quantitative RT-PCR

Tracy Aquilla aquilla at salus.med.uvm.edu
Mon Aug 28 10:14:02 EST 1995

In Article <41kmno$84 at news.acns.nwu.edu>, ikekim at merle.acns.nwu.edu. (Isaac
Kim) wrote:
>I have used quantitative RT-PCR for past publications. I have had no 
>problems with the reviewers. The main problem with RT-PCR is that 
>different primers have different efficiency of amplification. If you 
>choose the conditions of the standard correctly, RT-PCR can be more 
>accurate than Northern blot analysis, in my experience. Even more

This is an interesting topic for discussion. I'd like to know if you have
data to support this claim, or is this merely speculation? Specifically, how
did you (accurately) measure the samples to compare the accuracy of RT-PCR
versus Northern analysis?
>interestingly, a recent manuscript that I submitted to a reputable 
>journal, the reviewers' response was that I should use quantitative 
>RT-PCR over that of Northern blot analysis for all my studies as the 
>control for Northern, GAPDH, showed variations in the level. I would like 
>to direct you to two articles regarding this matter. 
>Siebert PD et al. 1992. Competitive PCR. Nature 359:557-558
>Forster E. 1994. Rapid generation of internal standards for competitive 
>PCR by low-stringency primer annealing.

(I think you forgot to include the actual citation here.)

So, are you (and/or these reviewers) saying the Northern appears to be less
accurate because it showed variation in GAPDH levels between samples? This
is an invalid conclusion. How do know there REALLY isn't variation in the
levels of GAPDH? In other words, maybe this is an indication that the
Northern is actually MORE accurate than RT-PCR.

Furthermore, GAPDH is actually a pretty lousy internal standard. GAPDH
levels vary significantly between individuals and with developmental stage,
as well as in various tissues and with numerous experimental treatments.
Furthermore, many different GAPDH-related sequences have been identified and
there appear to be approximately 300 copies of these in the rat genome and
100 copies in the human genome. To make matters even worse, some of these
pseudogenes are actually transcribed and fully processed (NAR 13:7,
2485-2502). This does not make for an ideal internal control! I'd say the
real problem is with your standard, and not necessarily the method of analysis.

More information about the Methods mailing list