Export bigwig files for groups of cells

Export Tn5 insertion coverage tracks as bigwig files, one per group of cells. Fragments are split by group, the genome is tiled, and the number of fragment ends (Tn5 insertion sites) falling in each tile is counted and (optionally) normalized before writing the bigwig files. Note that each fragment contributes two insertion events (one at each end), so the tracks represent insertion-site coverage rather than fragment pileup.

Usage

ExportBigwig(
  object,
  assay = NULL,
  group.by = NULL,
  idents = NULL,
  normMethod = "RC",
  tileSize = 100,
  minCells = 5,
  cutoff = NULL,
  chromosome = NULL,
  seqlengths = NULL,
  outdir = getwd(),
  temp.dir = tempdir(),
  cleanup = TRUE,
  verbose = TRUE
)

Arguments

object

A Seurat object

assay

Name of assay to use

group.by

The metadata variable used to group the cells

idents

Identities to include (defined by group.by parameter)

normMethod

Normalization method for the bigwig files. Default RC. RC will divide the number of insertions in a tile by the total number of fragments in the group. A scaling factor of 10^4 will be applied. ncells will divide the number of insertions in a tile by the number of cells in the group. none (or NULL) will apply no normalization. The name of a metadata column can also be passed, in which case insertions will be divided by the sum of that column over the cells in the group, with a scaling factor of 10^4 applied.

tileSize

The size of the tiles in the bigwig file

minCells

The minimum number of cells in a group for it to be exported. Groups with fewer than minCells cells are skipped.

cutoff

The maximum number of insertions for a single cell in a given genomic tile. Counts above this value are capped before summing across cells. Note that cells are identified by their fragment-file barcode, so when an assay contains multiple fragment files this cap is shared between any cells that have the same barcode in different files.

chromosome

A vector of chromosomes to export. If NULL, use all chromosomes present in seqlengths.

seqlengths

Chromosome lengths used to define the genomic tiles. Can be a named numeric vector of chromosome lengths, or any object with a seqlengths method such as a BSgenome or Seqinfo::Seqinfo() object. If NULL, the chromosome lengths stored in the object are used; note that these are frequently unset, in which case seqlengths must be supplied.

outdir

Directory to write the output bigwig files. Defaults to the current working directory.

temp.dir

Directory to write the intermediate per-group bed files. Defaults to a temporary directory (base::tempdir()).

cleanup

Remove the intermediate per-group bed files after writing the bigwig files. Default TRUE.

verbose

Display messages

Value

Returns a list of paths to the bigwig files that were created

Examples

if (FALSE) { # \dontrun{
# chromosome lengths can be supplied as a BSgenome object, a Seqinfo, or a
# named numeric vector
ExportBigwig(
  object,
  group.by = "celltype",
  seqlengths = BSgenome.Hsapiens.UCSC.hg38::BSgenome.Hsapiens.UCSC.hg38
)
} # }