Hi,
Briefly, I want to identify differentially expressed genes from RNA-seq data using DESeq2.
I have three conditions and I want to compare the three conditions to each other.
Usually, I perform pairwise comparison building the dds file with the two conditions that I want to compare and subsequently I run DESeq2 on the dds object. I would like to ask if there are any differences between two approaches:
- Approach 1: perform separately each pairwise comparison as I described above
- Approach 2: perform the analysis including the three conditions in the dds object and then run the contrast function to extract the comparison that I want to analyse.
In particular, are there any differences between the two approaches in the normalization of count data and in the differential expression analysis? For example, if I have a condition with more variability than the other two conditions, this can affect the identification of differentially expressed genes in Approach 2? In addition, I was wondering if the Approach 1 is able to highlight more differences, even the smallest one between the conditions, in comparison with the results obtained from the Approach 2.
Thank you!
Concetta
Hi, I'm also having this question because I get two very different adjusted p-value with pairwise comparison and contrast function (for the same dataset). Could you help me understand what's the better approach?
See the FAQ in the vignette.