About the subgroups in pseudobulk methods

[Xiangyi-Deng]

Link here:

I have a question about pseudobulk method when doing DEG analysis.

The prerequisites is that the replicates should be necessary, right ?

If so, the question comes:

if I have 3 donors with paired Bm and Blood, and for each sample, there are four kinds of different stimulations

These 6 samples, scRNA sequencing and processed until clusters annotation.

now, in order to do DEG analysis with lower false positive rate, I want to try the pseudobulk method.

At total donor level, between Bm and Blood, three donors can be regards as three replicates, pseudobulk works here.

But for each donor, each stimulation, between Bm and Blood, at all clusters level, in this case, all clusters can be replicates ?

So, in order to do DEG analysis, at different levels, there are different subgroups.

But for each analysis, I should create related subgroups with used conditions, or create subgroups with all conditions and use it for all DEG analysis?

For example:

3 donors, totally 6 samples, with four kinds stimulations, repectively.

at the beginning, do AggregateExpression() at donor, tissue(sample), stimulation, clusters, totally 4 levels ?

pseudobulk the counts based on donor-condition-celltype
pseudo_ifnb <- AggregateExpression(ifnb, assays = "RNA", return.seurat = T, group.by = c("stim", "donor_id", "tissue","seurat_annotations"))

because for this group information, the gene expression can be divided into all subgroups.

And then just select groups to do specific DEG analysis.

Or: just split into necessary subgroups:
pseudo_ifnb <- AggregateExpression(ifnb, assays = "RNA", return.seurat = T, group.by = c("stim", "donor_id", "tissue"))

I think it is different in some degrees.

Could somebody give me some advice on this problem.

Thanks a lot.

[Xiangyi-Deng]

Nobody respond to my question ?