You cannot perform ddCT with "not detected" data, because ddCT is a measure of fold change. The fold change from "not there" to "there at any level at all" is always going to be 'infinite fold', and this is of very little biological use.

The data you attached is for microRNAs, which are always going to be a bit more of a challenge than mRNAs, but even then, your Cq values are high: anything above 30 suggests you're into stochastic partitioning range (where you have single or low-double digit numbers of transcripts per well, and can quite easily see massive variation from well to well).

From the data attached, the conclusions (assuming all else is equal) I would reach would be:

MIR1244-1: not expressed

MIR125A: not expressed

MIR3198-2: barely expressed

housekeeping genes

MIR23A: not expressed

MIR24-1: barely expressed

Though frankly my first instinct would be to assume your cDNA synthesis was dodgy, or that your amount of starting material was low. Using the same cDNAs, do you have any miRs that give Cq values of 20-25? That would help reassure you that your samples are at least ok.

Obviously you cannot use miRs that are not expressed as reference miRs, so your next step would be to check whether your choices of reference are appropriate. The same applies to mRNAs, if you are looking at those too. Ideally your reference gene/miR should be of modest to strong expression (20-25 Cq), because again: Cqs in the 30s are highly variable.

Finally, you don't need to use the ddCT method for everything. It is a method that assumes you are measuring fold changes in expression, rather than "yes/no" changes (for simple yes/no, you could use regular PCR, for instance).

The Cq data can still tell you things, though: if you have a gene/miR that goes from "not detected" in controls to 35-37 Cq in treated conditions, view this with suspicion, since this is "not there" to "barely there", and only if this increase is very, very reliable should you view this as an increase in expression. A very, very modest increase in expression, for which you should be careful about making biological interpretations.

If you have a gene/miR that goes from "not detected" in controls to 20-25 Cq in treated conditions, this is a very robust expression response, representing increases from "none per well" to "hundreds of thousands per well", and you could report this accordingly. It would certainly qualify as a gene differentially expressed between treated and control.

Hello Tobias

You can simply report them as: "No expression detected" and they cannot be analyzed with delta-delta method. If this is contrary to your hypothesis, you may double check your reactions such as RNA input for making cDNA synthesis, qPCR reaction condition etc.

Regards

Hello Tobias,

Housekeeping genes miR-23A and miR-24-1 seems to be bad for your experiments. You should carefully select housekeeping for miRNAs because they varies a lot in different tissues ( https://www.nature.com/articles/srep41127 ). If your starting material is 50-100ng/ul then housekeeping should be somewhere around 18-20 Ct. Also there should not be more than 0.5-1 Ct variation in Ct values of housekeeping between treatment and control. Try to take almost equal amount of RNA and atleast 50ng/ul if you are looking for miRNAs.

Thanks