Hello, I need help in analyzing the significance of results obtained from RT-PCR
I have control group and three treatment groups and in each group I check the change in expression of 8 genes. I have made all calculations till the 2^-deltdelt CT level and I know that results more than one means overexpression and less than one means downregulation but I want to test the significance of resultsin the attached file.
1-I want to now what is the most suitable statistical test to test the significance
2- what are the best values that should be added in the statistical test whether using delta delta ct or 2^-delta delta CT ( fold change)
1 - essentially t-tests on the delta-ct values. The t-tests can address coefficients in any kind of linear model. Note that first-order coefficients are actually delta-delta-ct values.
2 - I don't really understand that question, but I think it is related to your first question, so my answer is: the delta-ct values are used for the tests.
many thanks Dr Jochen for your reply.
so I can understand from your reply that we use t-test as statistical method and I do the calculations using delta CT and not the fold change (2^-2deltadeltaCT), is that right?
Yes, that's right.
The simplest case is a straight-forward t-test. You have several delta-ct values in the control group and also several delta-ct values in the treated group. The t-test using these two groups of values tests the hypothesis H0:"delta-delta-ct = 0" (that is: the gene is not differentially expressed between the two groups).
The t-test is just an aspect of a a simple linear model Y = b0 + b1*X, with X being 0/1 coded (0 for "control" and 1 for "treated"). When Y is the vector of delta-ct values, the coefficient b0 estimates the mean delta-ct in the control group and b1 estimates the mean delta-delta ct (treated vs. control). The t-test is actually a test on b1=0.
You can use an "ANOVA model" (Y = b0 + b1*X1 + b2*X2 + ... , where X1=1 for the values belonging to the first treatment group,X2=1 for the values belonging to the second treatment group etc.). The coefficients b1, b2, ... estimate the delta-delta-ct values of all the treatment groups against the control, and they can be tested with t-tests (the model fit provieds the estimates together with their standard errors).
If your treatments differ quantitatively, like in the concentration of an applied drug or a in the time, you may consider a "regression model", which is also a linear model Y = b0 + b1*X, but where X codes the quantity of the treatment. The coefficient b1 will then still be an estimate of the delta-delta-ct value, and here this indicates the change in expression depending on the change of the treatment quantity (concentration or time or whatever).
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Cell Biology
- Culture Cells
- Cell Line