I have done my qPCR experiments and gave me some results, I used the DDCt method and I calculated the 2^(-DDCt), I transformed my data in base 10 logarithm and separated my samples between control and patients. I want to ask if I see that there is for example a fold change 4 times higher in patients for my gene of interest then I use one-tail or two-tail t-test, and what if the distribution is not normal, will I do non-parametric test, or I can skip the outliers and do the t-test. I am very confused in that statistical conundrum.
Hi friends,
If you have a non normal distribution, this can be avoided by using log10 values, so you can go ahead. But sometimes even using these values will not be helpful so you have to draw your histogram.
Analyze the dCt values. If the dCt values are paired, use a paired t-test, otherwise use a two-sample t-test.
No the dCt values are not paired and after log10 transformation the distribution remains non normal.
Dear Panagiotis N. Stoikos
If your data is not normally distributed, you should use non-parametric statistical tests such as Wilcoxon rank sum tests or Mann-Whitney U tests in order to compare the expression levels between the two groups.
Regarding the one tailed or two tailed. one tailed can specify the direction of the effect (positive or negative) but the two tailed one can be used for both direction at the same time.
Best...
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Peptides