3. MISC code

Contents

• MISC code
  • Renames nodes in a graph and combining several graph
    • combine with cellnopt.wrapper

miscellaneous

This section provides piece of code that are not finalised but could be useful.

3.1. Renames nodes in a graph and combining several graph

status:draft

This is a piece of code used to create a big network by combining three times the same network. For the renaming, you need networkx relabel_node function that returns a n graph. This is not inline modification. The function expect a mapping of new names to the old ones, which is done using a dict conversion of a list of tuples:

# create a graph
c1 = CNOGraph(a valid filename, midas name)
c1.add_edge("A", "B", link="+")
c1.add_edge("A", "C", link="+")
# create duplicates with different names
c2 = nx.relabel_nodes(c, dict([(x, x+"_a" )for x in c.nodes()]))
c3 = nx.relabel_nodes(c, dict([(x, x+"_b" )for x in c.nodes()]))

# create a full network with all previous edges/nodes
c = c1+c2+c3
#c.midas =

# finally add cross-talk edges from subPKN1 to 2 and 2 to 3
Nedges = 10

nodes1 = random.sample(c1.nodes(), Nedges)
nodes2 = random.sample(c2.nodes(), Nedges)
for n1, n2 in zip(nodes1, nodes2):
    if n1!=n2:
        c.add_edge(n1,n2, link="+")
nodes1 = random.sample(c2.nodes(), Nedges)
nodes2 = random.sample(c3.nodes(), Nedges)
for n1, n2 in zip(nodes1, nodes2):
    if n1!=n2:
        c.add_edge(n1,n2, link="+")

#c.midas =
c._signals += [x+"_a" for x in c1.midas.cnolist.namesSignals]
c._signals += [x+"_b" for x in c1.midas.cnolist.namesSignals]
c._stimuli += [x+"_b" for x in c1.midas.cnolist.namesStimuli]
c._stimuli += [x+"_a" for x in c1.midas.cnolist.namesStimuli]
c._inhibitors += [x+"_a" for x in c1.midas.cnolist.namesInhibitorsShort]
c._inhibitors += [x+"_b" for x in c1.midas.cnolist.namesInhibitorsShort]

c._node_set_attributes()
c.plotdot()
nx.write_dot(c, "test.dot")

3.1.1. combine with cellnopt.wrapper

from cellnopt.core import *
from cellnopt.wrapper import *
import tempfile

f1 = tempfile.NamedTemporaryFile(delete=False)
midasFilename = cnodata("MD-ToyMMB.csv")

c = CNOGraph(cnodata("PKN-ToyMMB.sif"), cnodata("MD-ToyMMB.csv"))
# play with the SIF file and save it
c.export2sif(f1.name)

b = CNORbool(f1.name, midasFilename)
b.preprocessing()
b.gaBinaryT1()

# the optimised model
c2 = CNOGraph()
reactions = list(b.expmodel.reacID)
bString = list(b.T1opt.bString)
s = SIF()
for r in reactions: s.add_reaction(r)
c2 = CNOGraph(s, midasFilename)


# TODO: be able to add AND gate in the SIF or CNOGraph data structure

# Need to color the edges now and/or add a label