runPCAprojection

runs a Principal Component Analysis on a random subset + projection

runPCAprojection(
  gobject,
  spat_unit = NULL,
  feat_type = NULL,
  expression_values = c("normalized", "scaled", "custom"),
  reduction = c("cells", "feats"),
  random_subset = 500,
  name = "pca.projection",
  feats_to_use = "hvf",
  return_gobject = TRUE,
  center = TRUE,
  scale_unit = TRUE,
  ncp = 100,
  method = c("irlba"),
  method_params = BiocParallel::SerialParam(),
  rev = FALSE,
  set_seed = TRUE,
  seed_number = 1234,
  verbose = TRUE,
  ...
)

Arguments

gobject

giotto object

spat_unit

spatial unit

feat_type

feature type

expression_values

expression values to use

reduction

cells or genes

random_subset

numeric. random subset to perform PCA on

name

arbitrary name for PCA run

feats_to_use

subset of features to use for PCA

return_gobject

boolean: return giotto object (default = TRUE)

center

center data first (default = TRUE)

scale_unit

scale features before PCA (default = TRUE)

ncp

number of principal components to calculate

method

which implementation to use

method_params

BiocParallelParam object

rev

do a reverse PCA

set_seed

use of seed

seed_number

seed number to use

verbose

verbosity of the function

...

additional parameters for PCA (see details)

Value

giotto object with updated PCA dimension recuction

Details

See runPCA and PCA for more information about other parameters. This PCA implementation is similar to runPCA, except that it performs PCA on a subset of the cells or features, and predict on the others. This can significantly increase speed without sacrificing accuracy too much.

• feats_to_use = NULL: will use all features from the selected matrix

• feats_to_use = <hvg name>: can be used to select a column name of highly variable features, created by (see calculateHVF)

• feats_to_use = c('geneA', 'geneB', ...): will use all manually provided features

Examples

g <- GiottoData::loadGiottoMini("visium")
#> 1. read Giotto object
#> 2. read Giotto feature information
#> 3. read Giotto spatial information
#> 3.1 read Giotto spatial shape information
#> 3.2 read Giotto spatial centroid information
#> 3.3 read Giotto spatial overlap information
#> 4. read Giotto image information
#> 
#> checking default envname 'giotto_env'
#> a system default python environment was found
#> Using python path:
#>  "/usr/bin/python3"

runPCAprojection(g)
#> "hvf" column was found in the feats metadata information and will be
#>  used to select highly variable features
#> Warning: ncp >= minimum dimension of x, will be set to minimum
#>                 dimension of x - 1
#> pca random subset: start
#> Warning: You're computing too large a percentage of total singular values, use a standard svd instead.
#> Warning: did not converge--results might be invalid!; try increasing work or maxit
#> pca random subset: done
#> pca prediction: start
#> pca prediction: done
#> An object of class giotto 
#> >Active spat_unit:  cell 
#> >Active feat_type:  rna 
#> [SUBCELLULAR INFO]
#> polygons      : cell 
#> [AGGREGATE INFO]
#> expression -----------------------
#>   [cell][rna] raw normalized scaled
#> spatial locations ----------------
#>   [cell] raw
#> spatial networks -----------------
#>   [cell] Delaunay_network spatial_network
#> spatial enrichments --------------
#>   [cell][rna] cluster_metagene DWLS
#> dim reduction --------------------
#>   [cell][rna] pca custom_pca pca.projection umap custom_umap tsne
#> nearest neighbor networks --------
#>   [cell][rna] sNN.pca custom_NN
#> attached images ------------------
#> images      : alignment image 
#> 
#> 
#> Use objHistory() to see steps and params used