How to choose calibrator (housekeeping) genes for RT-qPCR normalisation

Question

[Juan Cordero]

Hi,

I am using geNorm and normFinder methods (implemented in NormqPCR R package) to find housekeeping genes from AP (life tech) cards, that stay stable across samples. The reason behind is the (supposed) housekeeping genes do not behave as such, and they are highly variable across and even within samples. For instance U6 snRNA-001973 calibrator gene may vary between 4 and 10 Cts within samples, for all of them.

Therefore, I get several stable housekeeping genes (with geNorm and NormFinder), but they show high Ct values (around 33-34), and I do not really know if I should trust them.

Furthermore, I get very low correlations or nothing at all between biological replicates, before and after normalisation with the computed housekeeping genes, which makes very hard for me to trust differential expression of different conditions, when the processed data is so unreliable.

I am a newbie with RTqPCR data analysis and do not know if it is possible to make magic with these apparent noisy data, and would really appreciate some feedback, comments or advice.

[jotan]

Are you comparing across tissues or between experimental treatments?

Answer 1

[karl.stamm]

There's a pretty good investigation of the stability of housekeepers over here.