ranger_importances: ranger_importances

In jspaezp/sctree: Tree based analysis of single rna-seq clusters

Description

Calculates ranger-based variable importances for data frames and Seurat objects

Usage

ranger_importances(x, ...)

FindAllMarkers_ranger.Seurat(
  object,
  genes_use = Seurat::VariableFeatures(object),
  ...
)

Arguments

...	additional arguments to be passed to ranger
object	a Seurat or data frame object
genes_use	a character vector indicating which genes to use in the classification. currently implemented only for Seurat objects. (for data frames one can simply subset the input data frame)
cluster	a cluster name for which the markers will be found
pval_cutoff	p value cutoff for the markers
imp_method	importance method, either of "janitza" or "altmann"
num.trees	number of trees to be build using ranger
return_what	a subset of "ranger_fit", "importances_ranger", "signif_importances_ranger", defaults to signif_importances_ranger
warn.imp.method	logical indicating wether warning should be issued when few negative importances are found to calculate the p.values in ranger.
return	ranger_fit, importances_ranger, signif_importances_ranger

Value

by default returns a data frame with the importances and p values but this behavior can be modified by the

Functions

• FindAllMarkers_ranger.Seurat: Calculate variable importances to each cluster in a Seurat object

Examples

    head(ranger_importances(Seurat::pbmc_small, cluster = 'ALL', warn.imp.method = FALSE))
    
#          importance pvalue     gene
# HLA-DPB1   8.226410      0 HLA-DPB1
# S100A9     6.648748      0   S100A9
# S100A8     6.537872      0   S100A8
# HLA-DQA1   3.103721      0 HLA-DQA1
# GNLY       1.779237      0     GNLY

jspaezp/sctree documentation built on April 30, 2020, 10:36 p.m.