Run Lilace — lilace_fit_model • lilace

Runs Lilace on given data. Lilace will run on $normalized_data if it exists, otherwise it will use $data.

Usage

lilace_fit_model(
  lilace_obj,
  output_dir,
  control_label = "synonymous",
  control_correction = TRUE,
  use_positions = TRUE,
  pseudocount = TRUE,
  seed = NULL,
  min_total_counts = 15,
  n_parallel_chains = 4
)

Arguments

lilace_obj: initialized lilace object
output_dir: output directory to write scores and sampling logs to
control_label: label from $data$type column to use as negative controls to score against
control_correction: boolean for whether to use negative control scores as bias correction
use_positions: boolean for whether to use position hierarchy to improve estimation
pseudocount: boolean for whether to add pseudocount (+1 to all counts) for fitting model
seed: random seed for sampling process to get exactly reproducible results. A NULL value indicates no fixed seed.
min_total_counts: minimum total counts for a variant–anything less will be filtered out
n_parallel_chains: number of chains to run in parallel

Value

lilace object with $scores and $fitted_data entries

Examples

if (FALSE) { # \dontrun{
lilace_obj <- lilace_fit_model(lilace_obj, output_dir, control_label="synonymous",
                               control_correction=T, use_positions=T, pseudocount=T)
} # }