Since you cannot run a melting curve with TaqMan probes, you might want to run an agarose gel and see if there's really any DNA of any size in your reaction?

hi, I would have a look how the data look withour backgroudn normalization. The phenomnon you are mentioning is something happening sometimes e.g. if the wells were not porperly mixed prior to the PCR run. You should see it in the raw data

My guess is that your primers have an efficiency that is too low. What does your standard curve look like?

Thanks everyone for your responses. Olga Korolkova , since these are multiplexed reactions which include an internal reference gene for my samples, all of them have a clear band that just looks thicker in some cases but doesn't really seem to correlate with the amount of amplified targets. I do see dimers but I'm not sure if probes could be hybridizing to them, since when I look at my NTCs, only 1 presents a noise signal on a single channel. Also, gel resolution doesn't allow a clear signal to be compared to other samples, even with a higher exposure. Sven Reister since I'm using CFX96, there's no background normalization with a reference Dye like ROX. Although I did find a setting called "Fluorescence drift correction" which eliminated some of the noise from initial cycles. Katie A Burnette I have tested for efficiency in singleplex, multiplex with single amplicons, and multiplex with all amplicons. Efficiencies have reached 78% in some instances, but also 55%, when attempting certain changes or using samples that seemed to be degraded. Although I'm not sure as to how primer efficiency could cause probes to work like that. I'll keep digging into it though.

Those efficiencies are too low to be useful or to publish.

You have to be 95-105% for any decent journal to accept your data.

The primer efficiency determines the accuracy and sensitivity of your assay. The probe is just a way to measure the number of molecules. Don't multiplex if you can't get single reactions to work efficiently.

I'd recommend scaling back your experiment and optimizing it.

Good luck!

Hi Cristobal,

Your primers may not be as efficient as they should be.

Cristobal Urrutia even in the Biorad CFX or Maestro Software you have the option to remove baseline correction. Would suggest to do so to see the drift (The screenshot was pretty typical so no issue to see which Software it is, I was not talking about a passive ref dye)

This sounds like high background fluorescence. This can easily happen as many common lab artefacts (including labcoats and some kinds of disposable gloves) produce fluorescent dust.

The solution is probably to manually reset the background subtraction.

By the way, I would strongly advise displaying your amplification curves on a logarithmic scale. It's very difficult to identify the exponential part of an amplification curve on a linear scale.

I don't agree with Katie Burnett about efficiency. 95% efficiency is certainly desirable. > 100% efficiency is impossible; although an instrument may sometime report an efficiency over 100%, this means there is something wrong with the standard curve.

Sometimes, for example when designing taxospecific primers, you have to make compromises, accommodate mismatches, have a PCR product bigger than 150 bp or suchlike. In such cases, 95% efficiency is not achievable. It doesn't matter as long as the primers are specific and you run enough cycles.

Sven Reister I understand, I couldn't find anything about baseline correction in the equipment manual, that's why I thought you meant something else. If I understood your point correctly, I might've been able to correct this issue by doing what you said: I changed the Cq determination mode to "Regression" instead of "Single threshold", which eliminated the use of a baseline, and also without the use of the feature I mentioned earlier (Fluorescence drift correction), it changed the efficiencies of my primers to between 91.3 and 83.3% with an R^2 > 0.99. I know this is still suboptimal, as Katie A Burnette said, but I am confident that these primers were carefully designed, so the lower efficiencies from before couldn't be right. I did test lower concentrations for both primers and probes which also helped a little. Also, Andrew Jenkins I have read that primer efficiency can reach over 100% but doesn't mean greater efficiency, but probably contamination, since this can cause standard curve slopes to go above -3.32 by flatteing the curve.

Thanks everyone for your answers, you've all been very helpful !

It doesn't matter if the primers were "carefully designed" or not. You cannot use or publish these data.

Design new primers, test them, and only use them if they are 95-105% efficient. You are correct that above 100% is an artifact, but it's within the acceptable range for qPCR. Your reactions are not.

Look, you can't rush quality. We're given you good advice. It's up to you to decide how to proceed.

You're entitled to your opinion, Katie, but I disagree with you. In general terms, it's the specificity that matters. On the other hand, multiplexing is very tricky, and if you have wildly differing efficiencies, you risk having the most efficient reactions out-compete the less efficient ones.