Average Expression Profile Calculation

Calculating the Average Expression Profile From a scRNA-seq Reference Dataset

Because niche-DE compares observed gene expression against the average expression for a cell type, making an average expression profile matrix is critical.

From Raw Data

If you have raw data matrices, the expression profile matrix can be created using the function ‘CreateLibraryMatrix’. This function takes in 2 arguments

Arguments

• data: Single cell rna-seq counts matrix. Dimension should be #cells/spots by #genes
  • cell type: Cell type assignment matrix. First column is cell names and second column is cell type assignment.

#read in data
data('liver_met_data')
data('liver_met_CT')
#create library matrix
L = CreateLibraryMatrix(liver_met_data,liver_met_CT)
View(L)

From Seurat object

If your reference dataset is a seurat object, you can use the function ‘CreateLibrarymatrixFromSeurat’. This function takes in 2 arguments

Arguments

• seurat object: A seurat object
  • assay: The assay from which to extract the counts matrix to calculate the average expression profile Note that the cell types of the seurat object are assumed to be available via the command ‘Idents(seurat object)’

#read in data
data('liver_met_ref')
#create library matrix
CreateLibraryMatrixFromSeurat(liver_met_ref,assay = 'RNA')