Questions regarding the input data for scVelo

[denvercal1234GitHub]

Hi @zhanghao-njmu,

Thank you for the package!

Related to #109, I hope to get some clarification regarding the input data.

I had loom objects for different samples generated by velocyto, and a Seurat object that had been fully analysed (with clustering and UMAP etc).

I was able to concatenate my loom objects into 1 object and converted it to Seurat object, but this object has not been in anyway processed (e.g., normalise etc that are typically done in scVelo pipeline).

Q1. Do I need to process the spliced and unspliced counts somehow before or after adding to my Seurat cluster object?

Q2. How might I add the ambiguous, matrix, spliced, unspliced to my Seurat GEX cluster object? Do I do as below?

test[["ambiguous"]] <- CreateAssayObject(counts = concat_adata_Seurat@assays$ambiguous)
test[["matrix"]] <- CreateAssayObject(counts = concat_adata_Seurat@assays$matrix)
test[["spliced"]] <- CreateAssayObject(counts = concat_adata_Seurat@assays$splice)

Thank you for your help!

My Seurat GEX cluster object:

My loom-converted Seurat object: