TrainCellTypist: Train Celltypist

In bimberlabinternal/RIRA_classification: A package for the Rhesus Immune Reference Atlas (RIRA): A multi-tissue single-cell landscape of immune cells

Train Celltypist

Description

Train celltypist using the data in a seurat object and saves the resulting model

Usage

TrainCellTypist(
  seuratObj,
  labelField,
  modelFile,
  minCellsPerClass = 20,
  assayName = Seurat::DefaultAssay(seuratObj),
  tempFileLocation = NULL,
  dropAmbiguousLabelValues = TRUE,
  excludedClasses = NULL,
  featureInclusionList = NULL,
  featureExclusionList = NULL
)

Arguments

seuratObj	The seurat object
labelField	The field in [e-mail] holding the labels for training
modelFile	The path to save the model
minCellsPerClass	If provided, any classes (and corresponding cells) with fewer than this many cells will be dropped from the training data
assayName	The name of the assay to use
tempFileLocation	The location where temporary files (like the annData version of the seurat object), will be written.
dropAmbiguousLabelValues	If true, and label value with a comma will be dropped.
excludedClasses	A vector of labels to discard.
featureInclusionList	If provided, the input count matrix will be subset to just these features. If used, Seurat::NormalizeData will be re-run.
featureExclusionList	If provided, the input count matrix will be subset to remove these features. If used, Seurat::NormalizeData will be re-run.

bimberlabinternal/RIRA_classification documentation built on April 14, 2025, 5:59 p.m.