Skip to contents

Decomposing cell type for spatial transcriptomics data

Source: R/methods.R
dec_celltype.Rd

Identify the cellular composition for single-cell or spot-based spatial transcriptomics data with non-negative regression.

Usage

dec_celltype(
  object,
  sc_data,
  sc_celltype,
  min_percent = 0.5,
  min_nFeatures = 10,
  if_use_normalize_data = T,
  if_use_hvg = F,
  if_retain_other_genes = F,
  if_doParallel = T,
  use_n_cores = NULL,
  iter_num = 1000,
  method = 1,
  env = "base",
  anaconda_path = "~/anaconda3",
  dec_result = NULL
)

Arguments

object

SpaTalk object generated from createSpaTalk.

sc_data

A A data.frame or matrix or dgCMatrix containing counts of single-cell RNA-seq data as the reference, each column representing a cell, each row representing a gene.

sc_celltype

A character containing the cell type of the reference single-cell RNA-seq data.

min_percent

Min percent to predict new cell type for single-cell st_data or predict new cell for spot-based st_data. Default is 0.5.

min_nFeatures

Min number of expressed features/genes for each spot/cell in st_data. Default is 10.

if_use_normalize_data

Whether to use normalized st_data and sc_data with Seurat normalization. Default is TRUE. set it FALSE when the st_data and sc_data are already normalized matrix with other methods.

if_use_hvg

Whether to use highly variable genes for non-negative regression. Default is FALSE.

if_retain_other_genes

Whether to retain other genes which are not overlapped between sc_data and st_data when reconstructing the single-cell ST data. Default is FALSE. Set it TRUE to obtain the constructed single-cell ST data with genes consistent with that in sc_data.

if_doParallel

Use doParallel. Default is TRUE.

use_n_cores

Number of CPU cores to use. Default is all cores - 2.

iter_num

Number of iteration to generate the single-cell data for spot-based data. Default is 1000.

method

1 means using the SpaTalk deconvolution method, 2 means using RCTD, 3 means using Seurat, 4 means using SPOTlight, 5 means using deconvSeq, 6 means using stereoscope, 7 means using cell2location

env

When method set to 6, namely use stereoscope python package to deconvolute, please define the python environment of installed stereoscope. Default is the 'base' environment. Anaconda is recommended. When method set to 7, namely use cell2location python package to deconvolute, please install cell2location to "base" environment.

anaconda_path

When using stereoscope, please define the env parameter as well as the path to anaconda. Default is "~/anaconda3"

dec_result

A matrix of deconvolution result from other upcoming methods, row represents spots or cells, column represents cell types of scRNA-seq reference. See demo_dec_result

Value

SpaTalk object containing the decomposing results.