Run fastTopics and Generate Topic Annotations

This function runs the fastTopics model on the provided count matrix, performs differential expression analysis, computes p-values, and returns the factor matrices along with differential expression results.

run_fastTopics(
  count_matrix,
  nTopics = 10,
  n_s = 1000,
  n_c = 1,
  baseline_method = "constant",
  bl_celltype_cells = NULL,
  bl_celltype_peak_file = NULL,
  fdr_cutoff = 0.05,
  outdir,
  genome = "hg19",
  ...
)

Arguments

count_matrix

A numeric matrix or sparse matrix of counts with cells as rows and peaks as columns.

nTopics

An integer specifying the number of topics to fit. Default is 10.

n_s

Number of samples for the DE analysis (default: 1000).

n_c

Number of cores for parallel processing in DE analysis (default: 1).

...

Additional arguments passed to fastTopics::fit_topic_model().

Value

A list containing the factor matrices Fmat and Lmat, differential expression results de_res, p-values matrix p_jk and caculated baseline baseline.

Details

The function ensures that the count matrix has cells as rows and peaks as columns. It then runs the fastTopics model, performs differential expression analysis, computes local false discovery rates (lfdr), and calculates p-values for each peak-topic pair. If locfdr::locfdr fails, it binarizes Fmat based on a threshold and computes p_jk.

Examples

if (FALSE) { # \dontrun{
# Assuming 'counts' is your count matrix
results <- run_fastTopics(counts, nTopics = 5)
} # }