The four most commonly used reference gene determination methods for qPCR studiesare: the comparative delta-Ct method (Silver et al., 2006), BestKeeper (Pfaffl et al.,2004), GeNorm (Vandesompele et al., 2002) and NormFinder (Andersen et al., 2004).Gene expression stability of reference genes are evaluated differently in each of the fourmethods.The comparative delta-Ct method calculates the stability of each gene by obtaining thestandard deviation of Ct differences within each sample, for each pairwise comparisonwith the other genes and averaging them. If the delta-Ct value between two genesremains constant when analysed in different samples of RNA, it means that either bothgenes are stably expressed among those samples, or they are co-regulated (this approachassumes stability of both genes). However, if the delta-Ct fluctuates, then one or bothgenes are variably expressed. Introduction of a subsequent gene into the comparison willprovide more information on which pairs show less variability and therefore whichgene(s) has more stable expression among the samples tested. This method requiressimilar efficiency for the reference gene and the gene of interest during amplification.NormFinder is a Visual Basic application for Microsoft Excel, which automaticallycalculates the stability value for all candidate normalisation genes tested on a sample set.NormFinder gives user possibility to rapidly compare any number of samples organizedin any given number of groups. NormFinder is freely available for the researchers onrequest from the authors. The NormFinder as a model-based approach top ranks thecandidates with minimal estimated intra- and intergroup variation. For the first stepNormFinder estimates intergroup expression variations that allow for selection of thesuitable genes. Secondly, the intragroup variance is calculated to identify the number ofgenes to include. Having estimated both the intra- and intergroup variation, the two arecombined into a stability value. This value intuitively adds the two sources of variationand thus represents a practical measure of the systematic error that will be introducedwhen using the investigated gene (Andersen et al., 2004). The optimal number of genesis reached when addition of a further gene leads to a negligible reduction in the average17of the gene variance estimates (Andersen et al., 2004). One of the strengths ofNormFinder is that it can be used when data is collected from different sample groups,because both intra- and intergroup variation is taken into account for reference genestability assessment (De Spiegelaere et al., 2015).BestKeeper gene stability assessment software’s main statistical parameter is thestandard deviation of Ct values, which should be low in the case that the input materialused for all samples has an equal amount and quality. That data can be used to excludespecific reference genes when their standard deviation is too high (>1.5). After thatselection process, BestKeeper calculates the BestKeeper Index from the geometric meanof the suitable reference genes and performs Pearson correlation for each of the chosenreference genes to the BestKeeper Index to express the correlation of that gene with theIndex (Pfaffl et al., 2004).GeNorm gene stability assessment tool is based on the principle that the expression ratioof two ideal internal control genes is identical in all samples and, furthermore, this doesnot depend on experimental conditions or cell type (Vandesompele et al., 2002). Forevery reference gene candidate, GeNorm algorithm determines the pairwise variationwith all other reference genes as the standard deviation of the logarithmicallytransformed expression ratios, and defines the internal control gene-stability measure Mas the average pairwise variation of a particular gene with all other control genes. Geneswith the lowest M values have the most stable expression. On the assumption that thecontrol genes are not co-regulated, stepwise exclusion of the gene with the highest Mvalue results in a combination of two constitutively expressed HKGs that have the moststable expression in the tested samples (Vandesompele et al., 2002). Therefore,variations of expression ratios for pair of HKGs to all other reference genes reflect thefact that the gene is not constantly expressed with increasing variation ratiocorresponding to decreasing expression stability and usability as reference gene.From the four validation methods briefly described here, there is no commonly acceptedbest method. Each of the four validation methods represents viable strategies forreference gene validation, but still some problems can arise in certain experimentalscenarios (De Spiegelaere et al., 2015). BestKeeper uses a standard deviation for gene18ranking, but this approach doesn’t differentiate between variation that is stemming fromnatural variation of gene expression and from technical procedures (sampling, RNAextraction and RT steps etc), whereas GeNorm and the comparative delta-Ct methodanalyse the correlations between genes, assuming that the reference genes are not co-regulated (Vandesompele et al., 2002). It is risky to assume that genes are not co-regulated because this cannot be easily confirmed and as a consequence, two co-regulated genes could mislead to erroneous reference genes (De Spiegelaere et al.,2015). NormFinder algorithm is not susceptible to possible hidden co-regulated geneexpression, since it takes into account intergroup variation, which should be as low aspossible for a good reference gene. However, similarly to BestKeeper, a low overallintergroup and intragroup variation does not necessarily mean that it is a good referencegene.A number of studies have solved the disagreement between the four reference genevalidation methods by ranking reference genes according to the geometric mean of thefour ranking numbers for each gene: the lower the mean a gene gets the most stable it is(Robledo et al., 2014). This “cutting the Gordian knot” strategy can be argued, whetherall of the validation methods should have same contribution to final ranking, giving theirdifferent applicability, overlaps in approach and robustness in different experimentalconditions. RefFinder is a web-based tool that calculates the average gene stability usingthe four well known algorithms (GeNorm, NormFinder, BestKeeper, and thecomparative delta-Ct method). Based on the rankings generated by each algorithm,RefFinder assigns an appropriate weight to an individual gene and calculates thegeometric mean of their weights for obtaining the overall final ranking. In addition,RefFinder tool enables the user to evaluate results according to each of the fouralgorithms separately.