pbpk-lite

pbpk-lite is a lightweight Python package for basic physiologically based pharmacokinetic (PBPK) modeling.

It provides a simple programmatic interface for defining substance and metabolite properties, patient physiology, elimination kinetics, and solving the resulting ODE system.

The implementation uses nanograms (ng) for doses and amounts, milliliters (mL) for volumes, milliliters per minute (mL/min) for blood flows and clearances, nanograms per milliliter (ng/mL) for concentrations, and minutes for time. The simulation uses minutes internally, while the plotting helpers can display the x-axis in minutes, hours, or days via the optional time_unit argument.

Features

• Substance and metabolite partition coefficient calculation using logP and fraction unbound
• Patient blood flows and tissue volumes derived from body weight
• Linear liver and kidney elimination pathways
• ODE solution via scipy.integrate.solve_ivp
• Plotting helpers for whole-model, venous-blood, and selected-compartment concentration profiles
• Support for different administration routes, including intravenous, intra-arterial, and inhalation dosing

The model uses 16 compartments for each substance. Simulation results contain one block of 16 concentration rows per substance; administered doses enter the parent substance block.

Installation

Install from PyPI:

pip install pbpk-lite

Quick Start

from pbpk_lite import model

m = model()
m.set_substance_and_metabolites(
	log_ps=(6.97, 5.33, 5.24),
	fus=(0.0022448, 0.01209, 0.101),
	mms=(314.469, 330.468, 344.451),
)
m.set_patient(bw=70)
m.set_elimination(cl_ls=(247, 998, 8.94), cl_ks=(0, 0, 1.48))

dose = 25e6  # ng

doses = [dose]
times = [0, 60*24]

t, c = m.simulate(doses, times, route_of_administration='iv')

m.graph_whole('concentrations.png')
m.graph_venous('venous.png', limit_of_detection=0.15, time_unit='hours')
m.graph_compartments(['liver', 'kidney'], 'selected.png', time_unit='days')

# The underlying simulation still uses minutes internally; only the displayed axis changes.

Route of Administration

The simulate() method accepts a route_of_administration argument to control where each dose is introduced into the model.

Doses can only be administered to the parent substance. Metabolites are produced by the model and cannot be dosed directly.

Supported values are:

• iv: intravenous dosing into the venous blood compartment (default)
• ia: intra-arterial dosing into the arterial blood compartment
• inh: inhalation dosing into the lung compartment

Example:

m.simulate([dose], [0, 24*60], route_of_administration='inh')

Dosing Schedule

The simulate() method expects:

• doses: array-like of administered doses
• times: array-like of dosing times plus a final endpoint

Important: times must have exactly one more element than doses and must be strictly increasing. Each dose at index i is administered at times[i], and the final value in times is the last observation or endpoint. Times are interpreted in minutes, so dosing schedules and simulation endpoints should be provided in minutes.

Example with one dose:

# one dose at time 0, observation at 24 hours
doses = [dose]
times = [0, 60*24]

Example with two doses:

# two identical doses spaced one hour apart, with a final observation at 24 hours
doses = [dose, dose]
times = [0, 60, 60*24]

API Summary

pbpk_lite.model

set_substance_and_metabolites(log_ps, fus, mms)

Set the physicochemical properties and molecular masses for the substance and its metabolites. All three sequences should contain one value for each substance.

• log_ps: log octanol-water partition coefficients
• fus: fractions unbound in blood
• mms: molecular masses; the first entry is the parent substance

set_patient(bw)

Set patient physiological parameters using body weight in kilograms. Blood flows and tissue volumes are recalculated for the new body weight.

set_elimination(cl_ls, cl_ks)

Set linear clearance from the liver and kidney compartments for each substance.

• cl_ls: liver clearance values
• cl_ks: kidney clearance values

simulate(doses, times, route_of_administration='iv')

Simulate the PBPK model and return time points t and compartment concentrations c.

• route_of_administration: administration route used for each dose ('iv', 'ia', or 'inh')
• administered doses enter the parent substance block; returned concentrations have 16 rows per substance

graph_whole(name, time_unit='min')

Save a multi-panel plot of concentrations across all compartments. The optional time_unit argument can be 'min', 'hours', or 'days' to change the x-axis label and values.

graph_venous(name, limit_of_detection=None, log=True, time_unit='min')

Save a plot of venous blood concentrations, optionally marking a detection limit. The optional log argument controls whether the y-axis uses a logarithmic scale, and time_unit can be 'min', 'hours', or 'days' to change the x-axis label and values.

graph_compartments(compartments, name, time_unit='min')

Save a plot of selected compartments by name or index. The optional time_unit argument can be 'min', 'hours', or 'days' to change the x-axis label and values.

License

pbpk-lite is licensed under the MIT License.