Hi all,
I performed scRNAseq analysis and obtained DEGs of a cluster (relative to all other cells) by Seurat. I also performed bulkRNAseq of the same cell type (same tissue and condition) of this particular cluster and had processed the htseq counts using the DESeq2 pipeline.
I hope to now compare the genes that get picked up by both platforms and especially the genes that might be picked up by bulkRNAseq that were missed by scRNAseq.
I am thinking of plotting the VST-transformed data of the top highly expressed genes from bulkRNAseq across biological replicates (donors), and in a separate plot, visualise the SCT-transformed expression of DEGs of this cluster (that also present in RNAseq genes) from scRNAseq.
Anyone has any additional input?
Thank you for your help.
My take is that RNA-seq (be it bulk or single-cell) is inherently relative. I don't think you can really "compare" the absolute expression values. Did you do bulk of only a single celltype, so there is nothing in this experiment to compare to?
Thank you ATpoint for your response. In bulkRNAseq, I did have two cell types to compare. But, I want to also see if bulkRNAseq might have picked up genes that were missed by scRNAseq for the cell type of interest (given bulkRNAseq provided more sequencing depth).
With this goal, do you think it is reasonable to pull the top 100 DEGs by scRNAseq for the cluster of interest (most of them likely are highly expressed by these cells relative to other cells) and see if any additional genes (aside from the genes that shared between platforms) got picked up among the top 100 highly expressed genes by bulkRNAseq for this cell type? By a simple Venn instead of comparing directly the expression levels.
Note that the two cell types sequenced by bulkRNAseq were detected as two clusters among many clusters in my scRNAseq. But here, I just want to know whether the bulkRNAseq provided any additional genes to characterise these cell types beyond what scRNAseq already showed.