Run EM algorithm for transcript quantification

Main EM algorithm that estimates transcript abundances from short-read and optionally long-read RNA-seq data. This is an internal function called by the main quantification API.

run_em_algorithm(
  index,
  sr_counts,
  lr_counts = NULL,
  positional_weights = NULL,
  bias_type = "3p",
  post_weight_umi_correction = FALSE,
  max_iter = 100L,
  tolerance = 1e-04,
  use_prior = FALSE,
  use_regularization = TRUE,
  prior_predictions = NULL,
  covariate_matrix = NULL,
  prior_start = 25L,
  convergence_start = 25L,
  verbose = TRUE
)

Arguments

index

An mpaqt_index object

sr_counts

Short-read EC counts (named numeric vector)

lr_counts

Long-read transcript counts (optional)

positional_weights

Positional bias weights (optional)

bias_type

Type of positional bias ("3p" or "5p")

post_weight_umi_correction

Whether to normalize short-read probabilities after applying positional weights

max_iter

Maximum number of iterations

tolerance

Convergence tolerance

use_prior

Whether to use prior predictions

prior_predictions

Prior mean for each transcript (optional)

prior_start

Iteration to start using prior

convergence_start

Iteration to start checking convergence

verbose

Print progress

Value

List with quantification results