Standard curve in RT-PCR

[Kobrus]

Hello,
I have done 8 10-fold dilution series of PCR product to calibrate my amplification using Taqman methodology. Should my samples be in range of the calibration curve or not? To my statistical knowledge, one should get a standard curve and equation based on this and then use it as a standard for any sample range.

Thank you in advance.

[Aga]

Your samples should be within the range of your standard curve. The program I use does not allow me to apply my standard curves to the concentrations outside its range. Besides for quantification purposes you should not extrapolate the data from your standard curves to the concentrations outside its range because you'll get the error you can't really estimate.

[Kobrus]

Thank you for the reply!
I have another question if you may answer. Do you run standard curve together with your samples? I wonder if there is possibility to compile two files (1.samples and 2.st. curve) in MyiQ software? I have got my samples to be proper but the standard curve was not reliable in the same 96-well setup. I am thinking to repeat only standard curve and then compile output file with my 1st one. Or I can copy all information to Excel and perform statistical analysis.

Thank you in advance!

[Aga]

I don't always run standard curve with my samples, at least not in many replicates. Usually if I have the reaction normalized I run one replicate of standard and its dilutions. This gives me the confirmation whether I have similar reaction efficiency to the standard curve previously run (then I use that standard curve as external).
I don't use MyiQ soft do I don't know whether you can compile two files. Surely you can copy all the data to Excel if you have no other choice.

Remember - to use the standard curve as external, for any of your analysis - that analysis must be perform in conditions as much similar to those for standard curve as possible. Having standard curve for each of your separate analysis is ideal, but costly. Make sure you use some standards in your reactions as a reference.

If you see that your standard curve is not applicable to your samples then it's better not to calculate the results on the basis of it. You'll get too big error.

[Kobrus]

I must appreciate your careful answer.
What if I am running my standard and its dilutions for curve and my samples separately which makes the threshold different compare if we run them together simultaneously. We know that the threshold for Ct is set arbitrary at 10 standard deviations above the mean of base-line emission calculated from cycles 1 to 15. The baseline data is not equal for the stand. curve dilutions and samples, thus the thresholds will be different and Cts too.
Can you still analyse them both? Can you manipulate with the threshold levels?

Thank you,
Konstantin.

[Aga]

You can certainly manipulate with threshold levels. You can set them manually - always in the linear part of the reaction (you should be able to see logarithmic view of amplification curves using your software) - the threshold line is usually set high enough to eliminate noise and low enough to be still in the linear part of amplification.

The catch is that for one reaction in one experiment all samples should be analysed with the same threshold line. When having two experiments to compare you should set the threshold line for both of them at the same level - providing it is possible (that is the noise is eliminated in both experiments and threshold line is still in the exponential phase of amplification for both experiments).

[Kobrus]

There is a gap between the noise and plateau phase of the reaction. Manually it is possible move up or down the threshold line in this gap but how to set the same level for 2 experiments? The values for threshold line differ a little.

[Kobrus]

...one more question: do you use correlation coefficient for the PCR higher than 0.99, PCR efficiency higher than 90% and slope between -3.6 to -3?

[Aga]

Ok, lets start with the first questions I use Roche systems, and in fact I can change the threshold values after the PCR is performed, calculate the most important statistics (like mean values, standard deviations, standard error), so I guess I won't be able to help you with your software unless I see it.

The fact is you can't compare two different experiments, even with exactly the same PCR efficiency, having two different threshold levels, cause it affects Cts.

And the second question - yes, usually, yes Seldom, but for some types of samples I have problem with efficiency. In most reactions I get more than 90%, reaching 96% for PCR (not RT-PCR - for that I sometimes get higher values). But for some reactions I constantly get 88-89% and I can't do a lot about it and I don't. This is because for my samples, in which I have some inhibitors, I just have PCR efficiency constantly on the same 88-89% level and I leave it like that. This is just the most relevant to environmental sampes I analyse. I can't improve it on the cost of DNA purification efficiency and I don't care about the amount of the final product. As long as the efficiency is known to me, and it is close to 90%, everything is fine.

Off course there is a problem if your efficiency falls - that reaction conditions are not relevant for exact quantification.

[Kobrus]

Thank you very much for your help, Aga!
Which university are you studying if I may ask this question?