Protocol Online logo
Top : New Forum Archives (2009-): : PCR, RT-PCR and Real-Time PCR

About the qPCRs and statistics - (Nov/07/2012 )

Hello.
I make some efforts to describe change in expression of few genes in a time-course study. My experiment comprises one control and few experimental groups.
So far I`ve normalized the expression of each target gene for each single sample vs the ref. gene in all groups of the study. I used the dCt method and I collected groups of dCt data.
Then I determined the ratio of expression in the control group and in the test groups.
What I am supposed to do now?
Would it be enough if I plot the means of dCt of each group (control and test groups) with CI, SD or SEM bars and compute the P values between the groups, i.e. one-way ANOVA comparing more than two means with the appropriate multiple comparisons test?
I am reading a paper now: Statistical analysis of real-time PCR data. Yuan JS et al., BMS Bioinformatics 2006, 7:85.
I thing they recommend some SAS program and they discuss the use of the Wilcoxon test, which is for two groups of data.
I have some experience with the GraphPad Prism.
Isn`t it good enough for this job?

-Nephrite-

Graphpad should be able to do it. They use the Wilcoxon as it is for non-parametric data (samples where a normal distribution cannot be assumed, and/or low (<30) samples per group), it is more or less equivalent to the Student's t-test. A similar test is the Mann-Whitney, which is another variant of the t-test. You probably need to use the equivalent of an ANOVA, which is the Kruskal-Wallis test (one way ANOVA equivalent).

Plotting the normalized means relative to the reference gene should work.

-bob1-

Thank you!

-Nephrite-