Dear Vivian,

This is a known problem due to execs of DNA and it cannot be only related to RT mix. The reason is that even if you purify the DNA from perfect sample and quantify it to serve as standard you would observe the same phenomena if initial specific DNA in the first standard higher is than 5 pg. The most probable reason is that in highly concentrated DNA samples, the primers can non-specifically bind to DNA (if primer is has than more than 5 C’s or Gs or 5 AT’s in row) and from that moment lets says cycle 2 or 3 the wrong template would be amplified and get overhand. Also the unused primer in RT mix can give the same problem but in that case you would observe that amplification curve in these samples are different. In that case the linear part of amplification curve is less than 5 cycles instead of 7-9 cycles for high efficient PCR amplification.

So my advice should be do not worry and omit the first dilution for the standards. For the samples is the best initial test would be that you dilute the cDNA randomly from 3 samples (3X, 5x, 10X and 20x) and perform PCR with two or three housekeeping genes. Then you can choose the best dilution of cDNA.

Good luck

This is definitely an unusual situation. It is peculiar that the odd peak has an even lower melting temperature than the NTC primer dimer--that would seem to suggest that it must be very small---a gel analysis could confirm this. One possibility is that (at least) one of the primers has some homology to the internal portion of the amplicon---and when there is a greater concentration of the complete amplicon at early stages of the PCR--as in the undiluted sample--there is a greater probability of that primer initiating at the site of partial homology and creating a "deleted" amplicon. Another possibility is that the primers have low affinity for another gene sequence and this is only detectable when the cDNA concentration is high. If your Ct values are sufficiently low, you could just avoid this situation by diluting all your cDNAs. Primer re-design might be the best option though, it this is an assay you plan to use often.

Thanks very much for your reply.

The problem is that I have this peak in almost all of my assays of nine different genes. Therefore, I mentioned the possibility of having primer dimers from the RT reaction. Maybe I could try to take only Oligo-dT or only random or gene-specific primers in the RT reaction.

Anyway, diluting all samples could also be an option, thank you.

Ok, I did not understand that you were seeing this in assays with multiple different primer pairs. And the melt temp of the low dilution "extra" peak is always the same or does it vary with different primer pairs?

It is always 75°C.

Hi Vivian, I suggest you before performing the qRT-PCR assays, you could run Conventional PCR for amplification your gene of interests on the same volume of cDNA (for example 1 micro-L). This step will help you to see 1) if there is any contamination in your cDNA in case you see more than one in the gel. 2) The concentration of your gene of interests in cDNA , the intensity of PCR bands could give you an idea about how much of the gene of interests in the cDNA. In general, If you have more than one bands are being seen in the PCR product, that means the probability of contamination or your primers are not more specific and could match with other homologous genes. If you have only one band, that's mean that you may exclude the undiluted cDNA.

Are you quantifying your gene of interests from DNA template or library cDNA?

Could that be some product on genomic DNA? Are you DNasing your sample? Maybe there is low level DNA contamination, which you dilute out and then it is not detectable anymore. I would check (BLAST) what kind of product would your primers give on genomic DNA, maybe that's it?

You can differentiate the primer dimer from the non-specific products by simply running the agarose gel of your qPCR final products. Usually, use only 0.5ul or most 1ul of cDNA for PCR(the final dilution would be 25X if your PCR volume is 25ul). If you use more than 3ul of RT rxns for PCR, usually use the diluted RT rxn (at least 5X dilutions). If your primer pair does not span over 1kb intron or introns, you do need to remove your gDNA from your RNA samples before your RT rxns. You have multiple dilution of samples for your standard curve, you need to choose the best dilution range to make your standard curve in term of specificity and sensitivity.

Hi Vivian,

it is a very interesting situation. I have some questions for you.

First, could you describe how you do the reverse transcription from mRNA to cDNA? It is possible as many said above that you amplify some gDNA trash. Second, make sure that you use clean (non-contaminated) reagents (based on the chemistry you are using, that could be SybrGreen, dH2O or Master Mix). That is because you mentioned you have primer-dimers at NTCs at 79oC. Well, this is quite unusual because the melting temperature of a primer-dimer would be 65-75oC and it would appear later than the specific products because of reaction exhaustion. What you found likely suggests that the product is specific. So, you may also check the Ct value of this amplification. If it is relatively low, then I am not sure whether it is primer-dimers or another product your primers amplify. Third, the fact that the 75oC product is lost when you dilute cDNA is good because it shows that it is not in abundance in your samples. Besides, you will have to use diluted cDNA, anyways. May I ask again what its Ct is?

However, to keep things straight, I would suggest you also run a normal PCR, see the product size and conclude whether the 75oC product is a primer dimer (should be small) or specific. Make sure your reactions are clean (use new, clean reagents and check your NTCs). What is always useful is to run a non-reverse transcriptase control (omitting the enzyme in the RT after you eliminate gDNA). This will tell you if you have gDNA in your qPCR and therefore if you amplify gDNA trash.

If everything proves pessimistic (e.g. big 75oC product, low Ct, acquisition of the 79oC product with low Ct and big size, positive BLAST of your primer pairs with DNA databases), then the best thing to do is to re-design them.

Dear Vivian,

This is a known problem due to execs of DNA and it cannot be only related to RT mix. The reason is that even if you purify the DNA from perfect sample and quantify it to serve as standard you would observe the same phenomena if initial specific DNA in the first standard higher is than 5 pg. The most probable reason is that in highly concentrated DNA samples, the primers can non-specifically bind to DNA (if primer is has than more than 5 C’s or Gs or 5 AT’s in row) and from that moment lets says cycle 2 or 3 the wrong template would be amplified and get overhand. Also the unused primer in RT mix can give the same problem but in that case you would observe that amplification curve in these samples are different. In that case the linear part of amplification curve is less than 5 cycles instead of 7-9 cycles for high efficient PCR amplification.

So my advice should be do not worry and omit the first dilution for the standards. For the samples is the best initial test would be that you dilute the cDNA randomly from 3 samples (3X, 5x, 10X and 20x) and perform PCR with two or three housekeeping genes. Then you can choose the best dilution of cDNA.

Good luck

Thanks for your replies. I always waited for that point when someone would say: "I know this problem and solved it already" - special thanks to Behrouz :)

In conclusion, I would say that diluting the samples would be the best thing to do. However, in my case I have the problem of relatively late Cts for undiluted samples (between 28 and 35) which is true for most of my assays.

Today I tested my first assumption of RT-primer dimers (see pictures below). Interestingly, the additional peak does only occur in the reactions with Oligo-dT / random mix and with Oligo-dT only. The reactions with random only and gene-specific primers do not have it. Furthermore, gene-specific primers give me a much earlier Ct (ca. 20) than the others (28-30). Therefore, I am thinking about using gene-specific primers for my RT reaction and also dilute the cDNA samples to be sure to get rid of this phenomenon and at the same time having early Cts.

Non-specific amplification due to DNA contamination of your RNA is unlikely to be the cause of the 75C peak because a) you see the same 75 melt peak with several different primer pairs and b) the usual PCR product from gDNA contamination would be expected to have either the same or a higher melting temperature than the cDNA product. So the thought that there is some carryover from one of the RT reaction components, which is common to all of the PCR reactions--which include different PCR primers-- makes the most sense. Still, the suggestion that you try a reaction that has everything but the reverse transcriptase is good, if that is possible with the reagent kit that you are using-it will tell you whether there is a (separate) DNA contamination issue. If you can't do that, then you can also perform a q-PCR reaction where you just add the equivalent amount of your RNA sample directly as template into the q-PCR reaction. What is surprising about your new result is the huge difference in Ct between your gene-specific primed and your random or oligodT primed RTs. What is the temperature profile of your RT reaction? It it is too high (42-45 say) and does not include an initial low (20 or lower) step , then oligodT and especially random n-mers will not bind efficiently--but your specific primer--presumably your "reverse" PCR primer (?)--, has a much higher Ta and will bind efficiently. Obviously, if you can work out the conditions where you can use oligodT or random n-mers for your RT, it will save you a huge amount of work-- performing gene-specific RTs for every future qPCR you have to do! A reliable kit like Multiscribe:

http://products.invitrogen.com/ivgn/product/4368814 (without RNAse inhibitor)

http://products.invitrogen.com/ivgn/product/4374967 (with RNAse inhibitor)

or qScript

http://www.quantabio.com/product.php?base_id=95048

could be worth the investment.

By the way, although it is unusual, I have seen cases where the "NTC" ("primer-dimer") product has a higher melting temperature than the "correct" cDNA product. It seems possible to occur in cases where the primers have very high Ta values (are GC rich or very long).

Hi Vivian,

I have a suggestion ...try it when you do a reverse transcription next time.....it may work..

Check your oligodT conc. that you are using for the RT reaction......reduce it a little bit and you may get rid of the peak seen specifically in the oligodT reactions..

Thanks to all the contributers!

I am having a no RT and also an NTC control in all of my runs. The no RT control is negative, shows no peak, like it should be. The DNA contamination can be ruled out I think.

RT reaction is done with Agilent chemicals and 170ng oligo-dT and 30ng random primers in a 20µl reaction with the protocol:

5 minutes at room temperature

15 minutes at 42°C

5 minutes at 95°C

Of course it would be best to avoid the gene-specific RT reaction but what can I do else (except from trying new chemicals from a different company) to improve my protocol?

By the way: I have already sequenced some of my qPCR products and found that I have the right product and I also changed all chemicals I used before I wrote my question in this forum :)

I would just buy the small size Multiscribe kit (

I really appreciate your advice but I have decided not to buy new chemicals. It took me a long time to optimize my assays with the chemiclas I use now and if I would buy new chemicals, I would start all over again.

Today, I tried the RT reaction with a 25°C step for 10 minutes instead of 5 (random primers) but it gave me absolutely the same Ct as the 5 minutes reaction.

Therefore, I will go for the gene-specific RT strategy and additionally dilute my samples.

Thanks a lot

:)