Simulation of collaborative studies for real-time PCR-based quantitation methods for genetically modified crops.

To study impacts of various random effects and parameters of collaborative studies on the precision of quantitation methods of genetically modified (GM) crops, we developed a set of random effects models for cycle time values of a standard curve-based relative real-time PCR that makes use of an endogenous gene sequence as the internal standard. The models and data from a published collaborative study for six GM lines at four concentration levels were used to simulate collaborative studies under various conditions. Results suggested that by reducing the numbers of well replications from three to two, and standard levels of endogenous sequence from five to three, the number of unknown samples analyzable on a 96-well PCR plate in routine analyses could be almost doubled, and still the acceptable repeatability RSD (RSDr < or = 25%) and the reproducibility RSD (RSDR < 35%) of the collaborative study could be met. Further, RSDr and RSD(R) were found most sensitive to random effects attributable to inhomogeneity among blind replicates, but they were little influenced by those attributable to DNA extractions. The proposed models are expected to be useful for optimizing standard curve-based relative quantitation methods for GM crops by real-time PCR and their collaborative studies.