This function replicates counts from a real-world dataset.
sccomp_predict(
fit,
formula_composition = NULL,
new_data = NULL,
number_of_draws = 500,
mcmc_seed = sample(1e+05, 1),
summary_instead_of_draws = TRUE
)The result of sccomp_estimate.
A formula. The formula describing the model for differential abundance, for example ~treatment. This formula can be a sub-formula of your estimated model; in this case all other factor will be factored out.
A sample-wise data frame including the column that represent the factors in your formula. If you want to predict proportions for 10 samples, there should be 10 rows. T
An integer. How may copies of the data you want to draw from the model joint posterior distribution.
An integer. Used for Markov-chain Monte Carlo reproducibility. By default a random number is sampled from 1 to 999999. This itself can be controlled by set.seed()
Return the summary values (i.e. mean and quantiles) of the predicted proportions, or return single draws. Single draws can be helful to better analyse the uncertainty of the prediction.
A tibble (tbl) with the following columns:
cell_group - A character column representing the cell group being tested.
sample - A factor column representing the sample name for which the predictions are made.
proportion_mean - A numeric column representing the predicted mean proportions from the model.
proportion_lower - A numeric column representing the lower bound (2.5%) of the 95% credible interval for the predicted proportions.
proportion_upper - A numeric column representing the upper bound (97.5%) of the 95% credible interval for the predicted proportions.
message("Use the following example after having installed install.packages(\"cmdstanr\", repos = c(\"https://stan-dev.r-universe.dev/\", getOption(\"repos\")))")
#> Use the following example after having installed install.packages("cmdstanr", repos = c("https://stan-dev.r-universe.dev/", getOption("repos")))
# \donttest{
if (instantiate::stan_cmdstan_exists() && .Platform$OS.type == "unix") {
data("counts_obj")
sccomp_estimate(
counts_obj,
~ type, ~1, sample, cell_group, count,
cores = 1
) |>
sccomp_predict()
}
# }