Smooth expression along chromosomes (windowed running mean) with optional chunking/parallelism.

Usage

smooth_expr(
  obj,
  window_size = 101,
  step = 50,
  exclude_chromosomes = c("chrX", "chrY", "chrM"),
  smooth_with_ends = FALSE,
  use_chunk = TRUE,
  chunk_size = 5000,
  parallel = FALSE,
  n_cores = NULL,
  max_cores = 4
)

Arguments

obj: highSpaClone object.
window_size: Integer. Size of the running-mean window (in genes) used within each chromosome (default: 101).
step: Integer. Number of genes in a bin (default: 50).
exclude_chromosomes: Chromosomes that are excluded from analysis (default: c('chrX', 'chrY', 'chrM')).
smooth_with_ends: Logical. Whether to apply an end-aware smoother (default: FALSE).
use_chunk: Logical. If TRUE, split cells into chunks of size chunk_size
chunk_size: Integer. Number of cells per chunk when use_chunk=TRUE.
parallel: Logical. If TRUE, process chunks in parallel.
n_cores: Integer or NULL. Number of CPU cores to use when parallel=TRUE. If NULL, uses parallel::detectCores()-1. The effective cores are capped by max_cores and the number of chunks (default NULL).
max_cores: Integer. Upper bound on cores even if more are available (default 4).

Value

Return highSpaClone object containing smoothed gene expression data