It's controlled for by an internal control primer set that is used in both samples. The control primer set will likely be different depending on what you are looking for (cDNA, gDNA). There are also different ways of calculating your results following a real-time PCR but I will describe it in terms of the delta-delta-CT method.

For example, if you were interested in determining the copy number of gene X in a particular sample (sample A), you would need to compare it to a control sample for which you know that gene X has two copies (sample B). Now, there may be a small difference in the starting concentrations when you do the RT-PCR, which can be introduced by minor pipetting, dilution error etc, but you should aim to start with similar amounts of DNA or cDNA. You also use a set of control primers for which copy number is known to be 2. A common control gene for copy number validation is GPX7.

Let's exaggerate the differences and assume sample A is twice as concentrated at sample B, but let's assume that gene X actually is 2 copies in both samples. If you ran the real-time without an internal control, it would appear that sample A would reach a threshold cT value 1 cT value before sample B and without a control you would deduce that sample A has 4 copies of gene X and sample B has 2 copies of gene X, which would be incorrect. If you ran an internal control, even with a difference in the starting concentrations, you can use the delta-cT value between the internal control cT and the gene of interest cT and compare those values to the delta-cT of your control sample (sample B).

If delta-cT for sample B (cT GPX7 - cT geneX) subtracted from deltra-cT for sample A (cT-GPX7 - cT gene X) is 0, then you know there is no difference in the copy number of the two samples and they are both 2 copies. If the difference in the delta-cT is positive, say +1, then the copy number of gene X in sample A is 4 copies, and if the number is -1, then it is 1 copy.

This type of method can also be used to determine fold change gene expression between two or more samples.

I hope this helps but I can see how my explanation can be somewhat confusing even though the calculation isn't presented exactly as it would be done to generate the data.

It's controlled for by an internal control primer set that is used in both samples. The control primer set will likely be different depending on what you are looking for (cDNA, gDNA). There are also different ways of calculating your results following a real-time PCR but I will describe it in terms of the delta-delta-CT method.

For example, if you were interested in determining the copy number of gene X in a particular sample (sample A), you would need to compare it to a control sample for which you know that gene X has two copies (sample B). Now, there may be a small difference in the starting concentrations when you do the RT-PCR, which can be introduced by minor pipetting, dilution error etc, but you should aim to start with similar amounts of DNA or cDNA. You also use a set of control primers for which copy number is known to be 2. A common control gene for copy number validation is GPX7.

Let's exaggerate the differences and assume sample A is twice as concentrated at sample B, but let's assume that gene X actually is 2 copies in both samples. If you ran the real-time without an internal control, it would appear that sample A would reach a threshold cT value 1 cT value before sample B and without a control you would deduce that sample A has 4 copies of gene X and sample B has 2 copies of gene X, which would be incorrect. If you ran an internal control, even with a difference in the starting concentrations, you can use the delta-cT value between the internal control cT and the gene of interest cT and compare those values to the delta-cT of your control sample (sample B).

If delta-cT for sample B (cT GPX7 - cT geneX) subtracted from deltra-cT for sample A (cT-GPX7 - cT gene X) is 0, then you know there is no difference in the copy number of the two samples and they are both 2 copies. If the difference in the delta-cT is positive, say +1, then the copy number of gene X in sample A is 4 copies, and if the number is -1, then it is 1 copy.

This type of method can also be used to determine fold change gene expression between two or more samples.

I hope this helps but I can see how my explanation can be somewhat confusing even though the calculation isn't presented exactly as it would be done to generate the data.

The answer of Pawel is excellent. I'd suggest to try more than one primer pairs, as many housekeeping genes are affected from the experiment. You can see a previous question in my questions about the optimum housekeeping gene, however i suggest 18SrRNA.

Pawe's answer is very good, but I would suggest for calculations a method that takes into consideration the amplification efficiency of your primer pairs. With the delta-cT method you are assuming that your amplification efficiency is the same for all primer sets and this could not be always the real situation. I suggest this book chapter as an introduction: http:http://www.researchgate.net/publication/242568376_relative_quantification/file/72e7e51e56be24e05a.pdf

I will also suggest that you take into considertion the MIQE guidelines: http://www.gene-quantification.de/miqe-bustin-et-al-clin-chem-2009.pdf

Good luck.

Article A new mathematical model for relative quantification in real...

If you are able to measure the concentration of your samples, then you should be able to use a constant amount for your qPCRs. A further and often error-prone normalization with reference genes woulnd't be required then.

Thanks all for great answers. PB, will it be OK if I run standard curves for a control sample using internal control gene primers as a starting reference to my experiments to follow? Thanks very much again.

You should run a standard curve with dilutions for all primers as it can be used to calculate the primer efficiency (mentioned in other answers), which is important in interpreting and calculating your final results. Ideally your primer efficiency is 100%, meaning that each cycle (irrespective of the starting quantity of DNA), the signal intensity is doubling. This is not always the case, but choosing primers with near 100% efficiency is best. Other than the delta-delta cT method I mentioned, most if not all software on "modern" real-time PCR machines should have built-in algorithms for interpreting your results to generate RQ values using your standard curve data.

Even if you are able to measure your starting concentrations very accurately, you should always run an internal control primer to control for variations in pipetting when diluting or minor variations in concentration (NanoDrop is a good quick way of determining concentration but it's not very accurate). Of course there are potential differences in housekeeping genes so you need to be sure to select the proper control genes depending on what type of RT-PCR you are doing. If you wanted to be extremely fancy, you could always use multiple controls but I think for most applications that is not necessary as you could validate your RT-PCR results by other methods as well.

Good Luck!

If you know the concentration of the samples, then adjust them to the final concentration which you want to use (with DNAse free water). If you want to be sure that the concentration of different samples is similar after adjustment, measure the concentrations of different samples again using Nanodrop. Then you can use similar amount of each sample for your qPCR.

I agree with several on this, determine which sample has the lowest concentration, and then dilute all other samples to that concentration. Then you can add the same volume of sample to reach reaction, this makes setting up reactions a lot easier and your results will be more accurate.

Just to mention: it is not even required that you adjust all samples to same same concentration. As long as you know their concentrations, you can normalize the Ct values. This simple calculation is easier and quicker than adjusting concentrations by pipetting (where you have to do some calculations, too).

One caveat not yet addressed here is the idea of a 'sample-induced qPCR inhibition threshold.' Overly-concentrated samples can inhibit the reaction, so you will want to run a dilution profile of a representative sample (or, better yet, a representative sample mixture) to determine where this sample-introduced inhibitory threshold is (if it indeed exists in the first place for your particular sample set). Then, if you find the qPCR is indeed inhibited up to a certain point, this indicates that samples have to be diluted to at least to or slightly beyond that inhibitory point to avoid inhibiting the qPCR. This would provide the rationale for physically-diluting the samples before running them; wherein, if you're doing that, you may as well (after diluting just past the inhibitory threshold dilution) then dilute all to the same concentration (based on the least concentrated sample - if that limiting sample is not too low of a concentration to necessitate extreme dilutions of all others). Look for inhibition beforehand in all situations. Then, if you find inhibition does not exist, what Jochen has stated should hold true.

The transformation of Cq values would then be:

adjusted Cq A = observed Cq A + log base Eamp of (concentration of sample in reaction A/concentration in any other chosen reaction B to serve as the normalization point for all reactions).

Eamp is not always "2" for all targets, an 'estimate' of which can be obtained from the slope of the non-inhibited (good, log-linear) range of the dilution curve developed for each target using a representative sample mixture at the outset.

To clarify "overly-concentrated samples' in my above post - what I mean is two-fold:

1.) Carry-over gunk from sample RNA prep can inhibit the Reverse Transcriptase enzyme during RT, and the cDNA prep in turn can inhibit the 'Taq' or 'Taq-like' enzymes during the PCR. Or with one-step RT-qPCR (where RNA sample is added directly to the reactions), carry-over sample gunk in the RNA (if the RNA is not diluted to at least to or just beyond the inhibitory threshold up front) can inhibit both enzymes in the same reaction well,

and,

2.) Template inhibition of 'Taq' or 'Taq-like' enzymes during the PCR on account of there being too much target in the reaction (most often happening with things like 18S RNA or other hyper-abundant targets).

Again, one can alleviate the blindness going into these assays at the outset by running all targets on an extensive serial dilution profile of a representative sample mixture to see whether or not inhibition rears its head. Knowing this up front allows one the opportunity to side-step calamity before it runs you over from a blind-spot.