ModelPK is a package designed to extract basic information about the pharmacokinetic profile of a drug from experimental data.
Target users: for bench scientists who have little experience with PK modeling
Pharmacokinetics (PK) is the study of how a drug moves through the body. More specifically, it examines:
- Absorption: How does a drug get into the body?
- Distribution: Where does the drug go? For example, does it remain in the bloodstream, or does it partition into certain tissues?
- Metabolism: Does it get broken down by the body?
- Excretion: How does it leave the body? At any given point, the concentration of a drug in the body will be impacted by its absorption, distribution, metabolism, and excretion.
PK is crucial to understanding the safety and efficacy of a drug. Every drug has a concentration above which it can have serious side effects or be toxic to patients. Likewise, every drug has a concentration below which it no longer has a therapeutic effect. As such, every drug has a therapeutic window in which it actually has a therapeutic effect for patients. Understanding PK profile of a drug allows clinicians and physicians to understand how to keep drug concentrations within this therapeutic window.
To install ModelPK run the following command:
!pip install ModelPK
This should also install any required dependencies listed in the requirements.txt file.
- The dependencies below are required for ModelPK:
- numpy
- pandas
- sci-kitlearn
- tellurium
All required dependencies should be included when ModelPK or any of its modules is imported.
-
Import ModelPK as entire package:
import ModelPKOR import each module separately:
from ModelPK import extractPKparam as extractfrom ModelPK import simulatePK as sim -
See "examples" folder for sample code.
This package is currently only supports a 1 compartment model for a drug administered as an IV bolus. In other words, the drug must: i) have been administered intravenously as a single, large dose AND ii) remain in the bloodstream and do not partition into other tissues.
Future work will expand the number of PK models supported.