I am analyzing a data generated from an RT-qPCR experiment where different treatments, each with 3 different concentrations are applied (See attached snapshot). The CT value for around 50 microRNAs is measured. The aim is to find differentially expressed microRNAs in different concentrations of treatments.
For each concentration of the treatment there is 3 replicates and 1 control (See attached snapshot).
I applied the deltaCT method and generated the the fold changes. Are the fold changes, by filtering for those below 0.5 and 1.5 ,enough in this case to find the differentially expressed microRNAs, or do i also need to calculate the pValue per each microRNA, or per each concentration group vs control ? If so how exactly can i calculate and adjust the pvalues in this case, with only one control replicate and 3 treated replicates per concentration of the treatment?
Thank you
From what I understand from your question: You would want to use p-values.
A one-way ANOVA will determine the differences between means when you have one factor. ie. Treatment: treatment 1 vs treatment 2 vs treatment 3 (at a single concentration).
A two-way ANOVA will determine the differences between means when you have two factors. ie Treatment & concentration:
T1 C1
T1 C2
T1 C3
T2 C1
T2 C2
T3 C1
T3 C2
T3 C3
etc.
You will need one data value per exposure. Preferably fold change. However, if required you can use delta Ct.
So if you have one control value you would compare that to one value for each exposure. If you have multiple values for each exposure there are technical replicates. You would take the mean value of those three to represent that exposure.
You will need multiple n values, up to 6. This means you need to repeat that experiment 6 times using new flasks of cells and exposure preparations. Otherwise you cannot correctly apply statistics to your data.
When you do this you will have repeated observations for the same outcomes. This means you will use a repeated measures one, or two, way ANOVA. ANOVAs can be done in most statistical software above Excel. It can be done to a minimal degree using prism.
If you have dCt values from 50 miRNAs, I would suggest using limma (https://bioconductor.org/packages/release/bioc/html/limma.html). The advantage is that limma uses a moderated t-statistic for the tests, what increases power and robustness (by "borrowing" information about the variance from all genes for each test of one of the genes).
Arnaud Beurdeley Yes you are right concerning the experiment i will try the suggested thresholds
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Nucleic Acids
- RNA
- Small RNA
- microRNA