runMapping: Run single-cell mapping

In heilandd/SPATAwrappers: What the Package Does (One Line, Title Case)

Run single-cell mapping

Description

Mapping the output of "getSingleCellDeconv()"to the best matching cell from the reference dataset

Usage

runMapping(
  object,
  scDF,
  ref.new,
  cell_type_var = "annotation_level_4",
  model = c("oGA", "BF"),
  subset_ref = T,
  only_var = T,
  n_feature = 3000,
  max_cells = 10000,
  AE_norm = F,
  dropout = 0,
  bottleneck = 16,
  activation = "relu",
  layers = c(128, 64, 32),
  epochs = 20,
  multicore = T,
  workers = 16,
  iter = 200,
  iter_GA = 20,
  nr_mut = 2,
  nr_offsprings = 7,
  cross_over_point = 0.5,
  ram = 60
)

Arguments

object	SPATA2 object
scDF	Data.frame; Output of the getSingleCellDeconv()
ref.new	Seurat Object; Reference dataset used for runRCTD()
cell_type_var	Character value; The col of the Seurat meta data indicating the cell type annotations
model	Character value; Model "BF" randomly select cell compositions and select the best match. Model "oGA" will perform a conditional genetic algorithm to find the best match.
subset_ref	Logical, if TRUE, Seurat object will we downsized to increase speed.
only_var	Logical, if TRUE only consider variable genes for mapping
n_feature	Integer value: Number of variable genes
max_cells	Integer value: Number of cells for downsampling
AE_norm	Logical, if TRUE use an autoencoder for data integration and normalization (recommended)
multicore	Logical, if TRUE use multicore
workers	Integer, Number of cors
iter	Integer: For model BF: Number of random spot compositions; For model oGA size of initial population
iter_GA	Integer: Number of iterations of the oGA model
nr_mut	Integer: Number of mutations
nr_offsprings	Integer: Number of offsprings
cross_over_point	Numeric value: Percentage of cross-over cutt-off
ram	Integer: GB of ram can be used for multicore session

Author(s)

Dieter Henrik Heiland

heilandd/SPATAwrappers documentation built on Oct. 2, 2022, 1:40 p.m.