toxCRM {RFPM}

Concentration-Response Model, 4-parameter log-logistic curve

Description

Generate dummy toxicity data with strong concentration-response

Usage

toxCRM(x, max = 1, min = 0, steep, mid, eMean = 0, eSD = 0, seed = NULL)

Arguments

x

numeric vector of chemical concentrations (univariate)

max

numeric value, asymptotic maximum of model (default = 1)

min

numeric value, asymptotic minimum of model (default = 0)

steep

numeric value, steepness factor for model slope

mid

numeric value, inflection point concentration for CRM slope

eMean

numeric value, mean of random-normal error to add to dummy toxicity data (default = 0)

eSD

numeric value, standard deviation of random-normal error to add to dummy toxicity data (default = 0)

seed

numeric value, random seed to set for repeatable random error generating (default = NULL, i.e., no seed)

Details

toxCRM generates dummy toxicity data representing an idealized condition. This approach was used by the authors to simulate data and test FPM sensitivity and baseline variability. The user must specify all coefficients of the model (though default min/max values are provided) and may add some amount of random error, if desired. Random error adds noise to the dummy toxicity daata, increasing the uncertainty of toxicity predictions using floating percentile model benchmarks. The perfect, lowNoise, and highNoise datasets included in RFPM were all generated using toxCRM with increasing levels of eSD.

Value

numeric vector

See Also

perfect, lowNoise, highNoise

Examples

concentration = h.northport$Cu
toxVals <- toxCRM(concentration, 1, 0, 0.5, 
    median(concentration), 0, 0)
plot(concentration, toxVals, log="x", main="Perfect estimate")

toxVals_withNoise <- toxCRM(x = concentration,
    1, 0, 0.5, median(concentration), 0, 0.1, seed = 1)
plot(concentration, toxVals_withNoise, log="x", main="Noisy estimate")