# design of standard-curves with low variance at very low copy numbers - (Feb/27/2006 )

Hi,
We need to quantify very low copy numbers. The range is from about 1 to 100. For quantifying we use plasmid-DNA that contains our gene of interest. In our tests we got a standard curve. But at low copy numbers, the standard deviation becomes very high. We ran five replika for 10^1 copies and the standard deviation is 1,5 . So a quantification result will also be wrong by up to 80%.
Does everybody else also see similar problems?
How can I quantify a sample with a lower variance? Does it help to "improve" the efficiency of the PCR despite the efficiency is also at E=(1+0,9), so that in the initial cycles virually every template becomes copied and the cp-values are more reproducible? And how could I do this?
The annealing temperature and the primer-concentrations are already optimied. Would it help to increase annealing and elongation time? Annealing is at the moment at 10sec and the elongation time is also 10sec.

Thanks

Max

-westenmax-

We ran five replika for 10^1 copies and the standard deviation is 1,5 .

I assume, that your standard deviation is referring to your CT values; so this is actually a very high deviation. Are you able to measure your conc. photometrically? Is there a difference, so your realt-time results are due to incorrect pipetting.
I am also wondering that you have got 5 replicates; 3 of them should actually have the same CT, you can neglect the other 2 and calculate your stanard curve with those good 3 replicates.

The annealing temperature and the primer-concentrations are already optimied. Would it help to increase annealing and elongation time? Annealing is at the moment at 10sec and the elongation time is also 10sec.

Annealing and elongation times are very short. Usually 30 s to 60 s are used for real time pcr.

Silvia

-Sil-

QUOTE (westenmax @ Feb 27 2006, 03:30 AM)
Hi,
We need to quantify very low copy numbers. The range is from about 1 to 100. For quantifying we use plasmid-DNA that contains our gene of interest. In our tests we got a standard curve. But at low copy numbers, the standard deviation becomes very high. We ran five replika for 10^1 copies and the standard deviation is 1,5 . So a quantification result will also be wrong by up to 80%.
Does everybody else also see similar problems?
How can I quantify a sample with a lower variance? Does it help to "improve" the efficiency of the PCR despite the efficiency is also at E=(1+0,9), so that in the initial cycles virually every template becomes copied and the cp-values are more reproducible? And how could I do this?
The annealing temperature and the primer-concentrations are already optimied. Would it help to increase annealing and elongation time? Annealing is at the moment at 10sec and the elongation time is also 10sec.

Thanks

Max

Hi westenmax,

This is a common problem with low copy number standards which we also see here (using plasmid calibrators). Rather than PCR efficiency issues, it seems to caused more by the tendency of nucleic acids to stick to the inside of plastic tubes... This isn't noticable when the nucleic acid concentration is high, but causes problems at very low copy numbers. The solution (for us anyway) is to 'block' the tube surface using yeast tRNA (Sigma R5636) at 100ng/ul in the calibrator diluent.

Also, the effects of DNA degradation caused by poor pH control and freeze-thaw cycles have a more significant effect on low copy number calibrators, so make your calibrators in buffer (I use 0.2x TE - don't use 1x TE as EDTA will inhibit polymerases) and make single-use aliquots of each dilution.

good luck!

Del.

-del-

QUOTE (del @ Feb 28 2006, 12:00 PM)
QUOTE (westenmax @ Feb 27 2006, 03:30 AM)

Hi,
We need to quantify very low copy numbers. The range is from about 1 to 100. For quantifying we use plasmid-DNA that contains our gene of interest. In our tests we got a standard curve. But at low copy numbers, the standard deviation becomes very high. We ran five replika for 10^1 copies and the standard deviation is 1,5 . So a quantification result will also be wrong by up to 80%.
Does everybody else also see similar problems?
How can I quantify a sample with a lower variance? Does it help to "improve" the efficiency of the PCR despite the efficiency is also at E=(1+0,9), so that in the initial cycles virually every template becomes copied and the cp-values are more reproducible? And how could I do this?
The annealing temperature and the primer-concentrations are already optimied. Would it help to increase annealing and elongation time? Annealing is at the moment at 10sec and the elongation time is also 10sec.

Thanks

Max

Hi westenmax,

This is a common problem with low copy number standards which we also see here (using plasmid calibrators). Rather than PCR efficiency issues, it seems to caused more by the tendency of nucleic acids to stick to the inside of plastic tubes... This isn't noticable when the nucleic acid concentration is high, but causes problems at very low copy numbers. The solution (for us anyway) is to 'block' the tube surface using yeast tRNA (Sigma R5636) at 100ng/ul in the calibrator diluent.

Also, the effects of DNA degradation caused by poor pH control and freeze-thaw cycles have a more significant effect on low copy number calibrators, so make your calibrators in buffer (I use 0.2x TE - don't use 1x TE as EDTA will inhibit polymerases) and make single-use aliquots of each dilution.

good luck!

Del.

Hi. Max,

I agree with Del's comment on the dilution and sticking issue and the use of tRNA and 0.2x TE. However, in our experience plasmid DNA standards almost always show this problem at lower dilution. The way we solve this is by using linearized cDNA standards which are only slightly larger than the amplicons. These validated primers and standards are available at www.aanotech.com.

Regards,

Raja

-aanotechqpcr-