graphList = []
label = []
for model_name in sorted(os.listdir(model_library_folder)):
print('Loading', model_name)
model = cobra.io.read_sbml_model(model_library_folder + model_name)
label.append(model.name)
g = modelNet(model)
graphList.append(g)
print('Done')
GL = pd.DataFrame(list(zip(label, graphList)), columns=['organism', 'graph'])
#compute GK similarity matrix
kernel = gk.WeisfeilerLehman(base_kernel=gk.VertexHistogram, normalize=True)
GK = pd.DataFrame(kernel.fit_transform(GL['graph'].values))
GK.columns = GK.index = label
# Use 1-K as measure of Distance
DM_GK = DistanceMatrix(1 - GK.values)
#make GK tree
sktree = nj(DM_GK, result_constructor=str)
GK_tree = Tree(sktree)
GK_tree.name = 'AGORA network similarity tree'
# style
ts = TreeStyle()
ts.show_leaf_name = True
ts.mode = "c"
ts.arc_start = -180
def gKernel_DM(graphList):
# returns 1 - kernel similarity matrix (i.e. distance)
gkernel = gk.WeisfeilerLehman(base_kernel=gk.VertexHistogram,
normalize=True)
K = pd.DataFrame(gkernel.fit_transform(graphList))
return 1 - K