makePerCellDF: Create a per-cell data.frame

In LTLA/scuttle: Single-Cell RNA-Seq Analysis Utilities

Create a per-cell data.frame

Description

Create a per-cell data.frame (i.e., where each row represents a cell) from a SingleCellExperiment, most typically for creating custom ggplot2 plots.

Usage

makePerCellDF(
  x,
  features = NULL,
  assay.type = "logcounts",
  use.coldata = TRUE,
  use.dimred = TRUE,
  use.altexps = TRUE,
  prefix.altexps = FALSE,
  check.names = FALSE,
  swap.rownames = NULL,
  exprs_values = NULL,
  use_dimred = NULL,
  use_altexps = NULL,
  prefix_altexps = NULL,
  check_names = NULL
)

Arguments

x	A SingleCellExperiment object. This is expected to have non-NULL row names.
features	Character vector specifying the features for which to extract expression profiles across cells. May also include features in alternative Experiments if permitted by use.altexps.
assay.type	String or integer scalar indicating the assay to use to obtain expression values. Must refer to a matrix-like object with integer or numeric values.
use.coldata	Logical scalar indicating whether column metadata of x should be included. Alternatively, a character or integer vector specifying the column metadata fields to use.
use.dimred	Logical scalar indicating whether data should be extracted for dimensionality reduction results in x. Alternatively, a character or integer vector specifying the dimensionality reduction results to use.
use.altexps	Logical scalar indicating whether (meta)data should be extracted for alternative experiments in x. Alternatively, a character or integer vector specifying the alternative experiments to use.
prefix.altexps	Logical scalar indicating whether altExp-derived fields should be prefixed with the name of the alternative Experiment.
check.names	Logical scalar indicating whether column names of the output should be made syntactically valid and unique.
swap.rownames	String specifying the rowData column containing the features. If NULL, rownames(x) is used.
exprs_values, use_dimred, use_altexps, prefix_altexps, check_names	Soft-deprecated equivalents of the arguments described above.

Details

This function enables us to conveniently create a per-feature data.frame from a SingleCellExperiment. Each row of the returned data.frame corresponds to a column in x, while each column of the data.frame corresponds to one aspect of the (meta)data in x.

Columns are provided in the following order:

1. Columns named according to the entries of features represent the expression values across cells for the specified feature in the assay.type assay.

2. Columns named according to the columns of colData(x) represent column metadata variables. This consists of all variables if use.coldata=TRUE, no variables if use.coldata=FALSE, and only the specified variables if use.coldata is set to an integer or character vector.

3. Columns named in the format of <DIM>.<NUM> represent the <NUM>th dimension of the dimensionality reduction result <DIM>. This is generated for all dimensionality reduction results if use.dimred=TRUE, none if use.dimred=FALSE, and only the specified results if use.dimredis set to an integer or character vector.

4. Columns named according to the row names of successive alternative Experiments, representing the assay data in these objects. These columns are only included if they are specified in features and if use.altexps is set. Column names are prefixed with the name of the alternative Experiment if prefix.altexps=TRUE.

By default, nothing is done to resolve syntactically invalid or duplicated column names. check_names=TRUE, this is resolved by passing the column names through make.names. Of course, as a result, some columns may not have the same names as the original fields in x.

Value

A data.frame containing one field per aspect of data in x - see Details. Each row corresponds to a cell (i.e., column) of x.

Author(s)

Aaron Lun

See Also

makePerFeatureDF, for the feature-level equivalent.

Examples

sce <- mockSCE()
sce <- logNormCounts(sce)
reducedDim(sce, "PCA") <- matrix(rnorm(ncol(sce)*10), ncol=10) # made-up PCA.

df <- makePerCellDF(sce, features="Gene_0001")
head(df)

LTLA/scuttle documentation built on Oct. 28, 2024, 9:45 a.m.