表达谱基因芯片的可靠性验证分析

cDNA芯片是一项新兴的能评估检测全范围mRNA表达水平变化的技术。通过同种组织RNA自身比较实验及不同组织RNA的差异分析实验对cDNA芯片实验的重复性进行检验，利用相关系数(correlation coefficient，R)、变异系数(coefficient of variation，CV)和假阳性率(false positiver ate，FPR)分析eDNA芯片数据的可靠程度，对cDNA芯片实验数据作了整体的评估。结果证实，该芯片系统得到的cDNA表达谱数据相关系数一般大于0．9，平均变异系数15％左右，假阳性率控制在3％以内。还提出一致率(consistence rate，CR)的概念，作为衡量cDNA芯片系统重复性的新参数，同时阐述了该参数优于目前常用的相关系数及变异系数的特点。另外，通过比较芯片制备中点样浓度、mRNA和总RNA以及不同批次芯片和不同标记过程对实验的影响，来分析芯片数据的系统误差来源。并提出重复两次实验，可以克服绝大部分实验系统引入的假阳性。

Validation of cDNA Microarray Technology

cDNA microarray is a technological approach that has the potential to globally measure changes in mRNA expression levels.Self comparison experiments with the same kind of tissue and differential expression experiments with the different kinds of tissue have been done to verify the reproducibility and the accuracy of this technique.The parameter of the reliability and the reproducibility of the microarray data were analyzed by correlation coefficient (R),coefficient of variation (CV) and false positive rate (FPR) etc.Meanwhile,the error resource also has been inspected.These results showed that generally the correlation coefficient of data from this cDNA microarray system was more than 0 9,the coefficient of variation was about 15%,and the false positive rate was below 3%. The result proves the accuracy of the cDNA microarray data.Consistence rate (CR) was advanced here as a new parameter to evaluate the reproducibility of two replicate experiments.It has some advantages over correlation coefficient and coefficient of variation.The influence of some important factors in the experiments,such as different concentration of spotted DNA,mRNA and total RNA,different batches of slides and different processes of labeling,have been investigated by comparing the results.It was shown that most of the false position produced by the experiment system could be reduced by replicate experiments.