Perl-speaks-NONMEMPerl-speaks-NONMEM (PsN) is a collection of Perl modules and programs aiding in the development of non-linear mixed effect models using NONMEM. The functionality ranges from solutions to simpler tasks such as parameter estimate extraction from output files, data file sub setting and resampling, to advanced computer-intensive statistical methods. PsN includes stand-alone tools for the end-user as well as development libraries for method developers.
Use the top menu for information on how to Download and Install PsN. You can read about the different parts of PsN under Documentation and also learn about Known Bugs, how to submit a new bug and how to become a part of the PsN mailing list.
The latest release of PsN is version 5.3.1. Instructions on how to download it and a detailed list of changes can be found in the download section. A brief list of what is new can be found below in the news section.
PsN 5.3.1 released (2023-04-26)Changes in this version include:
- Bugfixes to qa
- Support for datasets with large IDs in vpc and mcmp
- Fix issue with adding $ETAS in scm for non-linearized models
PsN 5.3.0 released (2022-02-24)Changes in this version include:
- New tool m1find to find and remove m1 directories and NM_run directories
- Support for $DESIGN
- Further simplified installation by not depending on Moose, YAML:XS and File::Copy::Recursive perl packages
- Removed support for NMQual
PsN 5.2.6 released (2021-05-24)New features in this version include:
- The new machine readable results.json file created also for cdd, scm, simeval, bootstrap and qa
- scm plus was merged into PsN (Thanks Pharmetheus for the contribution)
- Installation on Linux and MacOS is simplified by not depending on some hard to install perl and R packages
PsN 5.0.0 released (2020-06-18)New features in this version include:
- A simplified installation procedure for Windows with all dependencies bundled together.
- Usage of the new Pharmpy python package. Read more about Pharmpy on its own webpage
- Updates to FREM: no need to run postfrem separately, automatic reordering of omegas and bugfixes
- New file formats for FREM results: results.json and results.csv. One machine readable and one human readable.
- Extended html report for qa
Tags for PsNR releases (2019-09-20)For rplots and qa to work properly each PsN version needs a specific version of PsNR. This will be handled by tagging PsNR versions with a PsN version tag. For PsN 4.9.0 this means that it is not adviced to install PsNR from setup.pl. Please instead follow the Psn installation instructions
PsN 4.9.0 released (2019-06-05)New features in this version include:
- Many improvements to the quality assurance tool qa
- New option parameters to sumo makes comparing model parameters and OFV easier
- R code for plotting has been moved to an R package PsNR
- Stratification for crossval
- New option
-debug_rmdto keep the tex and md files when using
- New clean level 5 to remove the entire run directory
PsN 4.8.1 released (2018-06-08)This version is a very minor upgrade from 4.8.0. To simplify installation of all R packages needed for qa and rplots an R-script was added. It has an option to install the needed R packages into a separate library. Then a setting in psn.conf can set this library to be the default for PsN. When having problems with getting R plots out from PsN it is recommended to use this script to get the right versions of all R packages.
Rscript PsN/R-scripts/install_R_packages.R myRlib
In psn-conf add:
PsN 4.8.0 released (2018-05-25)New features in this version include:
- A new tool qa for fast and automatic assumption assessment and quality assurance of models
- A new tool transform, to manipulate model files. For example to automatically add IOV or Box-Cox transform ETAs
- A new optional directory structure using a subdirectory for each model via the option -model_subdir
- The addition of support for vpc for mixture models
- Second order approximation added to the linearize tool
- The presentation of meta data in a new file, meta.yaml, which is in a both human and computer readable format
©2022-2023 by Mats Karlsson and Rikard Nordgren. All rights reserved.
©2020-2021 by Mats Karlsson, Rikard Nordgren and Sebastian Ueckert. All rights reserved.
©2018-2019 by Mats Karlsson, Rikard Nordgren, Svetlana Freiberga, Sebastian Ueckert and Gunnar Yngman. All rights reserved.
©2016-2017 by Mats Karlsson, Andrew Hooker, Rikard Nordgren, Kajsa Harling and Svetlana Freiberga.
©2013-2015 by Mats Karlsson, Andrew Hooker, Rikard Nordgren and Kajsa Harling.
©2008-2012 by Mats Karlsson, Niclas Jonsson, Andrew Hooker and Kajsa Harling.
©2006-2007 by Lars Lindbom.
©2000-2005 by Lars Lindbom and Niclas Jonsson.
Perl-speaks-NONMEM is a free and open source software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 2 of the License, or (at your option) any later version.
PsN is developed and maintained by Rikard Nordgren.
PsN was originally developed by Niclas Jonsson and continued by Lars Lindbom for his doctoral thesis.
Lead developer for PsN between 2008 and 2016 was Kajsa Harling.
Additional implementation has been done by Pontus Pihlgren, Jakob Ribbing, Kristin Karlsson, Maria Kjellsson and Joakim Nyberg.
Additional contributions by Radojka Savic, Paul Baverel, Martin Bergstrand, Elodie Plan, Yasunori Aoki, William Denney and many more.
References: Lindbom L, Pihlgren P, Jonsson EN. PsN-Toolkit--a collection of computer intensive statistical methods for non-linear mixed effect modeling using NONMEM. Comput Methods Programs Biomed. 2005 Sep;79(3):241-57.
Lindbom L, Ribbing J, Jonsson EN. Perl-speaks-NONMEM (PsN)--a Perl module for NONMEM related programming. Comput Methods Programs Biomed. 2004 Aug;75(2):85-94.
Acknowledgements: PsN was developed in parts with funding from the DDMoRe and IDeAl projects.
NONMEM® is a registered trademark of ICON plc. The Perl camel logo is a registered trademark of O'Reilly Media, Inc. and is used with permission.
All logos and trademarks in this site are property of their respective
Some computations needed for testing PsN were enabled by resources provided by the Swedish National Infrastructure for Computing (SNIC) at UPPMAX, partially funded by the Swedish Research Council.