R-BoolNet: Random BN generation, loading with biolqm and minibn

R-BoolNet: Random BN generation, loading with biolqm and minibn#

This notebook shows how to use R BoolNet package from python notebooks to generate random Boolean networks, and load them either with biolqm or with minibn.

Two different methods are used: the first with a pure python approach; the second using Jupyter to mix both python and R code.

The documentation for BoolNet is available here: https://cran.r-project.org/web/packages/BoolNet/BoolNet.pdf

import biolqm
from colomoto import minibn

Method 1: pure python interface#

We use rpy2 to provide an interface to the BoolNet R package:

from rpy2.robjects.packages import importr
boolnet = importr("BoolNet")

boolnet is a python object offering a bridge to the R package. Then, BoolNet functions can be directly invoked as if boolnet was a python module:

net = boolnet.generateRandomNKNetwork(10, 10, topology="scale_free", gamma=2, 
                                      linkage="uniform", functionGeneration="uniform")

We save the network to a temporary file:

netfile = "/tmp/random.bnet"
boolnet.saveNetwork(net, netfile)
netfile

'/tmp/random.bnet'

We can directly load it with minibn and show its influence graph:

bn = minibn.BooleanNetwork.load(netfile)
bn.influence_graph()

# computing graph layout...
../../_images/279d94e0d0d8fe68b7fd0ed7e011d81ee79ee91d4d63f28b378bc13bee0af5fd.svg

… or with biolqm for further processing and translation to other analysis tools:

lqm = biolqm.load(netfile)
lqm

JavaObject id=o19

Method 2: Mixing python and R in Jupyter notebooks#

We first load the R extension offered by rpy2.

%load_ext rpy2.ipython

R commands should be prefixed with %R.

First, let us load the BoolNet package:

%R require(BoolNet)

array([1], dtype=int32)

Multiline R code can be written in cells starting with %%R line, with optional input and output variables to translate to/from python:

%%R -o netfiles
n <- generateRandomNKNetwork(n=10, k=10, topology="scale_free",
                             linkage="uniform", functionGeneration="uniform")
netfiles <- "/tmp/rn.bnet"
saveNetwork(n, netfiles)

Then, similarly to the first method:

bn = minibn.BooleanNetwork.load(netfiles[0])

and so on.