Hello,
I am studying the expression of miRNA in different human body fluids. For this study I am using three reference genes (RG). I know how to calculate delta Ct value using single RG but don't know how to calculate delta Ct when using three RG's.
However, I have done little research and came across that we can take the average of all three RG's [CtRG = Mean (CtRG1 + CtRG2 + CtRG3)/3] and then calculate the delta Ct, which is delta Ct= CtRG - Ctgene of interest.
I want to know if this method is scientifically correct/rational or not? Also wanted to know is this method an alternative of geometric method for multiple RG calculation?
Many thanks.
It is good to use more than one RG. The more the better, because this increases the chance that (actually unwanted, but yet possibly present) treatment effects on the reference genes will cancel out. It also increases the precision of the measurement (and hence the power of the tests), because more information is collected and used (the information is not independent, so this has no effect on the degrees of freedom, but it reduces the standard errors).
Averaging the Ct values is natural and rational, as Ct values are approximately normal distributed and we are interested in differences between (d)Ct values. Under the given assumption of variance homogeneity it is equivalent to the "geometric method". The difference is that the full geometric method aims to directly estimate ratios of expected values, and analyzing the difference between Ct values is actually about the ratio of medians. This difference is of no practical relevance unless you have a fully quantitative model of gene regulation that translates to features/variables with a different distribution and/or a different scale of variance (coefficient of variation) where using the expected values matters (I don't know a single case like this, but maybe someone can give me an example here).
Jochen Wilhelm, Thank you so much for the elaborated and scientific answer :-)
- Badges
- Science method