You can choose any one of the stages to be your calibrator; just state which group it is: either the earliest stage or the latest stage (whichever stage gives non-zero expression for all targets).

Thank you all for your answer

I found that in my case the calibrator would be the sample with the highest Ct value.

This pdf. helped me so much

http://www.gene-quantification.de/lanz-qpcr-tutorial-bcm-2009.pdf

Heba as per your attached PDF, your calibrator would be sample with lowest Ct value.

please correct me if i am wrong,, i too had similar queries, my appreciation.

Below lines are from your PDF. page 9 

Comparing Samples with a Calibrator

The normalized amount of target (c-mycN) is a unitless number that can be used to

compare the relative amount of target in different samples. One way to make this

comparison is to designate one of the samples as a calibrator. In Table 1 on page 10,

brain is designated as the calibrator; brain is arbitrarily chosen because it has the

lowest expression level of the target.

Ref: User Bulletin #2: ABI PRISM 7700 Sequence Detection System Page 9 of 36

Dear Heba,

Suresh is right. your calibrator should be had lowest Ct. in this way, you can use comparative method using 2-deltaCT method.

Lowest expression level = highest Ct value.  Expression level and Ct value are inversely correlated.

Dear Mr. Suresh Oncosuresh

Sorry for being late in answer

please return to the attached pdf page number 18 you will find the following statement:

Calibrator cb may be a control treatment, or the sample with the highest Ct value.

The sample with the highest Ct value means that it has the lowest expression level since Ct value and expression levels are inversely correlated 

Dear all

Because I do real time PCR first time, I dont know too much about this.

How do we know which one will be highest Ct value or lowest expression level before we do real time PCR?

The machine that I use from Appied Biosystems 7500, They require to detail reference sample or calibrator and reference gene before do real time PCR but because of I dont know the expression of them how I can detail them?

And I would like to ask about this machine, Is it impossible to do real time PCR saparately between target gene and reference gene? Because of I have too many samples if I do reference gene same plate with target gene, I afraid that well it is not enough for all stage.

Thank you

Dear Karen,

In the case of your first question, you have to select reference sample (arbitrarily) and reference gene before starting qPCR. It is suggested that the calibrator can be sum of your control samples or cell line that expresses your gene of interest. In the case of reference gene, you must also find the best housekeeping gene with stable expression in your samples (controls and treatment samples) by using genorm/Normfinder or bestkeeper algorithms before starting analysis of your gene of interest.

To get answer for the last question, you can follow below link which supported by Dr. Joe Vandesompele:

https://blog.qbaseplus.com/four-tips-for-rt-qpcr-data-normalization-using-reference-genes

I hope that this explanation will be useful.

best

Dear Yasin Panahi

Thank you very much for your good explanation