actual laboratory work.

Description of the work performed:
Twenty-six padlock probes were designed. In all cases specific hybridisation was observed. The best 20 probes were mixed together for multiplex testing. These probes were specific for GM events, elements and endogenous plant genes. During the specificity testing of the probes all but one probe reacted specifically with their targets, when present, and the ones for which no target was present showed negative results. At the multiplex level we successfully detected 13 targets in one reaction.

Different kinds of multiplex ligation detection methods have been published with different reaction times, temperatures, number of repeats within the procedure, reaction buffers and design of probes. These have been compared with the Rikilt padlock probe (PLP) system. A statistical evaluation combining ANOVA and the Tukey HSD test was performed to decide which ligation protocol(s) worked the best. The best methods were the Rikilt protocol and a protocol published by Keith et Al. in the Plant Biotechnology Journal, issue 7, pages 375-390 in 2009. The results were also confirmed on microarray.

The original Rikilt protocol was chosen to test the sensitivity and the transferability of the method as an initial validation. A tenplex system was tested by the National Institute of Biology ((NIB) in Slovenia) to examine the transferability of the method in a double-blind approach. In this case a 0.1 % sensitivity was reached. Similar results were achieved from the two institutes (NIB and Rikilt) which proved the reproducibility of the method in another lab. An alternative method to PCR, rolling circle amplification (RCA), was also tested to see the effect on the sensitivity, but performed clearly not as well.

- Expected final results and their potential impact and use:
This research is very relevant for European consumers, because this method can provide an alternative tool for maintaining freedom of choice between GM-free and GM-containing foods and can also be used as an extra warranty of correct labelling of products. Further improvements are necessary for the event specific detection and for validation of the method in order to fully implement this elegant procedure in the routine analysis of food and feed samples. The transferability of the method was confirmed in cooperation with another institute (NIB, Slovenia) and the researcher is from Hungary so the project can improve the collaboration among these three countries.

Furthermore the results were presented at the 'Fourth international conference on coexistence between genetically modified (GM) and non-GM based agricultural supply' and is going to be published in a scientific paper which will help to reinforce the collaboration among the other EU member states regarding GMO detection. This project will help to maintain and increase the competitive scientific level in the EU, thus making the EU more attractive for researchers worldwide. Henceforth the decision tree model can be useful for institutes and companies who are working with GMO detection to minimise the number of tests and the cost of quality control.