Standard curve of cDNA

5 posts / 0 new
Last post
lsk
lsk's picture
Standard curve of cDNA

Dear all,

I have to perform a 2-step qRT-PCR & I have also constructed a set of in vitro transcribed RNA. My question is, do I make cDNA from the in vitro transcribed RNA & serial dilute the cDNA to plot a standard curve? I am quite lost because the protocols I've read so far either used plasmid DNA or RNA to plot standard curves. I have a problem with my plasmid DNA - the concentration is very low. Help needed please.

Thanks in advance.

Isk

Ckis
Ckis's picture
You don't need a standard

You don't need a standard curve at all if you calculate in-well efficiency. Besides, the ONLY way a standard curve will provide any information that is useful to your sample of interest is if the efficiency of the standard curve and the efficiency for your sample of interest match up. Unfortunately, most instruments don't do the latter and in many cases, if the source of nucelic acid is different (plasmid vs cDNA from RNA for example), the efficiency could very well differ. It is an issue that has not been addressed in the scientific community.

Look in pubmed for the following reference and you will see what I am talking about. I don't use standard curves any more. Eliniating them saves a ton of material and gets you better results anyway.

Quantitative real-time RT-PCR data analysis: current concepts and the novel "gene expression's CT difference" formula.
Schefe JH, Lehmann KE, Buschmann IR, Unger T, Funke-Kaiser H.
J Mol Med. 2006 Nov;84(11):901-10. Epub 2006 Sep 14. Review.

lsk
lsk's picture
Dear Ckis,

Dear Ckis,
Thanks for the manuscript. I have read it & found that the method you recommended isn't possible if I wanted to determine copy numbers instead of gene expression. It was mentioned that external standard curve is still the "gold standard" for such applications. The problem faced by me is my nanospectro quantified cDNA doesn't match my qPCR quantitation. I have low detection in qPCR when nanospectro showed at least 4-logs more product. What could be the reason for this? However, the methods from the article does provide us with a better alternative in gene expression analysis. In your experience, which is your preferred software, LinRegPCR or DART-PCR?
 

Ckis
Ckis's picture
     Truthfully told, it isn

     Truthfully told, it isn't really possible to determine copy number accurately unless you process your sample, convert RNA to cDNA and do the qPCR reaction all in the same well without removing anything. If you think about it, doing all kinds of washes, etc gradually removes material and your return is never 100% after every single step. If you took a 10cm plate of cells out of culture for example and processed it with a QIAGEN kit for RNA, you would get back about 80-90% of the original RNA if you are really good. Even at this point, you are down from the true starting material. cDNA conversion is also not 100%. Even the best reverse transcriptases convert 90-95% of your RNA into cDNA. Stuff like a generic Promega MMLV is terrible and only has a 15-20% conversion rate. Anyway, just imagine how all these steps gradual reduce your sample from the true "copy number". For this reason, any value you get from copy number is only relative to the standard curve and that's it. Odds are, that number is not really your true value.
     Here's another catch too.....standard curves only work if the efficiency value of the curve matches the efficiency of your sample. A difference of 5% is enough to throw off your value by 2 fold. The greater the percent difference, the more exponentially off you will be. I have done some research to show efficiencies don't always match, especially if the material you use for the standard curve is not the same as your sample. But if you are going to use your sample for a standard curve (which may not even be possible depending on how much material you have), then you might as well just do in-well efficiency analysis anyway if you can. Efficiency is a BIG factor and much of the scientific community is unaware of the things that need to be done and done right to really get accurate numbers. A biostatistician and I are going to be publishing some papers soon on all this. An external standard curve may be the current "gold standard" but I view it as a lemming effect...it is a gold standard because everyone does it but that doesn't necessarily make it right.
     Anyway, the fact that your quantified cDNA and qPCR quantification doesn't match up is in all probability due to a lot of the above factors I mentioned.
    Lastly, I don't use either of the software you mentioned. I designed my own excel spreadsheet but the analysis is based on the Schefe paper methods.

lsk
lsk's picture
Thanks Ckis. Could you please

Thanks Ckis. Could you please email me your papers on efficiency once they are published?