Single cell RNA-seq DEG analysis on multi-tissue

Question

[joonhong kwon]

Hi all,

I am analyzing the scRNA-seq data on immune cells in brain and PBMC of mouse. I performed clustering analysis by integrating brain and PBMC data. And I try to perform DEG analysis between control and case for each tissue and cluster (ex. Brain_Case_B-cell vs Brain_Control_B-cell or Blood_Case_T-cell vs Blood_Control_T-cell etc.). DEG analysis between Brain and Blood will not be performed. (The reason the two tissues are integrated and analyzed is to identify tissue-specific clusters.)

Is it OK for DEG analysis to be performed on the integrated data? Or is it better to perform DEG analysis after clustering again for each tissue?

Thanks in advance!

[ATpoint]

That all depends on the context and the question you want to answer. Technically you can run any DEG as long as the compared cells are within the same batch and you do not use the integrated values but raw (or subsequently normalized) counts, as the integrated values are not suitable for any differential analysis. Does that make sense?

[joonhong kwon]

Thanks,

'Brain+PBMC combined data' DEG or 'Brain data' DEG/'PBMC data' DEG separately --> depends on my context

Integrated values DEG vs raw counts DEG --> raw counts are suitable? (I have data with different batch and different tissue type)

Did I get it right?