General error and directional shift

There are two types of error that must be controlled for in a qRT-PCR assay. One is general error (noise) and directional shift: Figure 4.1 shows the many stages at which general error can be introduced into a qRT-PCR assay. Directional shift occurs when poor normalization either creates an arti-factual difference or obscures a real difference between experimental

Figure 4.1

Multistage process required to obtain a real-time RT-PCR result. Stars indicate the points at which normalization can be performed: 1: sampling, 2: RNA, 3: cDNA and 4: PCR dependent standard.

Figure 4.1

Multistage process required to obtain a real-time RT-PCR result. Stars indicate the points at which normalization can be performed: 1: sampling, 2: RNA, 3: cDNA and 4: PCR dependent standard.

Norm. floro.

Figure 4.2

Figure 4.2

Example of a triplicate real-time PCR reaction (using the same template). This demonstrates that real-time PCR can be highly reproducible with coefficients of variation of <5%).

groups. Real-time PCR has been reported to measure two-fold differences in nucleic acid amounts (Bubner et al., 2004) and is certainly capable of coefficient of variation (CV) of <1 % (Figure 4.2). However, achieving consistently low CVs in qRT-PCR assays is not as simple, with even the best CVs significantly higher than 1% (Melo et al., 2004). The main reason for this is that it is not feasible to determine directly the target RNA. Instead, RNA quantification requires the inclusion of several steps that eventually measure cDNA copy numbers, but that also serve as sources of systematic variation, increasing CV and limiting the resolution of the assay. Variability is introduced during sampling, RNA extraction, cDNA synthesis and comparison with PCR dependent controls, and appropriate normalization of each step is essential (Figure 4.1).

0 0

Post a comment