invitro_mc {httk}

Monte Carlo for in vitro toxicokinetic parameters including uncertainty and variability.

Description

Given a CAS in the HTTK data set, a virtual population from HTTK-Pop, some user specifications on the assumed distributions of Funbound.plasma and Clint, draw "individual" values of Funbound.plasma and Clint from those distributions. The methodology for this function was developed and described by Wambaugh et al. (2019) (doi:10.1093/toxsci/kfz205).

Usage

invitro_mc(
  parameters.dt = NULL,
  samples,
  fup.meas.mc = TRUE,
  fup.pop.mc = TRUE,
  clint.meas.mc = TRUE,
  clint.pop.mc = TRUE,
  fup.meas.cv = 0.4,
  clint.meas.cv = 0.3,
  fup.pop.cv = 0.3,
  clint.pop.cv = 0.3,
  caco2.meas.sd = 0.3,
  caco2.pop.sd = 0.3,
  Caco2.Fgut = TRUE,
  Caco2.Fabs = TRUE,
  keepit100 = FALSE,
  poormetab = TRUE,
  fup.lod = 0.01,
  fup.censored.dist = FALSE,
  adjusted.Funbound.plasma = TRUE,
  adjusted.Clint = TRUE,
  clint.pvalue.threshold = 0.05,
  minimum.Funbound.plasma = 1e-04
)

Arguments

parameters.dt

A data table of physiological and chemical-specific parameters

samples

The number of samples to draw.

fup.meas.mc

Logical – should we perform measurment (uncertainty) Monte Carlo for Funbound.plasma values (Default TRUE). If FALSE, the user may choose to provide columns for "unadjusted.Funbound.plasma" or "fup.mean" from their own methods.

fup.pop.mc

Logical – should we perform population (variability) Monte Carlo for Funbound.plasma values (Default TRUE)

clint.meas.mc

Logical – should we perform measurment (uncertainty) Monte Carlo for Clint values (Default TRUE)

clint.pop.mc

Logical – should we perform population (variability) Monte Carlo for Clint values (Default TRUE)

fup.meas.cv

Coefficient of variation of distribution of measured Funbound.plasma values.

clint.meas.cv

Coefficient of variation of distribution of measured Clint values.

fup.pop.cv

Coefficient of variation of distribution of population Funbound.plasma values.

clint.pop.cv

Coefficient of variation of distribution of population Clint values.

caco2.meas.sd

Standard deviation of the measured oral absorption - numeric value (Default 0.3).

caco2.pop.sd

Standard deviation of the population level oral absorption - numeric value (Default 0.3).

Caco2.Fgut

= TRUE uses Caco2.Pab to calculate fgut.oral, otherwise fgut.oral = Fgut.

Caco2.Fabs

= TRUE uses Caco2.Pab to calculate fabs.oral, otherwise fabs.oral = Fabs.

keepit100

= TRUE overwrites Fabs and Fgut with 1 (i.e. 100 percent) regardless of other settings.

poormetab

Logical. Whether to include poor metabolizers in the Clint distribution or not.

fup.lod

The average limit of detection for Funbound.plasma, below which distribution will be censored if fup.censored.dist is TRUE. Default 0.01.

fup.censored.dist

Logical. Whether to draw Funbound.plasma from a censored distribution or not.

adjusted.Funbound.plasma

Uses the Pearce et al. (2017) lipid binding adjustment for Funbound.plasma when set to TRUE (Default).

adjusted.Clint

Uses Kilford et al. (2008) hepatocyte incubation binding adjustment for Clint when set to TRUE (Default).

clint.pvalue.threshold

Hepatic clearance for chemicals where the in vitro clearance assay result has a p-values greater than the threshold are set to zero.

minimum.Funbound.plasma

Monte Carlo draws less than this value are set equal to this value (default is 0.0001 – half the lowest measured Fup in our dataset).

parameters

A list of chemical-specific model parameters containing at least Funbound.plasma, Clint, and Fhep.assay.correction.

Details

The Monte Carlo methods used here were recently updated and described by Breen et al. (2022).

Value

A data.table with three columns: Funbound.plasma and Clint, containing the sampled values, and Fhep.assay.correction, containing the value for fraction unbound in hepatocyte assay.

Author(s)

Caroline Ring and John Wambaugh

References

Breen M, Wambaugh JF, Bernstein A, Sfeir M, Ring CL (2022). “Simulating toxicokinetic variability to identify susceptible and highly exposed populations.” Journal of Exposure Science & Environmental Epidemiology, 32(6), 855–863.

Kilford PJ, Gertz M, Houston JB, Galetin A (2008). “Hepatocellular binding of drugs: correction for unbound fraction in hepatocyte incubations using microsomal binding or drug lipophilicity data.” Drug Metabolism and Disposition, 36(7), 1194–1197.

Pearce RG, Setzer RW, Davis JL, Wambaugh JF (2017). “Evaluation and calibration of high-throughput predictions of chemical distribution to tissues.” Journal of pharmacokinetics and pharmacodynamics, 44, 549–565.

Wambaugh JF, Wetmore BA, Ring CL, Nicolas CI, Pearce RG, Honda GS, Dinallo R, Angus D, Gilbert J, Sierra T, others (2019). “Assessing toxicokinetic uncertainty and variability in risk prioritization.” Toxicological Sciences, 172(2), 235–251.

Examples

#Simply generate a virtual population of 100 individuals,
#using the direct-resampling method
set.seed(42)

# Pull mean vchemical=specific values:
chem.props <- parameterize_pbtk(chem.name="bisphenolaf")

# Convert to data.table with one row per sample:
parameters.dt <- monte_carlo(chem.props,samples=100)

# Use httk-pop to generate a population:
pop <- httkpop_generate(method='direct resampling', nsamp=100)

# Overwrite parameters specified by httk-pop:
parameters.dt[,names(pop):=pop]
# Vary in vitro parameters:
parameters.dt <- invitro_mc(parameters.dt,samples=100)