We have a group of volunteers attended an intervention study. Samples were taken before and after the intervention and RNA was isolated to analyse change in gene expression, using qPCR. So in general we have two groups Control (before intervention) and Treatment group (after intervention), where every person is a control for itself.
The ΔΔCt values were used in Livak method for calculation of Relative Units/Relative Quantity mRNA expression.
In case when every person is a control for itself. Which is the correct way to calculate p-values? Which values should we use in t-test e.g.? We should use ΔΔCt or RU/RQ values of all patient?
Thank you in advance for your suggestions!
I would always suggest using linearised data (RQ etc), because this makes fewer implicit assumptions about your data (and because logarithms make everything just that little bit more terrifying).
If your qPCR is of good efficiency and your reference genes are well-validated, then your linearised RQ values should serve as a perfectly viable dataset.
You have paired data, so you should check your datapoints to see if the distribution is normal. If normally distributed, then paired t-test is fine.
If not normally distributed, then a Wilcoxon matched pairs signed rank test is more appropriate.
Hi Oskan,
2^^DDCT method needs efficiency of the PCR is close to 1 and the PCR efficiency of the target gene is similar to the internal control gene. If your qPCR data suitable for that, you can use 2^^ΔΔCt values of all patient.
Different examples in the article can help.
'Analyzing real-time PCR data by the comparative C(T) method''
Article Schmittgen TD, Livak KJAnalyzing real-time PCR data by the c...
You can do a paired-sample t-test using the ΔCt values (sample 1: before, sample 2: after). This is equivalent to a one-sample t-test using the ΔΔCt values (per individual, one ΔΔCt is calculated as the difference between the two ΔCt values [before and after]). The paired-sample test is simpler to do with software, as all you need are the ΔCt values [before and after]. The one-sample test is simpler when you want to do the calculations by hand.
At no time you need to/should transform the ΔCt values (like raise to the power of 2 or something). The whole test is just about comparing two samples of ΔCt values to check if the mean difference is discernible from zero. Since your ΔCt values are paired (before and after within each individual), you use a paired-sample test or a one-sample test on the individual differences (these individual differences would be called ΔΔCt values).
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis