tests/testthat/test-mcmc-nnls.R

In shackett/simmer: Systematic Identification of Meaningful Metabolic Enzyme Regulation

test_that("fit_reaction_equations applies the MCMC-NNLS algorithm to one reaction equation", {

  options(stringsAsFactors = F)

  rMech_summary_table <- suppressMessages(readr::read_tsv(system.file("extdata", "reactionEqn_fitting", "rMech_summary_table.tsv", package = "simmer")))
  reactionForms <- suppressMessages(readr::read_rds(system.file("extdata", "reactionEqn_fitting", "reactionForms.Rds", package = "simmer")))
  
  expect_equal(colnames(rMech_summary_table), c("reaction", "rMech", "modelType"))
  expect_equal(rMech_summary_table$rMech, names(reactionForms))

  markov_pars <- list()
  markov_pars$sample_freq <- 20 #what fraction of markov samples are reported (this thinning of samples decreases sample autocorrelation)
  markov_pars$n_samples <- 200 #how many total markov samples are desired
  markov_pars$burn_in <- 0 #how many initial samples should be skipped

  fitted_reaction_equations <- fit_reaction_equations(reactionForms[[1]])

  expect_equal(names(fitted_reaction_equations), names(reactionForms[1]))
  expect_equal(names(fitted_reaction_equations[[1]]), c("kineticPars", "all_species", "kineticParPrior", "markovChain",
                                                        "logLikelihood", "occupancyEq", "metabolites", "enzymes", "flux",
                                                        "specSD", "specCorr"))
  expect_equal(nrow(fitted_reaction_equations[[1]]$markovChain), markov_pars$n_samples)
})