Run_gene_precomputation_at_scale_round_1

This function runs the first round of gene precomputations. In particular, it computes the raw dispersion estimate and log geometric mean of each gene. It saves the results in the gene_precomp directory.

run_gene_precomputation_at_scale_round_1(
  pod_id,
  gene_precomp_dir,
  gene_matrix,
  covariate_matrix,
  regularization_amount,
  log_dir
)

Arguments

pod_id

pod id

gene_precomp_dir

location of the gene precomputation directory

gene_matrix

a gene-by-cell expression matrix; the rows (i.e., gene IDs) and columns (i.e., cell barcodes) should be named

covariate_matrix

the cell-specific matrix of technical factors, ideally containing the following covariates: log-transformed gene library size (numeric), log-transformed gRNA library size (numeric), percent mitochondrial reads (numeric), and batch (factor). The rows (i.e., cell barcodes) should be named

regularization_amount

non-negative number specifying the amount of regularization to apply to the negative binomial dispersion parameter estimates

log_dir

directory in which to sink the log file