Is the goup.by function of analyse_sc_clusters() available now?

[bravedu01]

The goup.by function of analyse_sc_clusters() will be highly helpful to compare the treatment with control group.

[jgriss]

Thanks for the feedback @bravedu01 !

As you probably now, this can currently be easily achieved by adapting the Ident (for Seurat objects). We are aware that this is an extra line of code but that's the reason why we haven't tackled this feature yet.

[bravedu01]

Thanks for the suggestion.

I know setting ident will be able to compare the treatment with control for the whole object. Is there a way to compare the treatment with control for all the subtypes. eg. analyse_sc_clusters(pbmc, group.by="Group"), could we get the comparison of pathways between treatment and control in B/T/NK/Mono subtypes of pbmc?

The advantage of doing this is that we could obtain a heatmap comparing the selected pathway across treatments and subclusters(two layers at the same time) and figure out which subcluster is the most interested one.

[jgriss]

Hi @bravedu01

Yes, you just have to adapt the primary ident:

# save the current ident
seurat_obj$old_ident <- Idents(seurat_obj)

# set "Group" as primary ident
Idents(seurat_obj) <- seurat_obj$Group

# now run ReactomeGSA