Personal tools
You are here: Home / Installing Biskit / Short instruction

Short instruction

...well, kind of short. These instructions should cover most cases.

Complete step-by-step instructions for a manual installation of Biskit and it's dependencies from source is described here: Install dependencies. This should rarely be needed any longer. But refer to that page if you get stuck with any of the steps below.

For the impatient, the following should work:

git clone https://github.com/graik/biskit.git biskit
pip install -r biskit/requirements.txt
pip install -e biskit

It is also highly recommend installing the biggles python package (Ubuntu package python-pybiggles; not available on PyPI) for plotting. gnuplot and pymol should come in handy, too. Read on for more detailed and alternative instructions.

Contents


1(a) Installation from source

Installation from source should work on any platform.

  • I) download or git/checkout (see below) the biskit source package:

    Select the latest biskit tar.gz from the github releases area.

    Or type in your terminal:

    wget https://github.com/graik/biskit/archive/v2.4.2.tar.gz
    tar xvf v*.tar.gz
    

    or check out the latest development snapshot:

    git clone https://github.com/graik/biskit.git biskit
    
  • II) Install dependencies

    pip install numpy scipy biopython
    
  • III) Install Biskit

    cd biskit
    python setup.py install
    

    This will install biskit into the "site-packages" folder of the python version that you used for calling setup.py. If you use a virtual environment (virtualenv), this will be venv/lib/python2.7/site-packages.

  • IV) Install Biggles

    Many Biskit modules use Biggles for generating publication-quality plots. Earlier or later, Biskit will likely raise import errors or Warnings if biggles is not installed. It is currently not available with pip. On Ubuntu, install biggles with:

    sudo apt-get install python-pybiggles
    

    Otherwise, you will have to install manually:

    You need X11 and plotutils (a.k.a libplot) development files for building biggles. (http://www.gnu.org/software/plotutils/). On some distros, this requires a separate x11-devel package. On Ubuntu / Debian:

    sudo apt-get install plotutils libplot-dev
    

    On Mac OSX:

    • you will first have to install XQuartz -- see: https://www.xquartz.org/

    • Then install plotutils:

      brew install plotutils --with-X11
      

    Check that x11 and libplot include+library files are installed:

    ~> ll /usr/lib/libXaw.so
       lrwxrwxrwx 1 root root 10 2007-05-16 14:36 /usr/lib/libXaw.so -> libXaw8.so
    

    Download and install biggles 1.6.6:

    wget http://downloads.sourceforge.net/biggles/python2-biggles-1.6.6.tar.gz
    tar xvf python2-biggles-1.6.6.tar.gz
    cd python2-biggles-1.6.6
    python setup.py build
    python setup.py install
    

    Test Biggles:

    python
    >>> import biggles
    >>>
    

    Then paste the following code snippet into your python interpreter:

    import biggles, numpy
    x = numpy.arange(0,10,0.3)
    p = biggles.FramedPlot()
    p.add( biggles.Curve(x, numpy.cos(x)) )
    p.show()
    

    Consult our biggles troubleshooting guide if you experience any problems!

    Other plotting and visualization-related tools that are useful:

    sudo apt-get install gnuplot pymol
    
  • V) Test your installation

    python Biskit/test.py -e exe old pvm
    

    This will run all Biskit tests which are not marked as requiring third-party programs (exe) or as old (old) and which do not require pvm parallelization (pvm). They should all pass.

    Update 06/2016:

    • On MacOSX, I encountered one test failure from PDBCleaner.test_capping_extra. This seems to be a mere numeric difference meaning the chain break detection is slightly more tolerant on MacOSX systems.
    • While Biskit can be used without Biggles, several tests strictly depend on it and fail if biggles is not installed.

1(b) Installation from Ubuntu / Debian package

Fetch the debian package (.deb) from the github releases area on: https://github.com/graik/biskit/releases/download/v2.4.3/python-biskit_2.4.3-1_all.deb

On Ubuntu, a click on the .deb file will automatically open the Ubuntu package manager. Just press "Install". Otherwise type:

gdebi python-biskit_2*_all.deb

This gives you a biskit installation that is lacking parallelization because there are no up-to-date debian packages for pypvm (see below).

1(c) Installation on Mac OS X

Marc-Michael Blum has compiled a Fink package for easy installation: http://biskit.pasteur.fr/install/osx_install

Update 06/2016: there is currently no fink package for the most recent biskit versions.

1(d) Installation from rpm package

Source and platform-independent "binary" rpms are provided in the github releases area. These files have not yet been tested. Please, report success or failure and the steps needed.

1(e) Installation on Windows

This is rather experimental. The core functions seem to work though. See Windows installation


2) setup PVM for parallelization (optional)

PVM is really useful and many Biskit scripts use it to distribute calculations. Most Unix distributions have a precompiled pvm package and you should use it. The python bindings for PVM (pypvm) are available from the biskit web site. See the PVM section in install dependencies for detailed instructions.


3) install helper applications

This depends on what you actually need. A list of all helper applications is given here: Helper applications. See also Install helper applications for more detailed installation instructions and troubleshooting. The Biskit library itself doesn't need any external programs but most people find the following programs useful: gnuplot, surfaceracer, and DSSP.


4) test Biskit

run:

~> python
Python 2.4.1 (#1, May 16 2005, 15:15:14)
[GCC 4.0.0 20050512 (Red Hat 4.0.0-5)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import Biskit

You should get several warnings about missing configuration files. That's OK. But make sure you don't get any exception.ImportError because the latter would indicate missing or broken libraries or an incorrect $PYTHONPATH.

Next, see Testing Howto for instructions on how to test all Biskit modules! In short:

cd /usr/lib/python2.5/site-packages/Biskit
python test.py -e old pvm exe

...will run all tests, except those that are depending on PVM and external helper applications (the intervor test is not tagged as 'exe' test case in release 2.3.1 and will report an error, please ignore this one).


5) Customize biskit

If at all necessary, adapt ~/.biskit/settings.cfg to your environment.

See Setup environment for details!


6) Removing Biskit

  • After installation from Debian packages:

    sudo apt-get remove python-biskit
    
  • After installation from source:

    cd /usr/local/lib/python2.5/site-packages
    sudo rm -r biskit* Biskit
    

    Biskit may also have ended up in /usr/lib/python2.5/site-packages or /usr/lib/python2.6/site-packages

7) List of all Biskit dependencies

See Install dependencies for detailed (cut-and-paste) instructions!

program/library comment or link
core functions --------- absolutely necessary ---------
python http://www.python.org/ (2.5 or higher, 3.x not supported)
python-dev python development files
python-numpy
the new incarnation of Numeric (for Biskit 2.1+)
python-scipy http://www.scipy.org/
plotting --------- recommended ---------
plotutils
required for building biggles
python-biggles
plotting library http://biggles.sf.net/
or use Ubuntu package python-pybiggles
gnuplot plotting program
homology modeling --------- recommended ---------
python-biopython
required for homology modeling (Biskit.Mod)
also required for fetching PDBs remotely from NCBI
parallelization --------- required by Biskit.PVM ---------
openssh-server usually already pre-installed
pvm http://www.csm.ornl.gov/pvm/ ParallelVirtualMachine
pvm-dev PVM development files
pypvm PVM python bindings -- use our repackaged pypvm!