Automatic Generic Models Creation with ESQhtpbpk • ESQhtpbpk

Introduction

The framework allows automatic detection of required processes and administrations, and creates the corresponding generic models.

First, the package need to be loaded:

library(ESQhtpbpk)

Creating Studies

Next, the user needs to define simulation scenarios (i.e. studies) to be simulated. Each study encompasses every component needed for the simulation individual to be used, protocol and formulation for each compound administration and compound properties.

To simplify this task, various functions have been implemented in the framework easing the creation of all required parts of a study.

Additionally, the user can create their custom function to generate a list of all Study objects to be simulated, tailored to their specific data structure, as this part is highly dependent on it.

Example

The following example shows how to create a single study. This mechanism can be wrapped into a custom function to create a list of studies to be simulated using specific database as input for parameter value.

Compound Definition

First, the user needs to create a compound object with the Compound$new() function and assign a specific compound ID:

compound <- Compound$new(ID = "Acetaminophen")

The compound is initialized with the main properties that are needed for the simulation. Those can easily be modified with the setPropertyValue method.

compound$setPropertyValue("Molecular weight", 151.16, unit = "g/mol")
compound$setPropertyValue("Solubility", 4.15, unit = "mg/ml")
compound$setPropertyValue("Lipophilicity", 0.91)
compound$setPropertyValue("pKa value 0", 9.46)
compound$setPropertyValue("Compound type 0", "Acidic")
compound$setPropertyValue("Fraction unbound", 80, unit = "%")

Additionally, various processes (as available in PK-Sim) can be added to the compound object using the addProcessProperty method. For example the following code adds a liver plasma clearance process by defining its plasma clearance:

compound$addProcessProperty(
  processType = "Liver Plasma Clearance",
  propertyName = "Plasma clearance",
  parName = "Plasma clearance",
  dimension = "Flow per weight",
  value = 0.27, unit = "L/h/kg"
)

Additional properties for the same process can be added and their values adjusted if required in a similar fashion:

compound$addProcessProperty(
  processType = "Liver Plasma Clearance",
  propertyName = "Lipophilicity",
  parName = "Lipophilicity (experiment)",
  dimension = "Log Units",
  value = 0.91
)
compound$addProcessProperty(
  processType = "Liver Plasma Clearance",
  propertyName = "Fraction unbound",
  parName = "Fraction unbound (experiment)",
  dimension = "Fraction",
  value = 80,
  unit = "%"
)

The partition coefficient and cellular permeability methods can be set easily with:

compound$PartitionCoefficientMethod <- "PT"
compound$CellularPermeabilityMethod <- "PK-Sim"

The compound properties that have been defined can be easily visualized with:

compound
#> • Compound Properties:
#>   • Lipophilicity: 0.91 Log Units
#>   • Fraction unbound: 80 %
#>   • Plasma protein binding partner: Albumin
#>   • Is small molecule: 1
#>   • Molecular weight: 151.16 g/mol
#>   • Bromine count: 0
#>   • Chlorine count: 0
#>   • Fluorine count: 0
#>   • Iodine count: 0
#>   • pKa value 0: 9.46
#>   • Compound type 0: Acidic
#>   • pKa value 1: 0
#>   • Compound type 1: Neutral
#>   • pKa value 2: 0
#>   • Compound type 2: Neutral
#>   • Reference pH: 7
#>   • Solubility: 4.15 mg/ml
#> • Compound Processes:
#>   • Liver Plasma Clearance
#>     • Plasma clearance: 0.27 L/h/kg
#>     • Lipophilicity: 0.91 Log Units
#>     • Fraction unbound: 80 %
#> • Compound Methods:
#>   • Partition Coefficient Method: Poulin and Theil
#>   • Cellular Permeability Method: PK-Sim Standard

Administration definition

Administration protocol and formulation can be easily defined as well.

The protocol can be defined with either SimpleProtocol$new() method for a simple protocol or AdvancedProtocol$new() for a more complex protocol. Here we will use the SimpleProtocol$new() method:

# define protocol used in study and associate formulation used
protocol <- SimpleProtocol$new(
  route = "Oral",
  dosingInterval = "24",
  dose = 1,
  doseUnit = "mg/kg",
  startTime = 0,
  startTimeUnit = "h",
  endTime = 48,
  endTimeUnit = "h"
)        
#> Warning: No `WaterVolPerBW` provided, using default value of 3.5 ml/kg.
#> Warning: No `Formulation` provided, using default of dissolved.
#> Formulation can be changed with `protocolObject$setFormulation(formulation)`.

For oral administration the formulation needs to be defined as well. Helper functions are available for each type of formulation and can be set as shown here for a Weibull formulation:

formulation <- createWeibullFormulation(
  name = "Weibull",
  lagTime = 30,
  lagTimeUnit = "min",
  dissolutionTime50 = 60,
  dissolutionTime50Unit = "min"
)

Then the formulation needs to be associated with the protocol defined previously using the setFormulation method.

protocol$setFormulation(formulation = formulation)

Similarly to compounds, administration protocol can be easily visualized:

protocol
#>   • Route: Oral
#>   • Dose: 1 mg/kg
#>   • Dose Interval: Once each 24 hours
#>   • Start Time: 0 h
#>   • End Time: 48 h
#>   • Volume of water per body weight: 3.5 ml/kg
#>   • Formulation: Weibull

Now that the protocol has been defined, it can be associated with the compound using the setProtocol method:

compound$setProtocol(protocol)

Study definition

Finally, the study can be defined with the Study$new() method. The study needs to be given an unique ID:

study <- Study$new(ID = "Acetaminophen_PO_QD", compounds = list(compound), individual = "Human")

Similarly to the compound and protocol, the study can be easily visualized with:

study
#> <Study>
#> ID: Acetaminophen_PO_QD
#> Individual: Human
#> Compounds:
#> 
#> • Compound with protocol Protocol
#>   • Compound Properties:
#>     • Lipophilicity: 0.91 Log Units
#>     • Fraction unbound: 80 %
#>     • Plasma protein binding partner: Albumin
#>     • Is small molecule: 1
#>     • Molecular weight: 151.16 g/mol
#>     • Bromine count: 0
#>     • Chlorine count: 0
#>     • Fluorine count: 0
#>     • Iodine count: 0
#>     • pKa value 0: 9.46
#>     • Compound type 0: Acidic
#>     • pKa value 1: 0
#>     • Compound type 1: Neutral
#>     • pKa value 2: 0
#>     • Compound type 2: Neutral
#>     • Reference pH: 7
#>     • Solubility: 4.15 mg/ml
#>   • Compound Processes:
#>     • Liver Plasma Clearance
#>       • Plasma clearance: 0.27 L/h/kg
#>       • Lipophilicity: 0.91 Log Units
#>       • Fraction unbound: 80 %
#>   • Compound Methods:
#>     • Partition Coefficient Method: Poulin and Theil
#>     • Cellular Permeability Method: PK-Sim Standard
#>   • Protocol Properties:
#>       • Route: Oral
#>       • Dose: 1 mg/kg
#>       • Dose Interval: Once each 24 hours
#>       • Start Time: 0 h
#>       • End Time: 48 h
#>       • Volume of water per body weight: 3.5 ml/kg
#>       • Formulation: Weibull

Running the studies

Once the user created the list of studies required, simulations can be run the using the runPredictions() function. This function creates the required models, associated them with the corresponding studies and run the simulation. All simulation generic models files will be saved in the outputFolder within a time-stamped subfolder under GenericModels and all simulations results can be saved under the same time-stamped subfolder under SimulationResults by setting the saveResults parameter to TRUE. Additionally the user can save the PKML files for each study with the saveSimulation parameter.

During this step, the user should also define the output paths to be simulated with the outputSelections parameter, as well as the simulationResolution defined by c(start time (min), end time (min), resolution in pts/min) to cover the longest simulation needed with the wanted time resolution.

Additionally the user can also defined the number of cores to be used for the simulations and the number studies to be queued (which can be needed to reduce the memory usage).

myresult <- runPredictions(
    studies = list(study), 
    outputFolder = "HT-PBPK",
    numberOfCores = 5,
    outputSelections = c("Organism|PeripheralVenousBlood|**|Plasma (Peripheral Venous Blood)"),
    simulationResolution = c(0, 10 * 24 * 60, 1),
    saveResults = TRUE,
    saveSimulation = FALSE,
    queueSize = 200
  )

Current limitations

The current implementation does not allows:

user defined administration route
solubility table
formulations with dissolution tables
custom individual (only default individuals are supported)
gene expression use or definition
transport or enzymatic processes
food administration or other events
multiple solubility values at different pH