Run the complete microbial dose-response analysis

Standardizes the input, evaluates its trend, fits the requested models, compares their fit, and optionally bootstraps uncertainty.

Usage

analyze_dose_response(
  data,
  models = c("exponential", "beta_poisson"),
  bootstrap_times = 10000L,
  exact_bootstrap_times = 10000L,
  resample = c("observed", "fitted"),
  seed = NULL,
  backend = c("auto", "sequential", "mirai"),
  compute = NULL,
  workers = NULL,
  check_trend = TRUE,
  alpha = 0.05
)

Arguments

data

Grouped dose-response data accepted by as_dose_response().

models

Character vector of models to fit and compare, a subset of "exponential", "beta_poisson", and "exact_beta_poisson". Defaults to the exponential and approximate beta-Poisson models; add "exact_beta_poisson" to include the exact model.

bootstrap_times

Number of bootstrap replicates for the exponential and approximate beta-Poisson models. Use zero to skip their bootstrapping.

exact_bootstrap_times

Number of bootstrap replicates for the exact beta-Poisson model, which is roughly an order of magnitude slower to fit. Only used when "exact_beta_poisson" is in models; use zero to fit and compare the exact model without bootstrapping it.

resample

Bootstrap method passed to bootstrap_dose_response().

seed

Optional integer random seed.

backend

Bootstrap execution backend passed to bootstrap_dose_response().

compute

Optional mirai compute profile name.

workers

Number of temporary local mirai daemons, passed to bootstrap_dose_response().

check_trend

Warn when an increasing trend is not detected.

alpha

Significance level for trend and goodness-of-fit tests.

Value

A qdr_analysis object containing standardized data, trend test, model fits, diagnostics, parsed assessment, comparison, and bootstrap results.