基于联合验证数据评定转化体定量检测结果测量不确定度  

Evaluation of Measurement Uncertainty of Transformation Quantitative Detection Results Based on Joint Validation Measurement Data

作　　者：赵新[1] 张华 武玉花[3] 李飞武[4] 沈平 兰青阔[1] 徐利群 王永[1] ZHAO Xin;ZHANG Hua;WU Yuhua;LI Feiwu;SHEN Ping;LAN Qingkuo;XU Liqun;WANG Yong(Tianjin Academy of Agricultural Sciences,Tianjin 300384,China;Development Center of Science and Technology,Ministry of Agriculture and Rural Affairs,Beijing 100176,China;Oil Crops Research Institute,Chinese Academy of Agricultural Sciences,Wuhan 430062,China;Institute of Agricultural Quality Standards and detection technology,Jilin Academy of Agricultural Sciences,Changchun 130033,China)

机构地区：[1]天津市农业科学院,天津300384 [2]农业农村部科技发展中心,北京100176 [3]中国农业科学院油料作物研究所,湖北武汉430062 [4]吉林省农业科学院农业质量标准与检测技术研究所,吉林长春130033

出　　处：《大豆科学》2024年第5期539-549,共11页Soybean Science

基　　金：农业生物育种重大项目(2022ZD040201103)。

摘　　要：随着转基因生物及产品定量标识制度的不断完善,大批转化体定量检测方法的建立迫在眉睫,为了在实验数据积累不足的情况下,在检测方法建立初期及时评定测量不确定度,本研究以耐除草剂大豆DBN9004定量检测方法为例,在检测实验室积累到足够检测数据之前,在8家实验室对5种转基因含量样品进行联合验证,分析试验数据的再现性标准偏差,形成“自上而下”的不确定度预评定方法。结果显示:经格拉布斯法和科克伦法检验8家实验室数据,实验室5的数据存在离群值和岐离值,故剔除,7家实验室5个含量样品的再现性相对标准差RSD_(jR)为6.025%~11.532%,均小于规定的35%,说明该方法具有良好的精密度。根据上述分析数据绘制回归曲线方程为s_(jR)=0.0783×c/≡_(j)-0.0104,计算的相对标准不确定度分量u_(r)=0.0783,u_(0)=0.012,进而将试样检测结果的平均值(c=)带入公式,计算定量结果的标准不确定度u=√u_(0)^(2)+(u_(r)×c/=)^(2)=√0.012^(2)+(0.0783×c=)^(2)。结果表明基于联合验证数据评定转化体定量检测结果测量不确定度,为转化体定量检测方法建立初期提供了科学有效的实验室自行评定方法,为转基因生物及产品的安全监管提供了技术保障。With the continuous improvement of the quantitative identification system of transgenic organisms and products,it is urgent to establish a large number of quantitative detection methods for transformation.In order to evaluate the measurement uncertainty in time at the early stage of method establishment when experimental data accumulation is insufficient,it is urgent to establish a scientific and effective evaluation method.This study takes the quantitative detection method of herbicide-resistant soybean DBN9004 as an example.It introduces a method of using 8 laboratories to conduct standard establishment joint validation on 5 kinds of transgenic content samples before the detection laboratory accumulates enough detection data.It analyzes the reproducibility standard deviation of experimental data,and conducts a"top-down"pre-evaluation of uncertainty.The results showed that after Grubbs and Cochran tests of the data from 8 laboratories,the data from laboratory 5 had outliers and were excluded.The relative standard deviation of reproducibility of the 5 content samples from the remaining 7 laboratories was between 6.025%and 11.532%,which was less than the specified 35%,indicating that the method has good precision.Based on the above analysis data,the regression curve equation was s_(jR)=0.0783×c/≡_(j)-0.0104,and the relative standard uncertainty component u_(r)=0.0783,u_(0)=0.012.Then,the average value of the sample detection results(c=)was substituted into the formula to calculate the standard uncertainty of the quantitative results u=√u^(2)0+(u_(r)×c=)2=√0.012^(2)+(0.0783×c=)^(2).The experiment shows that the evaluation of measurement uncertainty of transgenic quantitative detection results based on joint validation data provides a scientific and effective self-evaluation method for the establishment of transgenic quantitative detection methods in the early stage,and provides technical support for the safety regulation of transgenic organisms and products.

关 键 词：测量不确定度 预评定 联合验证 转基因 耐除草剂大豆DBN9004 安全监管 

分 类 号：S565.1[农业科学—作物学]