RT-qPCR plate set up NoRT controls - (Aug/15/2012 )

Dear all,

I'm quite new to RT-qPCR and I need some help.

I'm running RT-qPCR using SYBR Green. I use 96 well plates to run my qPCR reactions. However I need to save time and space in the plates to run as many samples as possible. I run them in triplicates(= "x3") by decision of my supervisor.

So, in a 96 plate I've been using 2 (x3) wells using a standard of a gene that is present in every plate I have and the another well (x3) for th NTC no template control.

Let's say I have 4 ref genes (gene1-4) to test and 3 different biological samples (A, B, C), from which I got RNA->cDNA samples from 3 different experimental time points (e.g. A1-3) (Fed-batch culture).
My question is how many NoRT (No reverse transcription) controls should I have in a plate.
Should I have 1 NoRT control per cDNA template that I did RT for or 1 NoRT control from each original biological sample?

(Each rectangle represents 3 replicates, meaning that this represents half-plate)

https://docs.google.com/spreadsheet/ccc?key=0AgexH37wJvUIdGZ5R3JaZUx5b0Y2T0FacmpHYThnbFE


Which plate set up is more correct (see link above)?

*Another question:
When measuring relative gene expression, do I have to have the reference genes in the same plate or it can be done in the sample plate? (As I mentioned in every plate I have a calibrator "ST1 and ST2"
Then I just make a ratio between the copy numbers of target gene/copy numbers of ref genes and then plot the ratios in a bar graphic?

Please ask if you couldn't understand what I meant.
Thank you in advance!

BR
lylah

Ideally you need RT- for each RNA you isolated (and hopefully treated with DNase) and each gene (some genes may detect DNA less than others) so in case of positive values you can substract them from the Ct you measured in your sample. (why? because every RNA you isolate contains it's own private amount of DNA that can screw your qPCR results, so you have to test every RNA, that you get rid of it right)
However, you can also have a standard, optimised, verified, DNase treatment protocol when you have stably negative values for RT- all the time. In such case you may decide to only include one RT- control for selected gene and sample to check if the DNase procedure is still OK. But in case of positive value in RT- , you can't use your results, because you don't know how it affected other samples.
Or you can have primers designed on exon-exon boundaries that don't detect DNA at all, like I have and thus have no RT-. It's also possible to check the melting for the presence of second "DNA" peak, but this method as those two before only can be used for checking, not for getting some data in case you have positive RT-.

If you don't know how your primers bind DNA and how your DNase procedure works, and it's one of your first experiments, I suggest having complete set of RT- controls. Better also in triplicates, in case you cross-contaminate one.

About the plate setting.. it was mentioned here many times, if you do relative quantification it's not necessary to include all GENES in one plate, but you always need to have all SAMPLES for one gene on the same plate.

But first some terminology for better understanding:
- in relative quantification which is what you're most likely are doing you need to compare your samples with a control sample called calibrator (that's probably a different thing than the one you mentioned), you have to decide what the calibrator is, in case of different samples and different timepoints I can hardly suggest when I don't know the experimental setting, the calibrator may be the first timepoint in each sample or one of the samples, unless you write more details on what you are comparing I can't tell, anyway.. calibrator is run as every other sample, on all genes you run (i.e not just once on a plate)
you have to decide what you actually want to compare with what before you even start pipetting something!
- relative quantification uses two types of genes, one we call target, those are the genes we are interested in, and then there are reference genes, that are used for normalisation of cDNA amount, you can have more targets and normalise to multiple references too, choosing the right reference genes for normalisation is a huge question itself, so I won't go further into it now, you are talking about "4 ref genes" and didin' mention any targets, either you call them wrong of you forgot to mention

You can run each gene with a complete set of samples, calibrator, NTC and RT- controls on separate plates. You don't need to include some interplate "calibrators" for that, it doesn't matter, all genes are then compared relatively within their sample set. If you have so much samples, that they don't fit on one plate even for single gene (meaning you have more that 32 samples including whatever controls, which is not your problem if I count correct) you need to place some samples on each plate, to calibrate among different plates.

So about your plate schemes.. do you want to compare samples A1,2,3, together with B1,2,3 or not? If not (like in case you actually only want to compare samples A2 and A3, to the first A1 timepoint, and then separately samples B2,3 to B1, etc.) you can put on one plate A1,2,3 their RT- and NTC, as on the first example and continue with other genes antil you run out of wells (just leave out the ST1,2 interplate calibrators or whatever it is).
If you want to compare changes between As and Bs, and Cs, you have to put them on the same plate, so none of your examples is right.


I'll just stop now and let you tell me what you think, what is going to be your calibrators and what you want to compare together? I'm sorry it's long, but it's a bit complicated topic to explain.