Hi qPCR geeks :-)
Since being unexperienced researcher (generally) and especially in the field of the analysis of my qPCR data I have run into problem for which I could not find a definitive answer on the internet.
Design of my experiment and samples from which I obtained my relative quantification values were as follows: 3X wild type control, 3X deletion mutant strain. These were bacteria grown in separate flasks – so in total 6 flasks, in other words I believe independent (non-paired) samples. I did qPCR of suspected genes influenced by deletion. Next I calculated relative expressions (including efficiencies even though they were quite good). For example for gene X I got three values: WT1 vs Δ1 = 1.6, WT2 vs Δ2 = 1.8, WT3 vs Δ3 = 2.6.
Now finally coming to my first question concerning student's t-test: If I have relative expression then it means that the value for wild type (WT) is always 1? Meaning my input for student's t-test will be 1,1,1 vs 1.6, 1.8, 2.6. If yes, is it not a problem when thinking about t-test as a tool how to disprove null hypothesis that is based on standard deviations (for WT samples it is all of the sudden zero).
And my second and more important question is: Shouldn’t I compare relative expression “everything with everything” instead 1vs1 2vs2 and 3vs3? The second approach would yield for relative expression 9 values: WT1 vs Δ1, WT1 vs Δ2, WT1 vs Δ3, WT2 vs Δ1, WT2 vs Δ2, WT2 vs Δ3, WT3 vs Δ1, WT3 vs Δ2 and WT3 vs Δ3. Because at the end of the day I chose consciously what is WT1 and what is WT2 and so on and I believe nature does not care what I wrote on those flasks.
Thank you very much in advance for your awesome answers!
Jan
Hi Jan Čapek ,
You can try this:
1) Take the mean of all 2-dCq of wild type and mutant - STEP 1 NORMALIZING
Control
2-Cq interest gene control 1 / 2-Cq reference gene control 1 = 2-dCq control 1
2-Cq interest gene control 2 / 2-Cq reference gene control 2 = 2-dCq control 2
2-Cq interest gene control 3 / 2-Cq reference gene control 3 = 2-dCq sample 3
Mutant
2-Cq interest gene mutant 1 / 2-Cq reference gene mutant 1 = 2-dCq mutant 1
2-Cq interest gene mutant2 / 2-Cq reference gene mutant2 = 2-dCq mutant 2
2-Cq interest gene mutant 3 / 2-Cq reference gene mutant 3 = 2-dCq mutant 3
2) Take the mean of all control 2-dCq. In this case: (2-dCq control 1 + 2-dCq control 2 + 2-dCq sample 3)/3 = 2-dCq mean control. - STEP 2 CALIBRATING BY CONTROL MEAN
Divide all samples 2-dCq (control and mutant strain) by the mean 2-dCq mean control of control group.
2-dCq control 1/2-dCq mean control = A
2-dCq control 2/2-dCq mean control = B
2-dCq control 3/2-dCq mean control = C
and the same for mutants 2-dCq mutant 1/2-dCq mean control = D... E... F
3) If you did right, your new control mean will be 1. Then, run a t test for control versus mutant. You will have 3 samples control (A, B, C) and 3 mutants (D, E, F).- STEP 3 STATISTICS.
Best
Jacó
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA