Representing RT-PCR Data - (Dec/28/2006 )
I am depiciting Real TIme PCR data as fold changes. The fold changes are calculated using the delta-delta Ct method. The controls are all 1 and the experimental values occupy different numbers. Can anyone suggest what values to use for error bars? Any suggestion is appreciated.
It is also a subject of great concern for me. I have been e-mailing with some staticians and they recommend NOT to add error bars unless you used those values for the statistic assay. As the representation with the delta-delta Ct (DDCt) method implies an exponential transformation (2-DDCt), at the end error bars originated at the DDCt step would be assimetrical.
Thus up to now I use the DCt (not DDCt) values for the statistical comparison and the 2-DDCt for the representation. However I have just discovered this forum and maybe I will find a better solution
I would strongly suggest to do the statistics on the dCT values and maybe even graph them. If you work on ddCT values, you are much more likely to produce false positive findings, as you virtually reduce the variance of your controls to 0! So allmost every deviation in the treated samples will result in a statistical significant result. But that is not sound. Graphing the deltaCT values with error bars provides even more information, than any ddCT graph will. So if you trust your experiments, share the information with the audience