Combine ExperimentSubset objects by columns or rows

cbind(...) and rbind(...) combine objects as for the class they extend, and keep the subsets of every object, with their own assays, rowData, colData, reducedDims and altExps. Subsets of the second and later objects are moved to where that object's columns (for cbind) or rows (for rbind) are in the result. Subset names must be unique across the objects, and differ from the assay names of all of them.

combineCols(x, ...) and combineRows(x, ...) combine objects whose rows (for combineCols) or columns (for combineRows) differ, as for the class they extend: the result has all of the rows (or columns), matched by name, and fills the gaps. They keep the subsets of every object too, matching each subset's rows (or columns) by name to the result's, or by position with use.names = FALSE.

Usage

# S4 method for class 'SubsetRangedSummarizedExperiment'
cbind(..., deparse.level = 1)

# S4 method for class 'SubsetRangedSummarizedExperiment'
rbind(..., deparse.level = 1)

# S4 method for class 'SubsetRangedSummarizedExperiment'
combineCols(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetRangedSummarizedExperiment'
combineRows(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetSingleCellExperiment'
cbind(..., deparse.level = 1)

# S4 method for class 'SubsetSingleCellExperiment'
rbind(..., deparse.level = 1)

# S4 method for class 'SubsetSingleCellExperiment'
combineCols(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetSingleCellExperiment'
combineRows(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetSummarizedExperiment'
cbind(..., deparse.level = 1)

# S4 method for class 'SubsetSummarizedExperiment'
rbind(..., deparse.level = 1)

# S4 method for class 'SubsetSummarizedExperiment'
combineCols(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetSummarizedExperiment'
combineRows(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetSpatialExperiment'
cbind(..., deparse.level = 1)

# S4 method for class 'SubsetSpatialExperiment'
rbind(..., deparse.level = 1)

# S4 method for class 'SubsetSpatialExperiment'
combineCols(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetSpatialExperiment'
combineRows(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetTreeSummarizedExperiment'
cbind(..., deparse.level = 1)

# S4 method for class 'SubsetTreeSummarizedExperiment'
rbind(..., deparse.level = 1)

# S4 method for class 'SubsetTreeSummarizedExperiment'
combineCols(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

# S4 method for class 'SubsetTreeSummarizedExperiment'
combineRows(x, ..., delayed = TRUE, fill = NA, use.names = TRUE)

Arguments

...

ExperimentSubset objects of the same class, which must have the same rows for cbind or the same columns for rbind. Convert a plain object, such as a SingleCellExperiment, with ExperimentSubset() first, which keeps all of its data: combining it as it is makes R warn about "multiple direct matches", and if it comes first, the result is a plain object without subsets. For combineCols and combineRows, the objects to combine with x.

deparse.level

Passed on to the method of the class the objects extend. Default 1.

x

For combineCols and combineRows, the first ExperimentSubset object to combine.

delayed, fill

Passed on to the method of the class the objects extend: whether to combine assays as DelayedArray objects, and the value for the gaps. Defaults TRUE and NA.

use.names

Whether to match rows (for combineCols) or columns (for combineRows) by name, as the method of the class the objects extend does; if FALSE, they are matched by position and the objects must have the same number of them. Default TRUE.

Value

For cbind and rbind, an object of the same class as the objects in .... For combineCols and combineRows, the ExperimentSubset class matching the result of the method of the class x extends; see Details.

Details

For some classes, combineCols() and combineRows() of the class the objects extend return a different class: combineCols() of a TreeSummarizedExperiment or SpatialExperiment returns a SingleCellExperiment, and combineRows() of a SingleCellExperiment, TreeSummarizedExperiment or SpatialExperiment returns a RangedSummarizedExperiment. The result is then the matching ExperimentSubset class, such as SubsetSingleCellExperiment, and each subset is converted to that class too, so it loses what the class can't hold, as the main object does: for example its reducedDims and altExps in a SubsetRangedSummarizedExperiment.

combineCols() and combineRows() use the method for the class of x only, so if x is a plain object, such as a SingleCellExperiment, the result is a plain object without subsets. Convert it with ExperimentSubset() first.

Examples

# two objects with the same 10 genes and 20 cells each
set.seed(1)
counts1 <- matrix(rpois(200, lambda = 5), nrow = 10)
rownames(counts1) <- paste0("gene", 1:10)
colnames(counts1) <- paste0("cellA", 1:20)
counts2 <- matrix(rpois(200, lambda = 5), nrow = 10)
rownames(counts2) <- paste0("gene", 1:10)
colnames(counts2) <- paste0("cellB", 1:20)
es1 <- ExperimentSubset(list(counts = counts1))
es2 <- ExperimentSubset(list(counts = counts2))
# a subset in each object
es1 <- createSubset(es1, subsetName = "firstA", rows = 1:5, cols = 1:8)
es2 <- createSubset(es2, subsetName = "firstB", rows = 1:5, cols = 1:8)
# combine the cells of both objects
esBoth <- cbind(es1, es2)
# both subsets are kept
subsetNames(esBoth)
#> [1] "firstA" "firstB"
subsetColnames(esBoth, subsetName = "firstB")
#> [1] "cellB1" "cellB2" "cellB3" "cellB4" "cellB5" "cellB6" "cellB7" "cellB8"

# a third object, with genes 6 to 10 and 5 genes the others don't have
counts3 <- matrix(rpois(200, lambda = 5), nrow = 10)
rownames(counts3) <- paste0("gene", 6:15)
colnames(counts3) <- paste0("cellC", 1:20)
es3 <- ExperimentSubset(list(counts = counts3))
# a subset of the genes only the third object has
es3 <- createSubset(es3, subsetName = "newGenes", rows = 6:10)
# combine the cells, keeping the genes of both; the gaps are NA
esAll <- combineCols(es1, es3)
dim(esAll)
#> [1] 15 40
# the subsets of both objects are kept
subsetNames(esAll)
#> [1] "firstA"   "newGenes"
# the new subset keeps its genes
subsetRownames(esAll, subsetName = "newGenes")
#> [1] "gene11" "gene12" "gene13" "gene14" "gene15"