I'm always a believer in using standards (to estimate amplification efficiency) that are most like the samples you will be assessing.

Context of template and matrix interference are highly important in these evaluations since primer efficiency is NOT a constant among the differing template types you present to the qPCR for the same target. Efficiency is a highly important parameter to evaluate and estimate with the utmost care for it determines the base of the log amplifications you massage mathematically into results.

I have spent the past 5 years or so trying to illustrate this point. One such publication is:

http://www.ncbi.nlm.nih.gov/pubmed/19776427

And my book chapter addressing this, mathematically, is:

http://books.google.com/books?id=oXoUkTSbnFgC&pg=PA23&lpg=PA23&dq=%22Jack+Gallup%22&source=bl&ots=VWbyoflgIQ&sig=L6jm7m1Xdhwm5Yc1TqNbAHIEQhg&hl=en&sa=X&ei=AUJhUOWdFLTDyAHl2IDoCw&ved=0CEMQ6AEwBjjIAQ#v=onepage&q=%22Jack%20Gallup%22&f=false

Cycles 1 thru 5 of any PCR, determines what happens next in the rest of the reaction. If the reaction doesn't get up on its feet due to matrix interference or non-linearized plasmid being used (for example) during cycles 1 to 5, you will have a hard time reconciling biological template amplifications (and efficiencies of amplification) with the chosen template used for standard curves if the template used for standards is different in kind and quality from the samples you purport to assess.

The context in which you present the target to the qPCR is highly important in my opinion, and thus, if examining cDNA, use cDNA (e.g. a cDNA sample mixture) for your standards.

It depends what do you want to do after.

If you're planning to quantified on cDNA use cDNA for your dilution series.

Then I would add that if you're only doing relative quantification, knowing your primer efficiency isn't really crucial (knowing their specificity is more important) as all your samples would be exposed to the same primer efficiency...

Cheers

A

Thank you for your answer Antoine

But, besides I want to know if template matrix is also relevant to choice among cDNA, gDNA or plasmid DNA. I mean, plasmid DNA matrix is less complex than cDNA or gDNA matrix. So we could expect obtain higher efficiency values using plasmid DNA than using cDNA or gDNA.

I think the best option is use cDNA to make serial dilutions to estimate amplification efficiency. However I need to be sure before to procede.

What do you think about it?

Thank you again!

Regards

I'm always a believer in using standards (to estimate amplification efficiency) that are most like the samples you will be assessing.

Context of template and matrix interference are highly important in these evaluations since primer efficiency is NOT a constant among the differing template types you present to the qPCR for the same target. Efficiency is a highly important parameter to evaluate and estimate with the utmost care for it determines the base of the log amplifications you massage mathematically into results.

I have spent the past 5 years or so trying to illustrate this point. One such publication is:

http://www.ncbi.nlm.nih.gov/pubmed/19776427

And my book chapter addressing this, mathematically, is:

http://books.google.com/books?id=oXoUkTSbnFgC&pg=PA23&lpg=PA23&dq=%22Jack+Gallup%22&source=bl&ots=VWbyoflgIQ&sig=L6jm7m1Xdhwm5Yc1TqNbAHIEQhg&hl=en&sa=X&ei=AUJhUOWdFLTDyAHl2IDoCw&ved=0CEMQ6AEwBjjIAQ#v=onepage&q=%22Jack%20Gallup%22&f=false

Cycles 1 thru 5 of any PCR, determines what happens next in the rest of the reaction. If the reaction doesn't get up on its feet due to matrix interference or non-linearized plasmid being used (for example) during cycles 1 to 5, you will have a hard time reconciling biological template amplifications (and efficiencies of amplification) with the chosen template used for standard curves if the template used for standards is different in kind and quality from the samples you purport to assess.

The context in which you present the target to the qPCR is highly important in my opinion, and thus, if examining cDNA, use cDNA (e.g. a cDNA sample mixture) for your standards.

If, for some reason your PI or a reviewer does not like that you are using samples to estimate target efficiency, I derived an equation that can theoretically be used to reconcile the two mathematically-disparate universes that result by using cDNA (RT'd from biological samples) in comparison to (i.e.) a purified (and 100%-linearized) plasmid for the same target of interest.

Anyone one of us can derive this equation - given enough time in Excel.

Many scientists have sought this equation:

I have merely expressed it in a form that makes the most sense - and that is to tell the user the Cq where 1 copy appears for a (trustworthy) absolute standard, and further, what that means in terms of what then, the Cq for 1 copy in the biological template would be.

Knowing these two key parameters, unravels the entire equation.

It is just LOG math "slight-of-hand" is all.

E.g., a Cq of 21 at a trusted absolute standard PCR efficiency (Eamp) of 1.7 is the same as a Cq of 16.19 at a trusted sample target PCR efficiency (Eamp) of 1.99.

When expressing one Cq (at a certain Eamp) in the "language" of the another Cq scale (at another Eamp) - the math is obvious. There's no secret.

I've attached the formula here...

I completely agree with Jack, template affects primer efficiency, I have tried it using both plasmid DNA and cDNA and I get differing values. In case you are doing relative quantification with house keeping genes I would stick to making the standard curve with a cDNA template.

Thank both of you for yours answers! I'm sure how to proceede now! :)

What was not mentioned are "intiating effects" of the template. In the first few cycles, where the original template strands are essential, it may be that these original templates are not copied with the efficiency that is obtained in later cycles. High.complexity regions adjacent to the target sequence may cause this, also steric hinerance in supercoiled plasmids may be a potatial reason. Such effects will cause a shift of the ct values of "biological samples" as compared to "PCR product standards", but the amplification efficiency form the first copied molecules will be similar.

Hi, I think the dilutions from the plasmid DNA (isolated though miniprep would be better) would be better option if you have the construct or gene of interest in a plasmid. Otherwise in our lab and most of the neighboring labs in the university, we are using dilutions of cDNA for measuring the efficiency of a primer pair, which is working very nicely.

Wishes

Hi, the best option is the use of DNA as similar as possible to your real samples, if not you'll have bias

maybe this paper could be interesting for you

https://www.researchgate.net/publication/236643667_Error_estimation_in_environmental_DNA_targets_quantification_due_to_PCR_efficiencies_differences_between_real_samples_and_standards

Article Error estimation in environmental DNA targets quantification...

I recommend using a composite sample of all of the samples you want to quantify. This will give you as close to the actual reaction efficiency of each sample and is also a good way to get a large pool of DNA to re-run reactions if you have to optimise your standard curves further.

Also... the following publication

http://aem.asm.org/content/early/2012/04/02/AEM.07878-11

arrives (independently, from another mathematical direction) at the same conclusion as I do (and as does Dr. Codony, also independently, above) as expressed by the formula in my 2011 book chapter. (Attached)

A convergence to the same conclusion by separate investigators seems to be a good thing.

The attachment shows a formula that can be used to reconcile absolute standards with biological templates that each amplify at different efficiencies for the same target...

In my experience, the use of circular plasmid with my specifi target gene, completely failed with any tested condition. When I linearized the plasmid, I started to see amplification curves and I started to work on the optimization of the stanadard used and the best way - in my lab experience - was the amplification of "long amplicons" by PCR, containing the qPCR-target region of the gene. So now I use only long amplicons for my standard curves. There are several advantages: - no need to clone, - possibility to store the small aliquots , - fast preparation of the serial dilution when your aliquots are finished, - good qPCR efficiency without problems related to the supercoiled plasmid...

you only need primers F/R to use in traditional PCR, a kit of PCR product purification and a nanodrop to quantify the dna concentration to estimate the gene copy numbers).

I was looking for an answer to a different question, and saw this discussion. Just wanted to comment on a point which is apparently getting quite frequent and consensus: "use the same kind of input for your standards as you have in your unknown samples". Sentences like "contexts should be the same" etc. are used to support this idea. This often leads to the conclusion "so if my unknown sample is cDNA, I use dilution of cDNAs for making standard samples", which is not a good idea indeed. I have experienced it, and I recently found it published years ago (attached to this message), that using dilutions of cDNA gives you higher-than-real efficiencies, since some components of RT reaction (most likely the RT enzyme) inhibit PCR reaction. This makes the amplification of more diluted cDNAs easier (more efficient), so they come up earlier than expected, and this leads to calculating higher efficiencies, which are sometimes above 100%. So if you want to have cDNA as std sample, to prevent such miscalculation you should spike a negative cDNA (name it A; e.g. from a tissue which does not express your gene of interest) with different amounts of a positive cDNA (name it B); for instance, use samples composed of 100% A, 25%A+75%B, 50%A+50%B, 75%A+25%B, and 100%B as standards; obviously if you want your std curve to fully cover your unknown Cts, B should have expression higher than any of your unknown samples. This is how you "keep context/structure the same", by avoiding "diluting the structure!". The next alternative way can be using dilutions of PCR product of the same primers (better if purified). Since you have to dilute them several orders of magnitude, there is virtually nothing else than the template DNA there in your std reaction (in my hands a 1:5000 dilution of a PCR product gives a Ct of ~10 in qPCR). You can even use serial dilutions of this to spike a negative cDNA sample and use to make stds. The only thing to worry about in this case is the difference in efficiency of primers annealing to that short DNA strand compared with real cDNA strand. The next way could be using plasmids as mentioned above, and the last (and furthest from ideal) will be diluting a cDNA!

The use of amplicons cloned into plasmids is not advisable under any circumstance. A plasmid is grown in bacteria, i.e. undergoes in vivo amplification. The likelihood of contamination is vastly increased, specially since you need to linearise the plasmid prior to incorporating it into a PCR reaction. By the time you have opened your tube, digested, reopened and run gels to check linearisation, you may have plasmid (+amplicon) floating around your lab. If your aim is to detect very low copy numbers, you generate a potential problem for yourself.

On another note, if a perfectly pure plasmid preparation gives you 100% amplification efficiency, this has little bearing on what you might get when you amplify a target from a complex mixture of other DNA molecules. It is particularly irrelevant, if you are carrying out RT-qPCR experiments, where you should use RNA-based standards.

why should the plasmid be linearized before doing qPCR? Maybe for a max.300-bp target piece of sequence it does not matter the rest (which is 3kb or more) at the end is circular or linear. At denaturan step (95C) everything is denatured. We never had contamination with plasmid, as we always check with NTC. Of course making the standards from RNA level is better, though not as easy as making them out of DNA.

The strands of denatured plasmids will not move apart, they are interwoven. Therefore their re-hybridization(naturation) won't follow a first-order but a zero-order kinetic of a very high rate. This can easily interfere (compete) with primer-annealing. Also, stand-displacement by the polymerase in a plasmid is sterically more difficult. This can result in a lower amplification efficiency.

Regarding Stephens comment: contamination is an important issue, no doubts that one should take as much care as possible. However, If the standard preparation is done in a different lab (best in a different building!) and only highly diluted "working aliquots" are taken to the lab where the PCRs are pipetted, a contamination risk is *relaively* low (given recommended standards in clean and careful working).

The standards should not be taken "as is", they should be spiked to "real samples" with no or negligible own amount of the target sequence.