DEseq on mixed cohort with and without replicates

Question

[5utr]

I have a cohort were multiple tumors, 5 from subtype A and 5 from subtype B, were sequenced. I want to run a DE analysis of subtype A vs subtype B. The non trivial part is that some tumor were sequenced in replicates while others as single samples. What is the best way to merge replicate samples before running DEseq?

[ATpoint]

How to collapseReplicate in DESeq2

[5utr]

Thank you I forgot to mention collapseReplicates but from my understanding it is meant to merge "multiple sequencing runs of the same library" and what I have are different preps from the same tumor biopsy. Would it still apply?

[ATpoint]

Depends on what you want to show. If you want to show differences between tumors then it might make sense to merge them since the difference between preps is probably much smaller than between biological replicates. Is the prep all with the same kit? You can of course ignore some of the samples and only use one sample per tumor. Depends all on the setup.