Seurat (version 2.3.4)

AssessSplit: Assess Cluster Split

Description

Method for determining confidence in specific bifurcations in the cluster tree. Use the Out of Bag (OOB) error of a random forest classifier to judge confidence.

Usage

AssessSplit(object, node, cluster1, cluster2, genes.training = NULL,
  print.output = TRUE, ...)

Arguments

object

Seurat object

node

Node in the cluster tree in question

cluster1

First cluster to compare

cluster2

Second cluster to compare

genes.training

A vector of genes to use to train the classifier, defaults to rownames(x = object@data)

print.output

Print the OOB error for the classifier

...

Arguments passed on to BuildRFClassifier

training.genes

Vector of genes to build the classifier on

training.classes

Vector of classes to build the classifier on

verbose

Additional progress print statements

Value

Returns the Out of Bag error for a random forest classifier trained on the split from the given node

Examples

# NOT RUN {
pbmc_small
pbmc_small <- FindClusters(object = pbmc_small, reduction.type = "pca",
                           dims.use = 1:10, resolution = 1.1, save.SNN = TRUE)
pbmc_small <- BuildClusterTree(pbmc_small, reorder.numeric = TRUE, do.reorder = TRUE)
# Assess based on a given node
AssessSplit(pbmc_small, node = 11)
# Asses based on two given clusters (or vectors of clusters)
AssessSplit(pbmc_small, cluster1 = 5, cluster2 = 6)
AssessSplit(pbmc_small, cluster1 = 4, cluster2 = c(5, 6))

# }

Run the code above in your browser using DataCamp Workspace