I have 3 disease groups that I would like to compare. Each disease group has 3 samples. Illumina microarray has been performed on each sample along with matched controls.
I have generated 9 DEG files for each sample (Affected 1A vs Control, Affected 1B vs Control, Affected 1C vs Control etc.). Each DEG file contains 23,835 genes with their respective logFC and Adj.P.Values. I would now like to combine the samples into their disease groups and then filter for <0.05 Adj.P.Val and +/- 2 LogFC to create disease group DEG profiles.
To do this, can I take an average of the 3 samples (LogFC and Adj.P.Vals) per disease group? I have heard a few people say that taking an average is the 'common' thing, but I cannot find it in the literature anywhere. I also am not sure how reliable taking an average of the Adj.P.Vals is.
Thanks in advance!
If I understand correctly, you have three biological replicates per treatment. Why are you performing the comparisons sample per sample, instead of performing group-wise comparisons? The limma Users Guide has sections on how to perform the analysis for comparisons between multiple groups.
Yes, you are correct! I completely forgot about performing group-wise comparisons within Limma, thanks for bringing that up. I will do this instead.