Macarthur Roy, Feinberg Max, Bertheau Yves
Food and Environment Research Agency, Risk and Numerical Sciences Team, Sand Hutton, YO41 1LZ York, United Kingdom.
J AOAC Int. 2010 May-Jun;93(3):1046-56.
A method is presented for estimating the size of uncertainty associated with the measurement of products derived from genetically modified organisms (GMOs). The method is based on the uncertainty profile, which is an extension, for the estimation of uncertainty, of a recent graphical statistical tool called an accuracy profile that was developed for the validation of quantitative analytical methods. The application of uncertainty profiles as an aid to decision making and assessment of fitness for purpose is also presented. Results of the measurement of the quantity of GMOs in flour by PCR-based methods collected through a number of interlaboratory studies followed the log-normal distribution. Uncertainty profiles built using the results generally give an expected range for measurement results of 50-200% of reference concentrations for materials that contain at least 1% GMO. This range is consistent with European Network of GM Laboratories and the European Union (EU) Community Reference Laboratory validation criteria and can be used as a fitness for purpose criterion for measurement methods. The effect on the enforcement of EU labeling regulations is that, in general, an individual analytical result needs to be < 0.45% to demonstrate compliance, and > 1.8% to demonstrate noncompliance with a labeling threshold of 0.9%.
本文提出了一种用于估计与转基因生物(GMO)衍生产品测量相关的不确定度大小的方法。该方法基于不确定度分布图,它是一种用于估计不确定度的工具,是最近为定量分析方法验证而开发的一种称为准确度分布图的图形统计工具的扩展。还介绍了将不确定度分布图应用于决策辅助和适用性评估的情况。通过多项实验室间研究收集的基于PCR方法的面粉中转基因生物数量测量结果遵循对数正态分布。使用这些结果构建的不确定度分布图通常给出了对于含有至少1%转基因生物的材料,测量结果的预期范围为参考浓度的50 - 200%。该范围与欧洲转基因实验室网络和欧盟共同体参考实验室的验证标准一致,可作为测量方法适用性的标准。对欧盟标签法规实施的影响是,一般来说,单个分析结果小于0.45%表明符合规定,大于1.8%表明不符合0.9%的标签阈值规定。