R framework for reproducible and quality-controlled modeling and simulation with the Open Systems Pharmacology Software
Sergei Vavilov (1), Pavel Balazki (1), Stephan Schaller (1), Juri Solodenko (2), Michael Sevestre (3)
(1) esqLABS GmbH, Saterland, Germany (2) Clinical Pharmacometrics, Bayer AG, Leverkusen, Germany (3) Design2Code Inc, Waterloo, ON, Canada
Objectives:
Physiologically-based pharmacokinetic (PBPK) and quantitative systems pharmacology/toxicology (QSP/T) modeling are crucial tools in drug development and regulatory decision-making [1]. The Open Systems Pharmacology Suite (OSPS) is a powerful open-source software suite [2,3] that offers PBPK and QSP/T modeling capabilities. The core software tools PK-Sim and MoBi enable model creation and simulation results generation in a graphical user interface (GUI) based environment. However, the GUI-based tools have limited functionalities when it comes to automated set up of simulation scenarios and reproducible systematic results analysis.
To address this gap, a landscape of R packages has been developed exclusively for the OSPS to streamline the process of PBPK and QSP/T model development. In this article, we present an overview of available packages and the developed workflow for controlled and reproducible modeling and simulation activities.
Methods:
The OSPS R landscape includes the following R packages: ospsuite [4], ospsuite.utils [5], tlf [6], ospsuite.parameteridentification [7], ospsuite.reportingengine [8], and esqlabsR [9]. The core ospsuite package is written in R and utilizes the rClr package [10] to communicate with .NET Core functionalities of the OSPS tools PK-Sim and MoBi. Simulations created with OSPS are exported to *.pkml file format and can be loaded in R. Parameter values and initial conditions can be changed, simulations run using the CVODES solver (same as PK-Sim/MoBi), and results processed. The esqlabsR package contains additional functions that allow to define simulation scenarios and output graphics using non-code Excel-based definitions.
Results:
We demonstrate an example workflow which is representative of a PBPK modeling project. It includes creating a model in PK-Sim/MoBi, exporting a ready model into a *.pkml file, loading this file into the R environment with the ospsuite package, changing the model parameters, running the simulation, comparing the simulation against the observed data with the tlf package, estimating model parameters with the ospsuite.parameteridentification package, and generating a report with the ospsuite.reportingengine package.
The esqlabsR package offers a variety of functions to facilitate reproducible and quality-controlled workflows. The workflow that we explored is based around a single PKML model file. The scenarios comprising the workflow and the figures to be generated are defined with a no-code approach using Excel files. The reports are generated in R Markdown format for easier reproducibility and traceability.
Conclusions:
The OSPS R landscape provides a set of powerful tools for controlled and reproducible modeling and simulation activities. The use of R packages allows for an automated set up of simulation scenarios and reproducible results analysis. The proposed workflow provides an efficient and effective way to manage modeling and simulation projects. The use of esqlabsR package further enhances the reproducibility and quality control of modeling and simulation activities.
References:
[1] Frechen, S., Solodenko, J., Wendl, T., Dallmann, A., Ince, I., Lehr, T., Lippert, J. and Burghaus, R. (2021), A generic framework for the physiologically-based pharmacokinetic platform qualification of PK-Sim and its application to predicting cytochrome P450 3A4–mediated drug–drug interactions. CPT Pharmacometrics Syst. Pharmacol., 10: 633-644. https://doi.org/10.1002/psp4.12636
[2] Open Systems Pharmacology Community. Open Systems Pharmacology [Internet]. 2018 [cited 2023 Mar 19]. Available from: www.open-systems-pharmacology.org
[3] Lippert J, Burghaus R, Edginton A, Frechen S, Karlsson M, Kovar A, et al. Open Systems Pharmacology community - an open access, open source, open science approach to modeling and simulation in pharmaceutical sciences. CPT Pharmacomet Syst Pharmacol. 2019 Oct 31
[4] Sevestre M, Balazki P, Solodenko J, Patil I, Vavilov S, Mil F (2023). ospsuite: R package to manipulate OSPSuite Models. https://github.com/open-systems-pharmacology/ospsuite-r
[5] Sevestre M, Balazki P, Solodenko J, Patil I (2023). ospsuite.utils: Utility Functions for Open Systems Pharmacology R Packages. https://github.com/open-systems-pharmacology/OSPSuite.RUtils
[6] Sevestre M, Chelle P, Hamadeh A (2023). tlf: TLF Library. https://github.com/open-systems-pharmacology/tlf-library
[7] Balazki P, Vavilov S (2023). ospsuite.parameteridentification: Open Systems Pharmacology Parameter Identification package. https://github.com/open-systems-pharmacology/ospsuite.parameteridentification
[8] Sevestre M, Solodenko J, Chelle P, Hamadeh A (2023). ospsuite.reportingengine: R package to create reports for the Open Systems Pharmacology models. https://github.com/open-systems-pharmacology/ospsuite.reportingengine
[9] Balazki P, Eitel J, Patil I, Vavilov S (2023). esqlabsR: esqLABS utilities package. https://github.com/esqLABS/esqlabsR
[10] Perraud, Jean-Michel. rClr package - low level access to .NET code from R. In: The R User Conference, useR! 2013; University of Castilla-La Mancha, Albacete, Spain. University of Castilla-La Mancha, Albacete, Spain; 2013. p. 15. csiro:EP132284. http://hdl.handle.net/102.100.100/97220?index=1