Basic question about qPCR and endogenous RNA control. - (Jul/14/2013 )

Let say I'm want to compare the gene Y expression for two different cell populations, cell A and cell B. I'll be using 18S rna as my endogenous control. Will the expression of the 18S gene differ in these two different populations when using about the same concentration of rna? We are not comparing completely unrelated cell populations, they are similar but differ slightly (e.g. something like EpCAM low vs EpCAM hi).


I recently performed a qPCR for three different cell population and the ct values for 18S differed quite a lot (e.g. something like ~19, ~22 and ~25). A colleague insists it is my fault (pipetting error) but I just have a hard time believing so for several reasons. Can anyone provide their opinion on this?


One other question, let's say I performed a qPCR on three different cell populations (A = GFP-, B = GFP low, and C = GFP hi) and used 18S as the endogenous control and GFP is my target gene. To analyze my results, am I right to think that I should calibrate the machine this way:


Population A GFP ---> Population A, B, C 18S and GFP samples (basically using one population control for all the samples).

Or should I be using Population A 18S as the calibrator for every sample?

I'm using the Applied Biosystems 7500 Real-Time PCR System if it matters.

You need to input the same amount of RNA to exclude that variation from analysis. If you for example input 1ug of RNA in all reactions, but then get varying Cq values in the qPCR you can say that 18s expression is differently expressed in those various cell populations. Furthermore if you are planning to use 18s for your normalisation, you will have a hard time to do so...
But have all those PCRs been run on the same plate with the same mastermix etc?

Re your 2nd question.

Not 100% sure what you are saying there. But you normalise all samples against the reference gene. And then you reference group B and C against group A for example. So group A would be 1 or 100%, Group B and C would be a then be and up/equal/down compared to your Group A.

Podge on Mon Jul 15 11:43:10 2013 said:

All the PCRs were ran on the same plate with the same mastermix. Nothing changed between the samples.
I used 6.4, 6.4, 7.8 and 10.4 nanograms per reaction (I was told that difference in concentration is not significant) and the ct values I got were:

Sample - - ct

P6 - 7.8ng - 31.591
P9 - 6.4ng - 19.806
P12 - 6.4ng - 15.919
P13 - 10.4ng - 26.637

If it's a concentration problem I would expect to see the ct value for P13 to be the lowest since it technically has more RNA but that's not what I get.

Would GADDH or HPRT1 be a better reference gene than 18S? My PI has been using 18S so that's why I am using it.

-Wek-

Wek on Mon Jul 15 14:55:38 2013 said:

Podge on Mon Jul 15 11:43:10 2013 said:

You need to input the same amount of RNA to exclude that variation from analysis. If you for example input 1ug of RNA in all reactions, but then get varying Cq values in the qPCR you can say that 18s expression is differently expressed in those various cell populations. Furthermore if you are planning to use 18s for your normalisation, you will have a hard time to do so...
But have all those PCRs been run on the same plate with the same mastermix etc?

Re your 2nd question.

Not 100% sure what you are saying there. But you normalise all samples against the reference gene. And then you reference group B and C against group A for example. So group A would be 1 or 100%, Group B and C would be a then be and up/equal/down compared to your Group A.

All the PCRs were ran on the same plate with the same mastermix. Nothing changed between the samples.
I used 6.4, 6.4, 7.8 and 10.4 nanograms per reaction (I was told that difference in concentration is not significant) and the ct values I got were:

Sample - - ct

P6 - 7.8ng - 31.591
P9 - 6.4ng - 19.806
P12 - 6.4ng - 15.919
P13 - 10.4ng - 26.637

If it's a concentration problem I would expect to see the ct value for P13 to be the lowest since it technically has more RNA but that's not what I get.

Would GADDH or HPRT1 be a better reference gene than 18S? My PI has been using 18S so that's why I am using it.

That is interesting that the Cq values don't correspond to the amount of RNA you put into the reaction. Is that a 1step RT-qPCR you are doing btw? But I would try to input same amounts of RNA and see what happens. but it does look like 18s RNA is regulated in those different conditions you are comparing.

Gapdh or HPRT1 could work as reference genes. You need to test. i am sure there is a paper out there comparing different references genes using geNorm etc. for your type of tissue. Have a look what other people use.

-Podge-

Well

Podge on Tue Jul 16 14:05:37 2013 said:

That is interesting that the Cq values don't correspond to the amount of RNA you put into the reaction. Is that a 1step RT-qPCR you are doing btw? But I would try to input same amounts of RNA and see what happens. but it does look like 18s RNA is regulated in those different conditions you are comparing.

Gapdh or HPRT1 could work as reference genes. You need to test. i am sure there is a paper out there comparing different references genes using geNorm etc. for your type of tissue. Have a look what other people use.

I just realize something. Using 18S as the reference gene in this particular experiment is fine. The cells have not been treated with anything. The four populations were sorted from C57BL6 mice and the RNA was isolated immediately. I will do a test to see if GAPDH and HPRT1 give me variable Ct values.

This is a 2 step RT-qPCR. I treat the RNA with random hexamers first, make the cDNA, and then do the qPCR.

I am also having a problem with the Qiagen rna isolation kit (column not holding the buffers properly) so maybe that's the root of it. Anyone experience this problem with Qiagen kits?

-Wek-