Quantitative PCR

William Alexander ALEXANDERW at cber.cber.fda.gov
Wed Sep 6 09:35:06 EST 1995


In Article <Pine.ULT.3.91.950831151350.24872C-100000 at rac5.wam.umd.edu>,
Frank Chen <yatsen at wam.umd.edu> wrote:
>
>        The one that argues for it always says: This has appeared in such 
>such publications and now patents, or this has been highly evaluated by 
>reviewers.  What is the scientific evidence? - Controls?

I do see the limitations of this forum now.  I can't scan the results in
easily yet (+ I'm a slow typist) and you have to actually go and get papers
(not yet on-line) and read them.  Yes, there are a lot of bad quantitative
PCR papers out there but this does not mean that the technique does not work
with the proper controls.  (Such as:  measure your results before the
plateau unless you want to add even more controls.)

>        The one argues against also has some sort of scientific 
>arguments.  For example, Dr. Miller's points in his reply are worth 
>citing here.
>        I will give you my simple controls.  I frequently did two 
>"identical" PCR reactions starting with the "same" amount of templates.  
>Most likely, I got different amplifications.  Which one had more template to 
>start with?
>        Suppose you have two samples for quantitative analysis by RT-PCR, 
>one gives you stronger signal than the other?  Are you going to be 95% 
>sure that there is more specific mRNA in the one that has stronger signal?

Try serial dilutions.  Bunn et al. (US. Pat # 5,213,961 written 1989) state
when talking about the necessary controls and the problems of conventional PCR:
"Further, even when these parameters are controlled precisely, there is
tube-to-tube variation which precludes accurate quantitation.  For example,
significant differences in yield occur in PCR samples which are prepared as
a pool and then aliquoted into separate tubes and amplified in the same run."
They go on to state that a competitive standard can overcome the problems.

>I will hesitate to conclude myself.  However, I will be quite confident 
>to conclude such a result if I repeat 3 times by convential Northern 
>and/or insitu analysis.
>
>        Sorry, if this offends you.

No problem, this is science, not politics.  I do remember this same
discussion taking place with regard to reproducible Northerns about 10 years
ago and by using the correct controls this method is now in wide use.  
>
>Frank Chen
>
>
>On Thu, 31 Aug 1995, William Alexander wrote:
>
>> In Article <41vo24$ok2 at miasun.med.miami.edu>, tmiller at newssun.med.miami.edu
>> (Todd Miller - Pharmacology) wrote:
>> >In article <thomas.kreuzer.62.0012F0A2 at kfunigraz.ac.at>,
>> >Thomas Kreuzer <thomas.kreuzer at kfunigraz.ac.at> wrote:
>> >>On Thu. 24 Aug Frank Chen wrote as a reply to a question concerning 
>> >>quantitative RT-PCR:
>> >>
>> >>" No matter how you do it and how good your controls are , your results
from 
>> >>quantitative PCR will be questionable."

No matter how?

>> >>
>> >>Since I am trying to put up a quantitative RT-PCR - method myself (and
being 
>> >>quite new to the method of PCR in common) I would be interested in the
major 
>> >>reservations against the principle of quantitative PCR and especially 
>> >>quantitative RT-PCR. What exactly makes itYs results questionable and why 
>> >>canYt these problems be avoided ?
>> >>It would be a great help if you could also give me some advice about 
>> >>literature dealing with these problems.

Try: Ferre, Francois.  "Quantitative or Semi-Quantitative PCR: Reality
Versus Myth"  PCR Methods and Applications (1992) Vol. 2, No. 1,  pp 1-9.

>> >>
>> >I would highly recommend reading an article by Luc Raeymaekers in
>> >_Analytical Biochemistry_ (I think it is volume 214, 1993, but you
>> >may have to check your favorite citation source).  He goes through
>> >all the mathematical theory of quantitative competitive PCR and comes
>> >to the conclusion that there are many papers published that have
>> >ignored some basic theory and predictions, which makes the data
>> >very suspicious.  Basically, the assumption that your competitor and
>> >your target have equal amplification efficiencies is almost never
>> >borne out, and this results in huge errors by the time you go through
>> >all the cycles.  

The Raeymaekers article deals almost exclusively with the situation in which
the  target sequence (T) and the standard sequence (S, the competitor)
amplify with different efficiencies.  Raeymaekers' math shows how even a
small difference in the efficiency of amplification (E) of the two templates
(T and S) can result in large errors at when the end of the reaction (not so
bad near middle of reaction if the E's of S and T are close).  A big problem
noted by some authors and ignored by others.  Raeymaekers hints that the
quantitation is possible with additional experiments (middle of last
paragraph).  This paper is really for those amplifying T and S with
different efficiencies.   Raeymaekers never addresses the situation in which
the standard differs by 1 base from the target.  In this case (and some
others) E should be the same and the products can be labeled during the
reaction, stopped at an appropriate point, restricted and run on a
sequencing gel and quantitated with radioimaging (mass spec may be possible
soon).  Is there a problem with one base difference creating a new
restriction site? 

<Snip>
>> >
>> >Todd Miller, PhD
>> >University of Miami School of Medicine
>> >tmiller at newssun.med.miami.edu
>> >
>> > 
>> >>Thanks in advance,
>> >>
>> >>Tom Kreuzer
>> >
>> >
>> Todd Miller, Did you read Isaac Kim's Post?  He is correct (great arguments
>> and the experiments were done) and so is Tracy.  Quantitative PCR does work,
>> you should not use a "house-keeping" gene (GAPDH) as your competitor and the
>> source of the nucleic acids has nothing to do with this topic.  
>> 
>> Todd said:
>> Basically, the assumption that your competitor and
>> >your target have equal amplification efficiencies is almost never
>> >borne out, and this results in huge errors by the time you go through
>> >all the cycles.  
>> 
>> The whole idea is to design the competitor so that it does amplify at the
>> same efficiency as the target.   (See Celi et al., "A rapid and versatile
>> method to synthesize internal standards for competitive PCR" Nucleic Acids
>> Research, 1993, Vol. 21, No.4, 1047.)  Yes, Todd is also correct, if the two
>> do not amplify at the same rate, the paper should be ignored.  
>> 
>> There are at least two simple ways to design good competitors:
>> 1.  Use the "same" product as your competitor by adding or removing a
>> restriction site so that the two products can be distinguished an a gel
>> after restriction.  (This extra restriction step can be a pain.)
>> 
>> 2.  Use a competitor that is slightly different in size to the original
>> target so that the only sequence difference is this extra Nucleic acid
>> fragment. 
>> This is the case where Todd gets caught up in mathematics.  The enzymes used
>> for the amplification are extremely fast (do you need references?).  
>> Therefore, the size difference can be designed so that it is detectable on a
>> gel but does not change the efficiency of the amplification of the target
>> and the competitor.
>> 
>> Two US Patents (there may be more) have been issued in this area.
>> 1. Wang et al., Patent # 5,219,727, Date of Patent:  June 15, 1993, Filed:
>> Sep. 28, 1989; "Quantitation of Nucleic acids using the polymerase chain
>> reaction"
>> 
>> 
>> 2. Bunn et al., Patent # 5,213,961, Date of Patent:  May 25, 1993, Filed:
>> Aug. 31, 1989; "Accurate quantitation of RNA and DNA by competitive
>> polymerase chain reaction"
>> 
>> These patents are great scientific papers and should be available around the
>> country (Frank Chen, U of MD College Park should have them in their lib.). 
>> As you can see, they were both filed in 1989 and issued in 1993. They are
>> both too "old" to be included in the full text biotech patents available on
>> the WWW.  (See:
>> http://www.inform.umd.edu:8080/EdRes/Topic/AgrEnv/Biotech/Biotechnology_Pate
>> nts-full_text)
>> 
>> I just don't think this area is so new anymore.
>> 



More information about the Methods mailing list