Assay missing after subset from scATAC object by feature

[RitataLU]

Hi Signac team,

I would like to perform the same idea from scRNA clustering with features (select_genes) on scATAC-seq, like in scRNA, there is a step after subset object,
object <- RunPCA(object = object, features=selected_genes, npcs = 30, verbose = FALSE), and then do the RunUMAP,FindNeighbors,FindClusters.

I would like to use the same way on scRNA, clustering without some genes, but I couldn't find the similar way on RunSVD or RunUmap.., so what I did is sub_atac_object <- subset (atac_object, feature = selected_genes)
and then do the standard method on clustering..,
but I found that after I got the new clusters in sub_atacobject, I couldn't do the find motif between clusters. since it looks like some assay missing.
originally ataco_bject has four assays, RNA, peaks,peaksinters, chromvar, but sub_atac_object only has RNA assay.

I was wondering how could I fix this issue, or do you have suggestions on my processing steps?

Thanks!
Rita

[timoast]

Those other assays are probably removed because they don't contain any of the features you asked to keep.

If you want to run SVD using a pre-defined set of features, you can use the features argument in RunSVD(). See the documentation for RunSVD() for more information.