subset-odm: Subset using the '[' operator.

In ondisc: Fast, Universal, and Intuitive Computing on Large-Scale Single-Cell Data

Description

Apply the [ operator to an ondisc_matrix, metadata_ondisc_matrix, or multimodal_ondisc_matrix to subset the object. You can pass logical, character, or numeric vectors to [; character vectors are assumed to refer to feature IDs (for rows) and cell barcodes (for columns).

Usage

## S4 method for signature 'ondisc_matrix,missing,missing,missing'
x[i, j, drop]

## S4 method for signature 'ondisc_matrix,ANY,missing,missing'
x[i, j]

## S4 method for signature 'ondisc_matrix,missing,ANY,missing'
x[i, j]

## S4 method for signature 'ondisc_matrix,ANY,ANY,missing'
x[i, j]

## S4 method for signature 'metadata_ondisc_matrix,ANY,ANY,missing'
x[i, j, drop]

## S4 method for signature 'metadata_ondisc_matrix,ANY,missing,missing'
x[i, j, drop]

## S4 method for signature 'metadata_ondisc_matrix,missing,ANY,missing'
x[i, j, drop]

## S4 method for signature 'metadata_ondisc_matrix,missing,missing,missing'
x[i, j, drop]

## S4 method for signature 'multimodal_ondisc_matrix,missing,missing,missing'
x[i, j, drop]

## S4 method for signature 'multimodal_ondisc_matrix,missing,ANY,missing'
x[i, j, drop]

## S4 method for signature 'multimodal_ondisc_matrix,ANY,ANY,ANY'
x[i, j, drop]

Arguments

x	an ondisc_matrix, metadata_ondisc_matrix, or multimodal_ondisc_matrix object.
i	a vector (numeric, logical, or character) indicating features to keep.
j	a vector (numeric, logical, or character) indicating cells to keep.
drop	not used

Details

You can subset an ondisc_matrix and a metadata_ondisc_matrix by cell and/or feature. You can subset a multimodal_ondisc_matrix by cell only (because the features differ across modalities).

Value

An appropriately subset object of the same class as x.

Examples

# NOTE: You must create the RDS files "expressions.rds" and
# "perturbations.rds" to run this example. Navigate to the help file of
# "create_ondisc_matrix_from_mtx" (via ?create_ondisc_matrix_from_mtx),
# and execute both code blocks.

# subset an ondisc_matrix
h5_fp <- paste0(tempdir(), "/expressions.h5")
if (file.exists(h5_fp)) {
odm <- ondisc_matrix(h5_file = h5_fp)
# keep cells 100-110
x <- odm[,100:110]
# keep all cells except 50, 100, 150
x <- odm[,-c(50, 100, 150)]
# keep genes ENSG00000188305, ENSG00000257284, and ENSG00000251655:
x <- odm[c("ENSG00000188305", "ENSG00000257284", "ENSG00000251655"),]
# keep the cells CTTAGGACACTGGCGT-1 and AAAGGATTCACATCAG-1:
x <- odm[,c("CTTAGGACACTGGCGT-1", "AAAGGATTCACATCAG-1")]
# keep all genes except ENSG00000188305 and ENSG00000257284
x <- odm[!(get_feature_ids(odm) %in% c("ENSG00000188305", "ENSG00000257284")),]
}

# subset a metadata_ondic_matrix
expressions_fp <- paste0(tempdir(), "/expressions.rds")
if (file.exists(expressions_fp)) {
expressions <- readRDS(expressions_fp)
# keep cells 100-110
x <- expressions[,100:110]
# keep genes ENSG00000188305, ENSG00000257284, and ENSG00000251655
x <- expressions[c("ENSG00000188305", "ENSG00000257284", "ENSG00000251655"),]
}

# subset a multimodal ondisc_matrix
expression_fp <- paste0(tempdir(), "/expressions.rds")
perturbations_fp <- paste0(tempdir(), "/perturbations.rds")
if (file.exists(expression_fp) && file.exists(perturbations_fp)) {
    expressions <- readRDS(expression_fp)
    perturbations <- readRDS(expression_fp)
    crispr_experiment <- multimodal_ondisc_matrix(list(expressions = expressions,
    perturbations = perturbations))
    # Keep all cells except 10,100, and 105.
    x <- crispr_experiment[,-c(10,100,105)]
    # Keep the first 5 cells
    x <- crispr_experiment[,1:5]
}

ondisc documentation built on March 5, 2021, 5:07 p.m.