Endpoint RT-PCR for quantification of mRNA?

Duncan Clark blackhole at abuse.plus.com
Thu Jun 1 10:56:27 EST 2006

Historians believe that in newspost 
<1149152614.170292.173910 at u72g2000cwu.googlegroups.com> on Thu, 1 Jun 
2006, paul_wary at yahoo.com penned the following literary masterpiece:
>Austin P. So (Hae-Jin) wrote:
>> paul_wary at yahoo.com wrote:
>> > As far as I know even under apparently equal or very similar conditions
>> > the variations in efficiency of reverse transcription and amplification
>> > are too big to allow quanititating mRNA levels that way. And using a
>> > housekeeping gene for normalization of the PCR step would still be
>> > insufficient because there is also variation in the reverse
>> > transcription step.
>> These issues you bring up apply whether or not you do normal PCR or
>> real-time PCR.
>I was thinking about this again and you are absolutely right. But
>assuming that the efficiencies for the reverse transcription of the
>housekeeping gene and the other genes of interest are approximately
>equal, the housekeeping genes should also provide some normalization
>for the reverse transcription step (after all, it is also used for this
>purpose in real-time RT-PCR, isn't it?). Yet, whether this assumption
>is true is difficult to determine.

The problem is that there is no such thing as a true housekeeping gene. 
It is more normal now to term them reference genes, to pick several and 
then use appropriate software to do the normalisation.

>> I think the thing to keep in mind is that there is no fundamental
>> difference between real-time PCR or just regular "end-point" PCR,
>> particularly if you do not hit the plateau phase of the reaction by
>> keeping the cycle number low, and you are interrogating a single transcript.
>OK, I interpret this in the way that there is no such thing as an
>"end-point" PCR for quantification of mRNA if "end-point" means
>amplification until the PCR hits the plateau phase.


> You have to stay
>within the exponential amplification phase.


> There is no correlation
>between the height of the plateau ("end-point") and the intial amount
>of cDNA, right?

Agree. If one looks at a real time SYBR plots the final plateau 
fluorescence varies amongst replicates, never mind different template 

>How low should the cycle number be? I know this depends on the initial
>amount of cDNA; with less cDNA the cycle number can be higher without
>reaching the plateau, but are there any estimates?

How long is a piece of string?

Depends also on primers used for cDNA synthesis; random hexamers, 
oligodT, specific primers etc. and then amplification efficiency for the 
primer pair chosen.  Add on what the transcript copy no. might be and it 
is difficult to suggest the maximum no. of cycles to use.

>> The advantage of real-time is that you do not have to run a gel, and it
>> doesn't matter if you extend your run beyond the plateau phase, because
>> you have all the low-cycle information stored already to get relative
>> abundances. You can also get "absolute" levels by having a standard
>> curve (which you could also do using an end-point method, but it would
>> be tedious as hell).
>Does the type of DNA detection method matter for the quantification?

SYBR detection is slightly more sensitive than Taqman but SYBR will also 
show dimers whilst Taqman won't.

>this case, the gel was stained with ethidium bromide (EtBr) and the
>bands were quantitated densitometrically. Does the intensity of EtBr
>stained bands correlate well with the amount of DNA in a gel band?

Yes. EtBr, SYBR Green I, SYBR Gold correlate well with amount of DNA. 
All have been or are currently used for real time PCR.

I love deadlines. I especially like the whooshing noise they make as
they go flying by.

Duncan Clark
GeneSys Ltd.

More information about the Methods mailing list