qpcR R package using output from ABI 7900HT (SDS 2.4)

Question

[piotr.grabowski]

Hello,

I am trying to use the qpcR package for R. It offers a few extra things that I don't have in the software from the 7900HT manufacturer such as well-specific efficiency estimation. I read the manual and wrote a script that takes an output file from SDS 2.4 and returns a list of fold-changes using the well-specific efficiency to correct for the differences between Taqman probesets (I can explain in detail why I am doing that, if anyone is interested).

In every well I have a duplex Taqman reaction taking place and I am trying to calculate a fold-change between the two (it's like running the copy-number calling qPCR in cancer etc.).

Now the problem is, all the fold-changes in all the wells turned out to be around 1,0 which is impossible since most of the samples have more than one copy number of analyzed gene. I went through my script and it seems that there are two major possibilities:

1) I am using the efficiency factors for calculating the fold-change between probesets in a wrong way (I am using the first equation on top in this manual, page 73)

OR

2) I am importing the wrong raw data file from SDS 2.4. I am importing the "clipped" data and use the deltaRn values for fitting the sigmoidal model and subsequent calculations (I am not using any extra background subtraction in the package). The package manual states that raw fluorescence data should be used, but SDS 2.4 offers a lot of different outputs and only the "clipped" looked like what I am looking for.

Does anyone have any experience with this package and could help me make some sense out of it ? I never processed qPCR data through R, but would love to get some skills in that area, it seems way more powerful than the GUI programs from machine vendors...

Best regards,
Piotr

Answer 1

[heather.bouzek]

Did you manage to figure this out? Several things occur to me. First, is that you are generating your own curves, but there is no background subtraction. Are you positive? ABI requires the use of the passive dye ROX. There is an "automatic" background subtraction due to this. Which is what kept ABI looking like they were doing a better job controlling background. When I was looking into this (years ago) there was no way to get access to the data before the background was subtracted. Second, at what point are you calculating Cq? Are you doing it through threshold or 2nd derivative maximum calculations? It may make a difference if the there is background that you are not accounting for. 'ratiobatch' seems to require threshold values. I am unfamiliar with the 'clipped' data. What exactly is the data?

As to the actual R. Are you using the 'ratiobatch' program or just the calculations? Next you are doing this without reference genes? How are you controlling for sample input?

If you are still working on this, let me know.