Error Bars--QPCR

[scienceluvr]

Hi,

I am a student working on a project using quantitative PCR at the Wayne State University School of Medicine. I would greatly appreciate it if anyone could answer a question for me about error bars.

I am trying to add estimated error bars to my graphs. The graphs are fold change in gene expression graphs that were constructed using the efficiency of PCR and delta delta cT values. There are 20 different genes I'm graphing on the X-axis, and the fold changes in expression are on the Y-axis. The gene I used for normalization was ACT, that gave me delta cT values. Then I compared effects of different drugs on the gene expression, for delta delta cT values. Do you know how I can estimate error, without replicate assays? I have done a dilution series along with each PCR run.

Thank you for your time.

[clevermizo]

Not sure, although I don't see the point of error bars without replicate experiments. Why not just perform replicate experiments?

[MrMistery]

there is no way of estimating the error, you have to do the assay multiple times and calculate the standard deviation. It's a way of showing people that it wasn't just that one experiment that went well; doing it multiple times is the whole point

[vastgenome]

Hi,

I think a typical qPCR run would consist of at least 3 technical repeats for each unknown sample. Once you get that, the standard deviation can be calculated from these technical repeats, and can be propagated using geometric averaging rule. the error propagated in the first dCt is carried over when doing ddCt, so they are identical. But at the last step when you are estimating the final 2^(-ddCt), you need to calculate the lower and upper bound of 2^(-ddCt) by incorporating the previously calculated linear errors. As a result, the final error bar will be somewhat asymmetric in the 2^(-ddCt) plot.

[JackBean]

the qPCR run does not have to have three replicates. That depends only on your setting