edgeR TMM normalization with different cell lines, same treatment, and no replicates

Question

[RZ]

Hi All,

I have an RNAseq dataset with 5 cell lines, untreated and treated with a drug (a total of 10 samples). We wanted to see if the treatment could produce the same effect in different cell lines, but we have no replicates for each sample. I understand this is definitely not ideal, but I am wondering whether it would still be reasonable to do TMM normalization on this dataset (How do you generate TMM normalized counts using EdgeR?), categorizing the samples into two groups, 'Treated' and 'Untreated', without specifying the cell lines. Thank you!

[ATpoint]

Sure, you can do that. TMM does not care about replicates and design.

[RZ]

Thank you so much!

[RZ]

I should add that eventually, I want to do GSEA on this dataset, which requires the normalization step.