Context: I am currently analyzing some Illumina RNA-seq data. These data are from treating two different genotypes (A & B) with four different treatments. There are three biological replicates for each treatment group.
- A treated with negative control
- A treated with a solution of some concentration
- A treated with the same solution but more concentrated
- A treated with positive control
Same for B.
What I have done: I used Salmon to quantify the reads and also create 100 bootstrapped samples, and then I used fishpond to perform differential gene analysis, where I was able to compare within a treatment group. For example, I was able to find differential expressed genes between A and B in the first treatment group, using A as a reference.
Question: The analysis I have done so far does not allow me to compare across treatment groups because the estimates of abundance are normalized within a treatment group. In other words, I cannot compare the log2(foldchange) from Treatment 1 to that from Treatment 2. However, I want to be able to compare across treatments. I want to be able to establish a basal level to which I can compare different treatments. But I do not know how to do so.
Attempt: I am thinking about putting all the A replicates from Treatment 1 and Treatment 2 into Group I, and all the B replicates from Treatment 1 and Treatment 2 into Group II, and use fishpond to compare Group I and Group II.
Is this the right way to go about doing this? Also, the mapping rates from salmon for all samples range from 68 to 81 with the majority being in the 70s. Is this typical? Any advice is appreciated! Thank you :)
https://support.bioconductor.org/p/9142044/