Simulate longitudinal case data with known cross-biomarker covariance

End-to-end simulator for validating Model 2a. It draws correlated subject-level parameters with sim_params_2a(), evaluates the Chapter 1 two-phase curve ab() at a set of visit times for two biomarkers, adds log-normal measurement noise, and returns a long data.frame in the serocalculator case-data layout that prep_data() / run_mod_2a() accept.

Decomposed on purpose: the statistical truth lives in sim_params_2a(), the curve in ab(), and this function only handles visit times, noise, and reshaping.

Usage

sim_case_data_2a(
  n,
  mu_g,
  mu_a,
  sigma_g,
  sigma_a,
  c_vec,
  visit_times = c(0, 7, 14, 28, 56, 90, 140, 200),
  noise_sd = 0.2,
  biomarkers = c("HlyE_IgG", "HlyE_IgA"),
  seed = NULL
)

Arguments

n: integer number of subjects.
mu_g, mu_a, sigma_g, sigma_a, c_vec: Model 2a truth, passed to sim_params_2a().
visit_times: numeric vector of sampling times (days) shared by all subjects. Default c(0, 7, 14, 28, 56, 90, 140, 200).
noise_sd: Residual SD on the log scale. Default 0.2.
biomarkers: Length-2 character biomarker labels. Default c("HlyE_IgG", "HlyE_IgA").
seed: Optional RNG seed.

Value

A list with data (the long case-data data.frame) and truth (the sim_params_2a() output, including rho).