2005 Volume 46 Issue 6 Pages 270-276
To investigate important factors affecting the analytical results, a laboratory-performance study was attempted for the Japanese official methods to detect genetically modified (GM) soybeans (40-3-2). Test samples containing 0, 1 and 5% GM soya powder in non-GM soya powder was prepared. A set of 3 test samples was sent to the participating laboratories along with the protocol. The data were collected from all laboratories and statistically analyzed. In the real-time PCR detection method, the average values of the GM 1% and 5% samples were both much lower than the spiked value because the laboratories using a silica-membrane DNA extraction method underestimated the GM value. On the other hand, the laboratories using other extraction methods, such as the CTAB method obtained values close to the spiked value. These results suggest that use of the silica-membrane DNA extraction method may result in underestimation of the GM content in the real-time PCR method. In the ELISA method, the average value of 5% spiked samples appears to be slightly higher than the fortified value. But, overall, it was considered that reported values were close to the spiked level.
