That looks like a problem with baseline correction. The baseline trend is estimated from a region of the amplification curve that shows a local increasing trend. This is extrapolated for background correction. This tilts the curve downwards, leading to the results you see. You can check the multicomponent view. Try to change settings for the baseline correction (try manual, using cycles where you have proper background signal without strong local trends).

Hi Andreas,

Jochen Wilhelm is right, this type of curve is invariably due to baseline correction. Altering the default analysis parameters for your green assay should make them look more typical.

Part of the problem is that you are using very strong template - the software needs to analyse fluoresence trends in a part of the programme where fluorescence is not expected in order to detract them from every other reading, any actual fluorescence in this region distorts the genuine trace. On top of that, concentrated template is not recommended as it increases to potential for contamination.

I would recommend trying to avoid to using any template that is likely to be called positive before cycle 20 - this will still leave at least 5-6 ten-fold dilutions to generate standard curves (each 1:10 dilution should be separated by 3.32 cycles if performed correctly and the assay has decent efficiency), reduces the potential for contamination and means the software should call the fluorescence without the need to tweak default settings

Thank you very much for your answers!

Unfortunately, the increasing signal exists as well in the raw data. It's not related with any algorism or baseline. It exist in all the dilutions we have prepared to get the standard curve about 9 logs.