Can I compare different RNAseq experiment DEG lists to see which experiments are the most similar?

Question

[rachel.walters.18]

Hi all, Is it possible to compare the differentially expressed genes from different RNAseq datasets to see which experimental conditions produce the most significantly similar expression changes in genes? I have 3 individual experiments (2 different conditions in each):

1. M1

1.2. M1 + compound

1. M2

2.2. M2 + compound

1. M3

3.2. M3 + compound

I want to see which experimental conditions produce the most statistically similar DEGs. If anyone could advise that would be great. Thank you!

[rpolicastro]

Before looking at the DEGs, I would suggest starting with a PCA plot and/or the euclidean distance between the samples. This will give you a good idea globally which samples are most similar.

[rachel.walters.18]

Hi and thanks for the reply! Would you recommend performing the PCA on normalised counts? thanks!

[rpolicastro]

Yea, I would run the PCA using the normalized counts. That would reduce the influence of library size and the mean-variance relationship which aren't interesting for this analysis.