Skip to content

Efficient coalescent simulations in a spatial continuum

License

Notifications You must be signed in to change notification settings

JereKoskela/discsim

 
 

Repository files navigation

Efficient coalescent simulation in continuous space

Simulates the coalescent for populations evolving in a spatial continuum under the extinction/recolonisation model. This package is a specialisation of the ercs package, and provides a much more efficient method of simulating the spatial coalescent for the disc model. A very similar (but not identical) interface to ercs is provided.

The simulations support:

  • A sample of n individuals with m loci at arbitrary locations on a torus of diameter L.
  • A fixed recombination rate between pairs of adjacent loci.
  • An arbitrary number of classes of event occuring at fixed rates.
  • Simulations in one and two dimensions.
  • Access to the locations of ancestors at any time in the past.
  • Simulations of the locations of all pedigree ancestors, as well as the genetic ancestors.

The discsim module supports Python 2 and 3.

Documentation

Here's a quick example for the impatient:

import ercs
import discsim
sim = discsim.Simulator(10)
sim.sample = [None, (3, 2), (6, 4), (7, 0)]
sim.event_classes = [ercs.DiscEventClass(u=0.5, r=1)]
sim.run()
pi, tau = sim.get_history()

Full documentation for discsim is available at http://pythonhosted.org/discsim.

Installation

Quick install for Debian/Ubuntu

If you are running Debian or Ubuntu, this should get you up and running quickly:

$ sudo apt-get install python-dev libgsl0-dev
$ sudo pip install ercs discsim

For Python 3, use python3-dev and pip3.

General instructions

The discsim module depends on the GNU Scientific Library, which must be installed before it can be built. Fortunately, this is straightforward on most platforms. For example, on Debian or Ubuntu use:

$ sudo apt-get install libgsl0-dev

or on Fedora:

$ sudo yum install gsl-devel

GSL is available on most packaging systems; if it is not available on your platform, it can be installed from source.

The discsim module also depends on the ercs Python module, which must also be installed, using the same methods as outlined below.

Once GSL has been installed we can build the discsim module using the standard Python methods. For example, using pip we have

$ sudo pip install discsim

Or, we can manually download the package, unpack it and then run:

$ python setup.py build
$ sudo python setup.py install

Most of the time this will compile and install the module without difficulty.

It is also possible to download the latest development version of discsim from github.

Potential problems

On platforms that GSL is not available as part of the native packaging system (or GSL was installed locally because of non-root access) there can be issues with finding the correct headers and libraries when compiling ercs and discsim. For example, on FreeBSD we get something like this:

$ python setup.py build
... [Messages cut for brevity] ...
_discsimmodule.c:515: error: 'sim_t' has no member named 'time'
_discsimmodule.c: In function 'Simulator_get_num_reproduction_events':
_discsimmodule.c:529: error: 'sim_t' has no member named 'num_reproduction_events'
_discsimmodule.c: In function 'Simulator_get_history':
_discsimmodule.c:743: error: 'sim_t' has no member named 'pi'
_discsimmodule.c:748: error: 'sim_t' has no member named 'tau'
_discsimmodule.c: In function 'Simulator_run':
_discsimmodule.c:789: error: 'sim_t' has no member named 'time'
error: command 'cc' failed with exit status 1

This can be remedied by using the gsl-config program to set the the LDFLAGS and CFLAGS environment variables to their correct values:

$ CFLAGS=`gsl-config --cflags` LDFLAGS=`gsl-config --libs` python setup.py build

Tests

discsim provides some test cases to ensure that the installation has gone smoothly. It is a good idea to run these immediately after installation:

$ python tests.py

Tested platforms

Discsim has been successfully built and tested on the following platforms:

Operating system Platform Python Compiler
Debian wheezy x86_64 2.7.3 gcc 4.7.2
Debian wheezy x86_64 3.2.3 gcc 4.7.2
Debian wheezy x86 2.7.3 gcc 4.7.2
Debian squeeze ppc64 2.6.6 gcc 4.4.5
Debian squeeze ppc64 3.1.3 gcc 4.4.5
Debian squeeze x86_64 2.6.6 gcc 4.4.5
Debian squeeze x86_64 3.1.3 gcc 4.4.5
FreeBSD 9.2 x86_64 2.7.5 gcc 4.2.1
FreeBSD 9.2 x86_64 3.3.2 gcc 4.2.1
Fedora 19 x86_64 2.7.5 gcc 4.8.1
Fedora 19 x86_64 3.3.2 gcc 4.8.1
SunOS 5.10 SPARC 3.3.2 gcc 4.8.0

About

Efficient coalescent simulations in a spatial continuum

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C 75.8%
  • Python 23.9%
  • Makefile 0.3%