In 2002, the relative expression software tool (REST, http://rest.gene-quantification.info) was established as a new tool (Pfaffl et al., 2002). The first REST version is Excel®-based and programmed in Visual Basic to compare several gene expressions on CP level. It compares two treatment groups, with multiple data points in the sample versus control groups, and calculates the relative expression ratio between them. The mathematical model used is published and is based on the mean CP deviation between sample and control group of target genes, normalized by the mean CP deviation of one reference gene as shown in eq. 7 (Pfaffl et al., 2002). Further an efficiency correction can be performed, either based on the dilution method (eq. 9) or an optimal efficiency of E = 2.0 is assumed. The big advantage of REST is the provision of a subsequent statistical test of the analyzed CP values by a Pair-Wise Fixed Reallocation Randomization Test (Pfaffl et al., 2002). Permutation or randomization tests are a useful alternative to more standard parametric tests for analyzing experimental data (Manly, 1997; Horgan and Rouault, 2000). They have the advantage of making no distributional assumptions about the data, while remaining as powerful as conventional tests. Randomization tests are based on one we know to be true: that treatments were randomly allocated. The randomization test repeatedly and randomly reallocates at least 2000 times the observed CP values to the two groups and notes the apparent effect each time, here in the expression ratio between sample and control treatment. The REST software package makes full use of the advantages of a randomization test. In the applied two-sided Pair-Wise Fixed Reallocation Randomization Test for each sample, the CP values for reference and target genes are jointly reallocated to control and sample groups (= pair-wise fixed reallocation), and the expression ratios are calculated on the basis of the mean values. In practice, it is impractical to examine all possible allocations of data to treatment groups, and a random sample is drawn. If 2000 or more randomizations are taken, a good estimate of P-value (standard error <0.005 at P = 0.05) is obtained. Randomization tests with a pair-wise reallocation are seen as the most appropriate approach for this type of application. In 2005 various new REST versions were developed, calculating with a geometric mean averaged REF index (Vandescompele et al., 2002; Pfaffl et al., 2004), according to the mathematical model described in eq. 8, which can analyze 15 target and reference genes (REST-384) (LightCycler® Relative Quantification Software,
2001). Specialized REST versions can compare six treatment group with one non-treated control (REST-MCS, REST - Multiple Condition Solver) (LightCycler® Relative Quantification Software, 2001), or take individual amplification efficiency into account, exported from the Rotor-Gene (REST-RG). A stand alone application REST-2005 was developed, running independent of Excel® or Visual Basic, comparing 'unlimited' target and reference genes, using newly developed bootstrapping statistical tool, and graphical output showing 95% confidence interval (TUM and Corbett Research, 2005) (Pfaffl and Horgan, 2005).
Was this article helpful?