实时荧光定量PCR检测限的统计推断

检测限是检测方法的重要性能指标,用Poisson分布的概率函数分式计算一次抽样检测出现的结果推论总体出现该结果的概率。以实时荧光定量PCR(FQ-PCR)检测乙型肝炎病毒DNA结果为例,计算可知,一次抽样反应管的检测结果≥4时,才可推论总体与0(空白)差异有显著性(P<0.05)。因而检测血清中病原体时扩增反应管中的检测限是4拷贝/ul计算也可知,1拷贝/ul表示一次抽样的结果为0的概率可达0.368,结果在0拷贝/ul～4拷贝/ul的概率为0.996,而1000拷贝/ml表示一次抽样结果为0的概率为0。结果在905拷贝/ml～1095拷贝/ml的概率为0.997;而100000拷贝/L表示一次抽样结果在997000拷贝/L～1003000拷贝/L的概率为0.997。抽样单位ul不能随意放大为ml甚至L。

Statistical Conclusion of Limit of Quqntitation in Real-time Fluorescence Quantitative PCR

Limit of quantiation is a important performance index of assay methods.Results of sample drawing are calculated through possiblity function equation of Poisson distribution,then the possibility of results emerged in population is deduced.For example,in Real-time Fluorescence quantitaive PCR to assay HBV DNA,only when sample drawing results≥4,significance of difference(P<0.05)could be deduced between population and blank.So limit of quantitation is 4 copies/ul denoting 0 in one sample drawing is 0,and Possibility of results between 905 copies/ml to 1095 copies/ml is 0.997.Possibility of 100 000 copies/L denoting results between 997 000 copies/L to 1 003 000 copies/L in one sample drawing is 0.997.Sample drawing unit ul could not be replace by ml or L.