Hi there!

I am performing differential expression analysis on embrionic cells. According to PCA (see below), I have found that samples are clustering according to Experiment day and not by Condition

Then, I have performed in parallel the analysis

• taking all samples
• taking all samples and corrected by the batch effect
• taking samples processed only at day 27th
• taking samples processed only at day 21th

When I have compared the obtained results, I have found that the 4 analyses have a low number of DE genes(see venn diagrams below).

I was wondering how the results of all analyses show low concordance and which one of these analyses reflects more the biological truth.

Best regards,

Giulia

The recommendation would be design=~batch + condition. If there are low DE number that is nevertheless the correct approach.