loom: A class for connections loom files
In mojaveazure/loomR: An R interface for loom files

Description Usage Format Fields Methods See Also

A class for connections loom files

1	lfile <- loomR::connect(filename = 'myfile.loom')

An R6::R6Class object

version: Version of loomR object was created under
shape: Shape of /matrix in genes (columns) by cells (rows)
chunksize: Chunks set for this dataset in columns (cells) by rows (genes)
matrix: The main data matrix, stored as columns (cells) by rows (genes)
layers: Additional data matricies, the same shape as /matrix
col.attrs: Extra information about cells
row.attrs: Extra information about genes

add.graph(a, b, w, name, MARGIN, overwrite), add.graph.matrix(mat, name, MARGIN, overwrite)

Add a graph to the loom object; can add either in coorindate format (add.graph) or matrix format (add.graph.matrix). Stores graph in coordinate format as [row, col]_graphs/name/a (row indices), [row, col]_graphs/name/b (column indices), and [row, col]_graphs/name/w (values)

a: Integer vector of row indices for graph, must be the same lengths as b and w
b: Integer vector of column indices for graph, must be the same lengths as a and w
w: Numeric vector of values for graph, must be the same lengths as a and b
mat: Graph provided as a matrix (sparse or dense) or data.frame
name: Name to store graph, will end up being col_graphs/name or row_graphs/name, depending on MARGIN
MARGIN: Store the graph in row_graphs (1) or col_graphs (2), defaults to 2
overwrite: Can overwrite existing graph?

add.layer(layer, chunk.size, overwrite)

Add a data layer to this loom file, must be the same dimensions as /matrix

layer: A named list of matrices to be added as layers
chunk.size: Number of rows from each layer to stream at once, defaults to 1000
overwrite: If a layer already exists, overwrite with new data, defaults to FALSE

add.attribute(attribute, MARGIN, overwrite)

Add extra information to this loom file.

attribute: A named list where the first dimmension of each element as long as one dimension of /matrix
MARGIN: Either 1 for genes or 2 for cells
overwrite: Can overwrite existing attributes?

add.row.attribute(attribute), add.col.attribute(attribute)

Add row or column attributes

get.attribute.df(MARGIN, attribute.names, row.names, col.names)

Get a group of row or column attributes as a data frame, will only return attributes that have one dimension

MARGIN: Either '1' or '2' to get row- or column-attributes, respectively
attribute.names: A vector of attribute dataset basenames
row.names: Basename of the rownames dataset
col.names: Basename of the colnames dataset

get.graph(name, MARGIN)

Get a graph as a sparse matrix

name: Name of the graph, can be either the basename or full name of the grpah
MARGIN: Load the graph from row_graphs (1) or col_graphs (2), defaults to 2. Ignored if full path to graph is passed to name

batch.scan(chunk.size, MARGIN, index.use, dataset.use, force.reset), batch.next(return.data)

Scan a dataset in the loom file from index.use[1] to index.use[2], iterating by chunk.size.

chunk.size: Size to chunk MARGIN by, defaults to self$chunksize
MARGIN: Iterate over genes (1) or cells (2), defaults to 2
index.use: Which specific values of dataset.use to use, defaults to 1:self$shape[MARGIN] (all values)
dataset.use: Name of dataset to use, can be the name, not group/name, unless the name is present in multiple groups
force.reset: Force a reset of the internal iterator
return.data: Return data for a given chunk, if FALSE, returns the indices across MARGIN for said chunk

apply(name, FUN, MARGIN, chunk.size, dataset.use, overwrite, display.progress, ...)

Apply a function over a dataset within the loom file, stores the results in the loom file. Will not make multidimensional attributes.

name: Full name ('group/name') of dataset to store results to
FUN: Function to apply
MARGIN: Iterate over genes (1) or cells (2), defaults to 2
index.use: Which specific values of dataset.use to use, defaults to 1:self$shape[MARGIN] (all values)
chunk.size: Size to chunk MARGIN by, defaults to self$chunksize
dataset.use: Name of dataset to use
overwrite: Overite name if already exists
display.progress: Display progress
...: Extra parameters to pass to FUN

map(FUN, MARGIN, chunk.size, index.use, dataset.use, display.progress, expected, ...)

Map a function onto a dataset within the loom file, returns the result into R.

FUN
MARGIN: Iterate over genes (1) or cells (2), defaults to 2
chunk.size: Size to chunk MARGIN by, defaults to self$chunksize
index.use: Which specific values of dataset.use to use, defaults to 1:self$shape[MARGIN] (all values)
dataset.use: Name of dataset to use
display.progress: Display progress
...: Extra parameters to pass to FUN

add.cells(matrix.data, attributes.data = NULL, layers.data = NULL, display.progress = TRUE)

Add cells to a loom file.

matrix.data: A list of m2 cells where each entry is a vector of length n (num genes, self$shape[1])
attributes.data: A list where each entry is named for one of the datasets in self[['col_attrs']]; each entry is a vector of length m2.
layers.data: A list where each entry is named for one of the datasets in self[['layers']]; each entry is an n-by-m2 matrix where n is the number of genes in this loom file and m2 is the number of cells being added.
display.progress: Display progress

add.loom(other, other.key, self.key, ...)