Skip to contents

Annotate Single-Cell Data Using SingleR

Source: R/75-SingleRAnnotate.R
SingleRAnnotate.Rd

Performs automated cell type annotation using the SingleR package. Supports built-in references (e.g., Human Primary Cell Atlas) or user-provided SingleCellExperiment objects. Integrates seamlessly with Seurat objects by adding prediction results as metadata columns.

Usage

SingleRAnnotate(
  sc,
  verbose = getFuncOption("verbose"),
  ensembl = FALSE,
  cell.ont = c("all", "nonna", "none"),
  legacy = FALSE,
  ref = c("HPCA", "custom"),
  labels = NULL,
  ...
)

Arguments

sc

A single-cell dataset, either:

  • A Seurat object (recommended), or

  • A SingleCellExperiment object.

verbose

Logical. Whether to print progress messages. Default: inherits from package option getOption("SigBridgeR.verbose").

ensembl

Logical scalar indicating whether to convert row names to Ensembl IDs. Genes without a mapping to a non-duplicated Ensembl ID are discarded.

cell.ont

String specifying whether Cell Ontology terms should be included in the colData. If "nonna", all samples without a valid term are discarded; if "all", all samples are returned with (possibly NA) terms; if "none", terms are not added.

legacy

Logical scalar indicating whether to pull data from ExperimentHub. By default, we use data from the gypsum backend.

ref

Reference data for annotation. One of:

  • "HPCA": Uses the Human Primary Cell Atlas from the celldex package.

  • A SingleCellExperiment object containing pre-labeled reference cells.

Note: The placeholder option "custom" is not valid—users must provide an actual reference object.

A numeric matrix of (usually normalized and log-transformed) expression values from a reference dataset, or a SummarizedExperiment object containing such a matrix; see SingleR::trainSingleR for details. Alternatively, a list or List of SummarizedExperiment objects or numeric matrices containing multiple references. Row names may be different across entries but only the intersection will be used,

labels

A character vector of cell type labels corresponding to the columns of ref. Required when ref is a custom SingleCellExperiment. For "HPCA", defaults to ref$label.main.

Alternatively, if ref is a list, labels should be a list of the same length. Each element should contain a character vector or factor specifying the labels for the columns of the corresponding element of ref.

...

Additional arguments passed to SingleR::SingleR() (e.g., assay.type.ref, genes).

Value

  • If input is a Seurat object: returns the same object with three new metadata columns:

    SingleR_labels

    Predicted cell type labels.

    SingleR_delta_next

    Confidence score (difference between top and second-best scores).

    SingleR_pruned_labels

    Labels after pruning low-confidence assignments.

  • If input is a SingleCellExperiment: returns the SingleR prediction result

Requirements

  • The celldex package is required when using ref = "HPCA".

  • All gene names must be consistent between test and reference datasets (e.g., Ensembl or symbol).

See also

Other Single_Cell_Annotation_Method: CellTypistAnnotate(), RegisterAnnoMethod(), SCAnnotateStrategy, mLLMCelltypeAnnotate()

Examples

if (FALSE) { # \dontrun{
# Using HPCA reference (requires celldex)
annotated_seurat <- SingleRAnnotate(seurat_obj, ref = "HPCA")

# Using custom reference
library(celldex)
ref_sce <- BlueprintEncodeData()
annotated_sce <- SingleRAnnotate(
  sc = my_sce,
  ref = ref_sce,
  labels = ref_sce$label.main,
  de.n = 30
)
} # }