Related Experiment Videos
PsN-Toolkit--a collection of computer intensive statistical methods for non-linear mixed effect modeling using
Lars Lindbom1, Pontus Pihlgren, E Niclas Jonsson
1Department of Pharmaceutical Biosciences, Division of Pharmacokinetics and Drug Therapy, Uppsala University, Box 591, SE-751 24 Uppsala, Sweden. lars.lindbom@farmbio.uu.se
Computer Methods and Programs in Biomedicine
|July 19, 2005
Summary
The PsN-Toolkit simplifies pharmacometric data analysis by offering object-oriented Perl tools for NONMEM. It integrates Bootstrap, Jackknife, profiling, diagnostics, and model building, enhancing workflow efficiency.
Area of Science:
- Pharmacometrics
- Computational Statistics
- Software Development
Background:
- Pharmacometric data analysis often requires complex statistical methods.
- Non-linear mixed-effects modeling (NONMEM) is a key technique in pharmacometrics.
- Efficient implementation of statistical tools is crucial for reproducible research.
Purpose of the Study:
- To introduce the PsN-Toolkit, a collection of statistical tools for pharmacometric analysis.
- To facilitate the use of advanced statistical methods within the NONMEM framework.
- To provide a flexible and efficient software solution for pharmacometricians.
Main Methods:
- The PsN-Toolkit is object-oriented, written in Perl, and utilizes the Perl-speaks-NONMEM (PsN) library.
- It includes classes for Bootstrap, Jackknife, Log-likelihood Profiling, Case-deletion Diagnostics, and Stepwise Covariate Model building.
- Tools support integration into user scripts and stand-alone command-line applications, with parallel execution capabilities.
Main Results:
- The toolkit offers integrated functionalities for common pharmacometric analysis tasks.
- It supports parallel processing across various computing environments, including SMP, Mosix/openMosix clusters, and NorduGrid.
- The object-oriented design promotes code reusability and streamlined workflow management.
Conclusions:
- PsN-Toolkit enhances the accessibility and efficiency of key statistical methods in pharmacometric data analysis.
- The toolkit simplifies the application of Bootstrap, Jackknife, Log-likelihood Profiling, Case-deletion Diagnostics, and Stepwise Covariate Model building.
- It provides a robust and scalable solution for pharmacometricians utilizing NONMEM.