qRT-PCR: Data analysis

Question

[johannes.berstein]

Hi everyone,

I am trying to analyse data from qRT-PCR. My experimental design was the following: I have two groups of cells (A and B) and those were infected with a virus. What we want to know now is the fold change of B in comparison with A. I believe that a relative quantification using the normal methods are not possible since the reference and calibrator are not expressed normally in these cell (only upon virus infection). Would someone have any suggestion?

Thank you for your helping

Jo

[VHahaut]

If it is a virus integrating to the DNA you could use a cell line that has one provirus per cell as control.

[johannes.berstein]

Hi, Unfortunately it is not a DNA virus. :/ Regards

Answer 1

[Devon Ryan]

You have two options:

1. Use absolute quantitation (this will require cell counting).
2. Find a useful control gene (or more than one of them) that can be used for comparison and then use that for relative quantitation.

[johannes.berstein]

Hi Absolute quantification is not possible once we dont have a standard curve. As control gene I use beta actin to normalize the genes. Even though the results seems to be quite strange after analysis but our raw data shows clearly a huge difference. :/ The normal methods ddCT or 2^-ddCT seems not to work in this case as mentioned before due to the absence of a calibrator and the target gene is only expressed if the cells are infectef (viral gene)

Grüße

[Devon Ryan]

You can always make a standard curve...

[johannes.berstein]

Ja sure,

Just an aditional question: is it acceptable to calculate the fold change like this: 2^(CT target / CT Beta actin). If so this would give us results acceptable results and in accordance to the CT values.

[Devon Ryan]

No, 2^(CT target/CT Beta actin) would produce wrong results.

[johannes.berstein]

Ok, thanks

Any other suggestion? I am wondering if I throw this set of experiments.... :/

[Devon Ryan]

Nope, if your reference gene isn't reliable then you'll need to redo the experiment.

[johannes.berstein]

So, Ok thanks my reference gene is really reliable and I have few variation... the whole problem is that I cannot apply the normal formulas available because this gene is not expressed in normal cells (only virus infected ones)

[Devon Ryan]

Sure you can apply the normal function, you just give the "failed detection" samples an arbitrary Ct value (e.g. 40) and make sure to use a non-parametric test.