Low level or no level expression???

3 posts / 0 new
Last post
Jason King
Jason King's picture
Low level or no level expression???

Our lab has been doing qPCR using Taqman sets from AB. We use an 18S-VIC set as an internal housekeeping control.
Generally we get 18S signals at about cycle 10-12. If we get our gene of interest (hERG) coming up at cycle 25 or so, we are happy so say that the gene is expressed. But what if it come up at cycle 33, 37, 39?!
I have heard (must still check) that on average a cell has 2500 copies of the 18S RNA (although I must say that this does seem low to me). But if this is true and one assumes a 100% reaction efficiency, then a gene of interest that crossed the threshhold at cycle 21 would only have an average copy number of 1.25. Extrapolating with the same set of assumptions, a positive at cycle 39 would indicate that there was only one copy of the mRNA for the gene of interest in every 256,000 cells!
I realise that the PCR reaction efficiency will not be 100%, but even if it is 70-80%, this would not change my concern about low copy numbers and when exactly one can say that a result is positive or not.
AB state somewhere that their Taqman sets do not detect other mRNAs within the standard 40 cycles.
Any suggestions from people who actually do probe based qPCR would be most welcome. Is the 2500 copies of 18S per cell a myth?

Ivan Delgado
Ivan Delgado's picture
Hi parvoman,

Hi parvoman,
Unfortunately it is close to impossible to find published sources that define what is acceptable and what is not. As far as Ct measurements are concerned, most experts agree that anything beyond cycle 35 is something that should be taken with a grain of salt. Assuming 100% efficiency (doubling of your template at every cycle, or a ten-fold increase every 3.321 cycles), a single DNA molecule will be amplified 1 billion-fold after 30 cycles. In other words, after 35 cycles you've amplified a single DNA molecule by almost 100 billion-fold. If you try to make sense out of signals appearing beyond Ct 35 you are inviting the interpretation of false positives. 
Even this rule has to be taken with a grain of salt. A related rule is that as long as your signal is 10 Ct units below the signal of your NTC, you are fine in using it. For example if you run an assay for 50 cycles (not uncommon), in theory you could use a CT of 40 if your NTC did not show up until after Ct 50.
I have never heard of the 2,500 copies of 18S per cell myth. Trying to normalize the expression of your gene based on how many cells you start with is a very tricky proposition since there are so many steps (RNA extraction, RT, qPCR, data analysis) between your starting material and your final signal. Ribosomal RNAs by nature are found as multi-copy genes and consequently are expressed very highly in cells (the fact that you get Ct signals in the 10-12 range shows this). While it is fine to normalize the expression level of hERG using 18S, I strongly recommend you to consider the following: 
1. When comparing two qPCR assays to each other (hERG and 18S), you should be very careful that the efficiencies of your assays are comparable. If your 18S assay is 98% efficient and your hERG assay was 96% efficient, you will see a greater than 2-fold difference in your readings.
2. When choosing a normalizer (in this case 18S), you want the signal of the normalize to be comparable to that of your GOI. In other words, the Cts should be similar (with the normalizer having slightly smaller Cts to make the ddCt calculation easier). These days most experts agree that you should run multiple normalizers in order to minimize the error inherent in using a single normalizer gene.
Hope this helps.

Jason King
Jason King's picture
We run the 18S-VIC and hERG

We run the 18S-VIC and hERG-FAM sets in the same PCR reaction wells so that there is no variation in the amount of input cDNA. Therefore, it should be possible to get a relative expression level of hERG (dCT). ie. the steps involved in making the RNA etc should not be a problem. So knowing roughly how many copies of the control gene are expressed should allow you to calculate how many copies per cell you have of the gene of interest. At the very least this should act as a sort of common sense check as expressions of significantly less than one copy per cell can probably be classed as no expression.

As you say, it is important that the efficiency of each reaction (18S and hERG) is very similar. Where large dCT exist, differences in the reaction efficiencies between the gene of interest and the internal control would be greatly amplified, so I can see that it would make sense to use an internal control that gives a Ct closer to the gene of interest's Ct.

What do other people think about this?