predict_expr_loc: Predict gene expression at locations

In lambdamoses/novoSpaRc: Spatial Reconstruction of Tissues from scRNA-seq Data

Description

With the probabilistic assignment of cells to locations from optimal transport, we can predict gene expression at each location based on the gene count matrix. Let X denote the gene count matrix with genes in rows and cells in columns, and T denote the probabilistic assignment of cells to locations, with cells in rows and locations in columns. Gene expression at each location can be predicted by XT. This function will use this to predict gene expression. It will also scale the prediction so the mean gene expression across locations will match the mean gene expression among cells in scRNA-seq.

Usage

predict_expr_loc(X, cell_loc, transposed = FALSE, scale = TRUE)

Arguments

X	Numeric matrix (can be sparse) with genes in rows and cells in columns. If genes are in columns, then set transposed = TRUE.
cell_loc	Matrix that probabistically assigns cells to locations, made from the function gw_assign.
transposed	Logical, whether the matrix has cells in rows rather than in columns. Defaults to FALSE.
scale	Logical, whether predictions should be scaled so their means are the same as in single cell RNA-seq data supplied in x. Defaults to TRUE.

Value

A numeric matrix with genes in rows and locations in columns.

lambdamoses/novoSpaRc documentation built on May 12, 2019, 3:14 p.m.