News | Downloading GNU MCSim | Documentation | Installation | Mailing lists | Contributed material and related links | Getting involved | Licensing
GNU MCSim
GNU MCSim is a simulation package, written in C, which allows you to:
- design your own statistical or simulation models (eventually dynamic, via ODEs),
- perform Monte Carlo stochastic simulations,
- do Bayesian inference through Markov Chain Monte Carlo simulations.
News
January 1st, 2016 - Happy new version of GNU MCSim: version 5.6.3.
Version 5.6.3 essentially fixes a small bug in output scheduling and add various checks (checking that output variables are actually computed for example).
Version 5.6.2 fixes problems with reading restart files (introduced in version 5.6.0) and a bug in the MCMC proposal kernel weighting. That bug had a minor impact on posterior sampling in case of bounded prior distributions in the absence of informative data.
"Events" (state variables discontinuities) in GNU MCSim
State variables disconstinuities can now be requested at given times. See the online user's manual.
Delay differential equations
GNU MCSim can now easily solve delay differential equations. A new example (examples/delay_diff_eqns/) is available.
GNU MCSim and R (even under Microsoft Windows)
The basic tools needed to build and run GNU MCSim are not available to
most users of Windows systems.
Fortunately, the R software (see
http://www.r-project.org/), when properly installed with its Rtools
(this requires administrator's rights), can compile and run GNU MCSim
models. We have developed code and R scripts to compile GNU MCSim
models to C code usable by the R package deSolve (to do that use the
GNU MCSim "mod" model generator with the -R option) or by itself. This
works under any operating system. You can then use deSolve to perform
simulations of your models. The integators provided by deSolve are
improved implementations of the lsode family of integrators (used also
by GNU MCSim), and provide a few more options than GNU MCSim (in
particular, "events", see the deSolve user's manuals). However, if you
need raw speed (say, for Markov chain Monte Carlo simulations) GNU
MCSim is the fastest option.
Follow this link to download the corresponding scripts and code.
You need to have R and its Rtools (and optionally the deSolve package) installed. Follow the simple instructions given in the included README text file.
Downloading GNU MCSim
The latest version of GNU MCSim (5.6.3) can be
found on any
GNU mirror.
You can also download it from the main GNU ftp server:
- via HTTP at
http://ftp.gnu.org/gnu/mcsim/
- via FTP at
ftp://ftp.gnu.org/gnu/mcsim/.
- A User's Manual (in Texinfo, pdf and HTML formats)
- The ASCII source code of the Mod program (to preprocess your models)
- The ASCII source code of the GNU MCSim routines (to link with your models)
- Examples of model and simulation files
Older versions can also be obtained on the GNU ftp mirrors and main page.
GNU MCSim development is hosted on savannah.gnu.org. See the MCSim project page on Savannah, where the latest development sources are publicly available.
Documentation
The User's manual for GNU MCSim is available online, as is documentation for most GNU software. You may also find more information about GNU MCSim by running info mcsim, or by looking at /usr/share/mcsim/doc, or similar directories on your system.
The pdf version of the manual can be found here here
Installation
To install GNU MCSim, you will need a C compiler and linker to compile the sources and obtain executables. For any machine we recommend the GNU gcc compiler, but standard C compilers should also work.
Versions 5.0.0 up to 5.3.1 required GNU Scientific Library (gsl, and its companion gslcblas) to be installed. With version 5.4.0, GNU gsl becomes optional (but recommended).
Versions above 5.4.0 also prefers libSBML v4.2.0 or above to be installed, but that is not needed if you do not want to read SBML model files.
The user's manual (online here) provides detailed instructions for installation for several platforms.
Mailing lists
GNU MCSim has the following mailing lists:
- bug-mcsim is used to discuss most aspects of GNU MCSim, including development and enhancement requests, as well as bug reports.
- help-mcsim is for general user help and discussion.
Announcements about GNU MCSim and most other GNU software are made on info-gnu (archive). See also info-mcsim.
Security reports that should not be made immediately public can be sent directly to the maintainer. If there is no response to an urgent issue, you can escalate to the general security mailing list for advice.
You can also look at the same lists, GNU MCSim help, and GNU MCSim bug report, on the gmane portal which has spam control.
Contributed material and related links
-
Supplementary material for the article "GNU MCSim: Bayesian statistical inference for SBML-coded systems biology models" Bioinformatics, 1 June 2009; 25: 1453 - 1454. That paper demonstrates the application of GNU MCSim MCMC sampling, optimal design and multilevel modeling to SBML models.
-
Supplementary material containing all model and input files for predicting interactions between benzene, toluene, ethylbenzene and xylene via PBPK and systems biology coupling.
-
A PDF software validation report for version 5.0.0 can be found here. (The bug with the half-normal distribution it mentions has been fixed in version 5.1.0 and later versions.)
Getting involved
Development of GNU MCSim, and GNU in general, is a volunteer effort, and you can contribute. For information, please read How to help GNU. If you'd like to get involved, it's a good idea to join the discussion mailing list (see above).
- Development
- For development sources, issue trackers, and other information, please see the GNU MCSim project page at savannah.gnu.org.
- Maintainer
- GNU MCSim is currently being maintained by Frederic Y. Bois. Please use the mailing lists for contact.
Licensing
GNU MCSim is free 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 3 of the License, or (at your option) any later version.