Hi Maha,

How many cycles did you perform ? One possibility among others, there is a saturation of your revelation system. A factor 10 of dilution is only about 3 cycles. Dilute your PCR products (2-4-8) and dose again.  The threshold of your detection at 2% is reached, around 10e-7 and 10e-8.

There are multiple factors that affect PCR efficiency:

1. Amounts of PCR inhibitors in the sample like SDS, excessive proteins, hemoglobin, phenol/ethanol, etc.

2. PCR primer and/or probe design 

3. Inaccurate sample or reagent pipetting esp if doing serial dilutions

4. Standard curve problems: like Ct outliers or sometimes you have to drop points for example if the most concentrated sample of the standard curve occurs at a later CT value than expected, but the rest of the curve looks normal, it is likely that this dilution point is showing inhibition and should be dropped

Are you diluting a parasite culture or blood? Then be cautious that different RBC's have varying amount of parasite genome and this also depends on the stage of growth. Assuming that your PCR conditions are perfect, this can cause variations in Ct.

Please post a plot of the dilution curve.

What detection format (signal generation) have you used? SYBR Green?

Have you analyzed the PCR products (melting curve and/or PAGE) ?

As I understand your most diluted samples have the same (high?) ct-value.

There are only three possible reasons that I can think of:

- contaminations

- amplification and detection of unspecific products

- Poisson variation*

*if the dilutions are so high than only a few molecules will get into a PCR, the distribution of molecules will follow a Poisson distribution. Example: If the dilution is so that -on average!- only 3 molecules will be in the reaction, particular PCRs may contain 0 molecules with 5% probability, 1 molecule with 15% probability, 2 with 22%, 3 with 22%, 4 with 17%, 5 with 10% and so on. The higher dilution will give an average of 0.3 molecules per PCR, so there is a 74% probability that the PCR contains 0 mlecules, 22% for 1, 3% for 2, and 0.36% for more than 2 molecules. So there is a resonable chance to get the same amount of molecules into the PCRs for both dilutions.

Yes I agree with Jochen, I had the same problem a year ago and it was the Poisson variation. I just prepared dilution series pipetting much greater volumes and put more template in the starting dilution sample (the most concenrated one) which the dilution series is based on.

I think you have to purify the parasite's DNA at the highest point of curve, do it by dúplex or triplex,and measure the loud you amplify, do the dilutions pertinents and make the curve amplifyng by triplicate.So you will see when the excess of DNA is inhibiting and the LOD  of your method.I believe it is a cleanest way to make a curve.

I agree with Jochen. Most likely Poisson variation, which he explains very well. 

Joshen and Andrew, is it possible that Poisson variation happens at dilutions 7&8 and fades out in 9 and 10?

Yes. At dilutions 7-8 you can have some "effective" Poisson-varying concentration, at higher dilutions the expected conc. it too much below 1 molecule per PCR, so that essentially all the PCRs are actually negative. And if you measure unspecific amplification, it can be that the Ct values at these hiher dilutions are all the same, because the unspecific amplification is all the same.

This is all speculation. Just a possible explanation. You need to carefully check if this can apply (or actually does apply) in your specific case.

The Poisson effect only occurs in the last two dilutions that give a specific signal.

If 9 and 10 are amplifying as well, check the melting temperatures - they may be giving nonspecific products.

Also, dilution 10 is a very high dilution; most PCR's cut out before this. If it's giving a signal, the levels of parasitemia will have to be high enough to give you 10 billion parasites per aliquot in your undiluted DNA, which seems an awful lot to me. 

I see Andrew and I still fully agree :)

I too agree with Sir Andrew.

I think you are not done real time PCR, so your visualizing the band in agarose  gel. In this band one can not accurately evaluate the PCR efficacy. Minimum copy number of template enough to give visible band. But we can realize the reducing the level of fluorescent bands according to dilution. If it is not reducing, all the dilution producing same level of fluorescent band, that may be do to contamination.

I encounter the same question these days, maybe Jochen's answer will be helpful

.