Plot prior kinetic trajectories

Draw population-level kinetic curves from the configured priors. The band shows the pointwise 95% prior interval and the line shows the pointwise prior median.

Construct and inspect the priors used by fit_epikinetics(). Population priors are Normal distributions parameterised by c(mean, sd). Positive kinetic quantities use Normal priors truncated at zero. Their prior means must be non-negative so that the inverse-CDF parameterisation remains numerically stable.

Usage

# S3 method for class 'epikinetics_priors'
plot(
  x,
  ...,
  times = 0:150,
  ndraws = 1000,
  probs = c(0.025, 0.975),
  reference_value = 1,
  scale = c("model", "response")
)

epikinetics_priors(
  baseline = c(mean = 6, sd = 2),
  time_to_peak = c(mean = 10, sd = 5),
  waning_duration = c(mean = 50, sd = 20),
  boost_rate = c(mean = 0.25, sd = 0.1),
  early_waning_rate = c(mean = 0.02, sd = 0.01),
  late_waning_rate = c(mean = 0.002, sd = 0.005),
  participant_sd = c(baseline = 0.75, time_to_peak = 0.35, waning_duration = 0.5,
    boost_rate = 0.35, early_waning_rate = 0.5, late_waning_rate = 0.75),
  covariate_sd = c(baseline = 0.5, time_to_peak = 0.35, waning_duration = 0.5, boost_rate
    = 0.35, early_waning_rate = 0.5, late_waning_rate = 0.75),
  observation_sd = 1
)

Arguments

x

An epikinetics_priors object.

...

Reserved for compatibility with the plot() generic.

times

Non-negative prediction times.

ndraws

Number of prior trajectories to simulate.

probs

Lower and upper probabilities for the pointwise prior interval.

reference_value

Positive reference value used to convert model-scale values to the response scale when scale = "response".

scale

Output scale: "model" for log2-relative values or "response" for natural measurement values.

baseline, time_to_peak, waning_duration, boost_rate, early_waning_rate, late_waning_rate

Numeric length-two vectors giving the population prior mean and standard deviation.

participant_sd

Named positive vector giving half-Normal scales for participant-level standard deviations.

covariate_sd

Named positive vector giving Normal standard deviations for regression coefficients.

observation_sd

Positive half-Normal scale for observation error.

Value

A ggplot2::ggplot() object.

An object of class epikinetics_priors.

Details

Baseline is expressed on the model's log2 scale. Time parameters are in the same units as the input time columns (normally days). Rates are log2 units per time unit. Participant standard deviations act additively for baseline and on the log scale for positive time/rate parameters. Covariate priors use the same convention.

Examples

priors <- epikinetics_priors(
  time_to_peak = c(mean = 12, sd = 4),
  early_waning_rate = c(mean = 0.025, sd = 0.01)
)
priors
#>          parameter population_mean population_sd participant_sd_scale
#>           baseline           6.000         2e+00                 0.75
#>       time_to_peak          12.000         4e+00                 0.35
#>    waning_duration          50.000         2e+01                 0.50
#>         boost_rate           0.250         1e-01                 0.35
#>  early_waning_rate           0.025         1e-02                 0.50
#>   late_waning_rate           0.002         5e-03                 0.75
#>  covariate_sd
#>          0.50
#>          0.35
#>          0.50
#>          0.35
#>          0.50
#>          0.75
#> 
#> Observation SD half-Normal scale: 1