runUMAP: runUMAP

In bernard2012/Giotto: Spatial single-cell transcriptomics toolbox

Description

run UMAP

Usage

runUMAP(
  gobject,
  expression_values = c("normalized", "scaled", "custom"),
  reduction = c("cells", "genes"),
  dim_reduction_to_use = "pca",
  dim_reduction_name = "pca",
  dimensions_to_use = 1:10,
  name = "umap",
  genes_to_use = NULL,
  return_gobject = TRUE,
  n_neighbors = 40,
  n_components = 2,
  n_epochs = 400,
  min_dist = 0.01,
  n_threads = 1,
  spread = 5,
  set_seed = T,
  seed_number = 1234,
  verbose = T,
  ...
)

Arguments

gobject	giotto object
expression_values	expression values to use
reduction	cells or genes
dim_reduction_to_use	use another dimension reduction set as input
dim_reduction_name	name of dimension reduction set to use
dimensions_to_use	number of dimensions to use as input
name	arbitrary name for UMAP run
genes_to_use	if dim_reduction_to_use = NULL, which genes to use
return_gobject	boolean: return giotto object (default = TRUE)
n_neighbors	UMAP param: number of neighbors
n_components	UMAP param: number of components
n_epochs	UMAP param: number of epochs
min_dist	UMAP param: minimum distance
n_threads	UMAP param: threads to use
spread	UMAP param: spread
set_seed	use of seed
seed_number	seed number to use
verbose	verbosity of function
...	additional UMAP parameters

Details

See umap for more information about these and other parameters.

• Input for UMAP dimension reduction can be another dimension reduction (default = 'pca')

• To use gene expression as input set dim_reduction_to_use = NULL

• If dim_reduction_to_use = NULL, genes_to_use can be used to select a column name of highly variable genes (see calculateHVG) or simply provide a vector of genes

• multiple UMAP results can be stored by changing the name of the analysis

Value

giotto object with updated UMAP dimension recuction

Examples

# 1. create giotto object
expr_path = system.file("extdata", "seqfish_field_expr.txt", package = 'Giotto')
loc_path = system.file("extdata", "seqfish_field_locs.txt", package = 'Giotto')
VC_small <- createGiottoObject(raw_exprs = expr_path, spatial_locs = loc_path)

# 2. normalize giotto
VC_small <- normalizeGiotto(gobject = VC_small, scalefactor = 6000)
VC_small <- addStatistics(gobject = VC_small)

# 3. dimension reduction
VC_small <- calculateHVG(gobject = VC_small)
VC_small <- runPCA(gobject = VC_small)
VC_small <- runUMAP(VC_small, dimensions_to_use = 1:5, n_threads = 2)
plotUMAP(gobject = VC_small)

bernard2012/Giotto documentation built on Sept. 22, 2020, 10:29 a.m.