Running BONITA on a Windows Desktop

If you just want to run steps 1 and 3 on a desktop and use a compute cluster, stop reading this and go to Analysis of BONITA data on a Desktop. For Unix, please read, Running BONITA on a Unix Desktop.

First, we must install a unix emulator, cygwin. Go to https://www.cygwin.com/ and download the installer. Be sure to install the following packages when prompted to select packages during installation:

• liblapack-devel
• libopenblas
• gcc-fortran
• python-gtk2
• tcl-tk
• libpng
• pkg-config
• gcc-core
• gcc-fortran
• gcc-g++
• make
• python2
• python2-bs4
• python2-cython
• python2-devel
• python2-lxml
• python2-numpy
• python2-pip
• python2-requests
• python2-setuptools
• python2-six
• python2-wheel
• wget
• git

Then complete python package installation with networkx1.11, pandas, seaborn, deap, matplotlib, and scipy:

• pip2 install --upgrade pip
• pip2 networkx==1.11
• pip2 install pytz python-dateutil
• pip2 install pandas
• pip2 seaborn
• pip2 deap
• pip -v install numpy
• pip -v install matplotlib
• pip2 install scipy
  Be careful because the scipy installation can take a very long time (hours). It is recommended this be left overnight. \

Now, we are ready to install BONITA itself.
git clone REPOSITORY NAME FROM PAPER
cd BONITA
make

BONITA is now fully built, but special instructions must be used to run as well since no SLURM queue is available. Simply follow README instructions for running BONITA, but run find_rules_desktop.sh rather than find_rules_pathway_analysis.sh at step 2. If you are not planning to use your computer for a very extended period (at least 24 hours), consider using the parallel script find_rules_desktop_parallel.py Please note that BONITA will be orders of magnitude slower when run on a single desktop and a single core rather than many cores on a compute cluster.