test_classifier: Testing process.

In grisslab/scAnnotatR: Pretrained learning models for cell type prediction on single cell RNA-sequencing data

Testing process.

Description

Testing process.

Usage

test_classifier(
  classifier,
  test_obj,
  assay,
  slot = NULL,
  tag_slot,
  target_cell_type = NULL,
  parent_classifier = NULL,
  parent_tag_slot = "predicted_cell_type",
  path_to_models = "default",
  zscore = TRUE,
  ambiguous_chars = NULL
)

## S4 method for signature 'scAnnotatR'
test_classifier(
  classifier,
  test_obj,
  assay,
  slot = NULL,
  tag_slot,
  target_cell_type = NULL,
  parent_classifier = NULL,
  parent_tag_slot = "predicted_cell_type",
  path_to_models = "default",
  zscore = TRUE,
  ambiguous_chars = NULL
)

Arguments

classifier	scAnnotatR classification model
test_obj	object that can be used for testing
assay	name of assay to use in test_object
slot	type of expression data to use in test_object. For Seurat object, some available types are: "counts", "data" and "scale.data". Ignore this if test_obj is SingleCellExperiment object.
tag_slot	string, name of annotation slot indicating cell tag/label in the testing object. Strings indicating cell types are expected in this slot. Expected values are string (A-Z, a-z, 0-9, no special character accepted) or binary/logical, 0/"no"/F/FALSE: not being new cell type, 1/"yes"/T/TRUE: being new cell type.
target_cell_type	vector indicating other cell types than cell labels that can be considered as the main cell type in classifier, for example, c("plasma cell", "b cell", "b cells", "activating b cell"). Default as NULL.
parent_classifier	scAnnotatR object corresponding to classification model for the parent cell type
parent_tag_slot	string, name of tag slot in cell meta data indicating pre-assigned/predicted parent cell type. Default field is "predicted_cell_type". The slot must contain only string values.
path_to_models	path to the folder containing the list of models. As default, the pretrained models in the package will be used. If user has trained new models, indicate the folder containing the new_models.rda file.
zscore	boolean, whether gene expression is transformed to zscore
ambiguous_chars	List of characters to indicate ambiguous cells. For more details see filter_cells.

Value

result of testing process in form of a list, including predicted values, prediction accuracy at a probability threshold, and roc curve information.

Note

Only one cell type is expected for each cell. Ambiguous cell type, such as: "T cells/NK cells/ILC", will be ignored. Subtypes used in testing model for parent cell types can be indicated as parent cell type, or can be indicated in target_cell_type. For example, when testing for B cells, plasma cells can be annotated as B cells, or target_cell_type is set c("plasma cells").

Examples

# load small example dataset
data("tirosh_mel80_example")

# train the classifier
selected_marker_genes_B = c("CD19", "MS4A1", "CD79A")
set.seed(123)
classifier_b <- train_classifier(train_obj = tirosh_mel80_example,
assay = 'RNA', slot = 'counts', marker_genes = selected_marker_genes_B,
cell_type = "b cells", tag_slot = 'active.ident')

# test the classifier, target cell type can be in other formats or
# alternative cell type that can be considered as the classified cell type
classifier_b_test <- test_classifier(classifier = classifier_b,
test_obj = tirosh_mel80_example, assay = 'RNA', slot = 'counts',
tag_slot = 'active.ident', target_cell_type = c("B cell"))
classifier_b_test

grisslab/scAnnotatR documentation built on March 20, 2023, 2:42 a.m.