SpatialAnno.

SpatialAnno, an efficient and accurate annotation method for spatial transcriptomics datasets, with capability of effectively leveraging a large number of non-marker genes with “qualitative” information about marker genes, without using a reference dataset.

SpatialAnno(
  X,
  Adj_sp,
  markers,
  initial = c("SCINA", "scSorter"),
  Unknown = TRUE,
  xi_int = 1.5,
  xi_grid = seq(0.1, 2.5, by = 0.2),
  mode = "full",
  q = 15
)

Arguments

X

a normalized matrix. The row is for spots/cells, and the column is for genes.

Adj_sp

a integer vector used in SC.MEB. The default is 2:10

markers

a list that contain cell types and corresponding cell-type specific markers.

initial

character specifying which method to generate initial value. In general, We recommend to use the method of SCINA for high-dimensional non-markers and the method of scSorter for low-dimensional non-marker. The default is SCINA.

Unknown

logical value specifying whether or not include unknown

xi_int

a initial value of smooth parameter

xi_grid

a pre-defined vector of smooth parameter

mode

a string specifying which mode to use. It has four modes: "full", "annotation", "spatial+annotation" and "annotation+factor". The "full" mode means SpatialAnno runs with all three components, marker component, spatial component, and factor component. The "annotation" mode means SpatialAnno runs with only marker component. The "spatial+annotation" means SpatialAnno runs with marker and spatial components. The "annotation+factor" mode means SpatialAnno runs with marker and non-marker components. The default is "full".

q

the dimension of embedding generated by SpatialAnno. The default is 15

Value

a list.

Details

SpatialAnno, an efficient and accurate annotation method for spatial transcriptomics datasets, with capability of effectively leveraging a large number of non-marker genes with “qualitative” information about marker genes, without using a reference dataset.