Hello dear collegues. Im using sybr green 1 qpcr assay for evaluation of relative mitochondrial copy numbers in dna samples from human blood. Im using primers for nuclear and mitochondrial genome that showed good efficiency from standard curves (pic 1,2,3).
I use 96 well plates and aliquots of primers and 2 control samples in each run (both control samples were tested by running standard curves). Problem is that each run shows large differences in average dRn value of all samples (pic 4,5,6 - 2 control samples; pic 7 - all samples in plate). Even both targets (Mito and Nuclear) show differences in dRn in the same run, so dCq between Mito and Nuclear targets of control samples and ddCq of 2 control samples change every run (Up to 30%). And multicomponent plot looks strange - what is that increase of fluorescence at 40 cycle? I thought maybe reaction efficiency varies between runs, but what could be the problem? I checked amplification products from plates with tenfold difference in fluorescence (pic 4 and 6) in agarose gel and bands had same intensity. I dont use reference dye but i doubt it can solve the tenfold difference in dRN problem. Also i have evaporation from some wells in each run and dRn of these wells are lower but could it decrease dRn of whole plate tenfold.
Not quite sure what your problem is. The amplification curves look fine (you should use a log-scale of the y-axis to get a better view on the exponential phase!), the standard curves are fine, and the multicomponent plots are also ok. What you see in the multicomponent plot is the series of all measurements taken from a well. The first 40 measurements are the ones taken after each amplification cycle (at 72°C). The further data points are supposedly from a melting curve. I assume you have started the curve at a temperature considerably lower than 72°C. The lower the temperature, the higher is the fluorescence signal ("temperature quench"). This explaines the steep step upwards directly after "cycle 40". During the melting curve, the temp is continuosly increased and the wells are repeatedly measured (the axis should actually show the temperatures, not "cycles"). As the temp increases, the fluorescence decreases (temp. quench). At "cycle 100" (that means, about 60 measurements after end of the 40 PCR cycles), the fluorescence decreases rapidly. This is where the temerature reaches the melting temperature of the PCR products.
Jochen Wilhelm, thank you for the explanation of temperature quench! As you said i start melting curve from Ta of primers.
The main problems is that such differences in dRn change the ddCq of 2 control samples (and ddCq of all other samples respectively), so i cant realy use this runs for analysis. Its like if i put 2 standart cuves (for nuclear and mitochondrial targets) in each run and the slope will change every time, so i need to use formula for correction. But i can correct 2 efficiencies that differ not more than 5% (as recommended). But actually both reactions are ok - the large decrease of fluorescence causes the increase of slope i think, but why? I know that sybr green + transparent plates are not the best choice for evaluation of 2fold difference in copy numbers (as in my case), but i didnt see such differences in dRn between plates of my collegues.
RN normals shouldn't influence Ct/Cq results, unless the threshold is set incorrectly. Instead, ct results are influenced by template abundance (there are other considerations ofc). Also note that after around 30cts, the amplification for your triplicates may vary. This is a known phenomena.
You should be meticilous about
1) Primer efficiency
2) Triplicate tightness
3) Primer specificity
http://www.bio-rad.com/en-au/faq/Do-levels-of-RFUs-fo_1384541547/normalization-of-real-time-pcr-fluorescence-data-with-rox-passive-reference-dye
Can Kiessling, when i did some weak calculations from this article -
Article Critical Evaluation of Methods Used to Determine Amplificati...
, i found that such differences of dRn between my plates change the eficiencies by about 30%, and as i mentioned above i have differences in ddCq of my control samples between runs about 30% too.
- Badges
- Science method