def find_adjacencies_command(args):
"""
Infer co-expression modules.
"""
LOGGER.info("Loading expression matrix.")
ex_mtx = _load_expression_matrix(args)
tf_names = load_tf_names(args.tfs_fname.name)
n_total_genes = len(ex_mtx.columns)
n_matching_genes = len(ex_mtx.columns.isin(tf_names))
if n_total_genes == 0:
LOGGER.error(
"The expression matrix supplied does not contain any genes. Make sure the extension of the file matches the format (tab separation for TSV and comma sepatration for CSV)."
)
sys.exit(1)
if float(n_matching_genes) / n_total_genes < 0.80:
LOGGER.warning(
"Expression data is available for less than 80% of the supplied transcription factors."
)
LOGGER.info("Inferring regulatory networks.")
client, shutdown_callback = _prepare_client(args.client_or_address,
num_workers=args.num_workers)
try:
network = grnboost2(expression_data=ex_mtx,
tf_names=tf_names,
verbose=True,
client_or_address=client)
finally:
shutdown_callback(False)
LOGGER.info("Writing results to file.")
network.to_csv(args.output, index=False, sep='\t')
def Create_Graph(idList, labelList, percentage,
netthreshold): #CREACIÓN DEL GRAFO
dfz = load_gexpressions(idList, labelList, percentage)
#Preparo Dataframe de forma que contenga todos los genes en las columnas
dfinvert = dfz.transpose()
#Obtengo la lista de genes
TF_names = list(dfinvert)
client = Client(processes=False)
network = grnboost2(expression_data=dfinvert,
tf_names=TF_names,
client_or_address=client) # generate network
#networkG = genie3(expression_data=dfinvert, tf_names=TF_names) # generate network
#ts = time.time()
#st = datetime.fromtimestamp(ts).strftime('%Y-%m-%d_%H%M%S')
#network.to_csv(path + "\\figures\\network_" + st + ".csv")
#network=pd.read_csv("network.csv")
#labels=list(dfinvert)
limit = network.index.size * netthreshold
G = nx.from_pandas_edgelist(network.head(int(limit)),
'TF',
'target', ['importance'],
create_using=nx.Graph(directed=False))
print(nx.info(G))
return G
def process(mtx_fname, tfs, net_fname, client):
network = grnboost2(expression_data=pd.read_csv(mtx_fname,
sep='\t',
index_col=0).T,
tf_names=tfs,
verbose=True,
client_or_address=client)
network.to_csv(net_fname, index=False)
def generate_grnets(idList, labelList, percentage, netthreshold):
path = os.getcwd()
dfz = load_gexpressions(idList, labelList, percentage)
## generate network
netdata = dfz.T # rotate matrix
network = grnboost2(expression_data=netdata,
tf_names=list(netdata)) # generate network
#network = genie3(expression_data=netdata, tf_names=list(netdata)) # generate network
network.rename(columns={'importance': 'value'}, inplace=True)
# Build your graph
limit = network.index.size * netthreshold
G = nx.from_pandas_dataframe(network.head(int(limit)),
'TF',
'target',
create_using=nx.Graph())
#G=nx.from_pandas_dataframe(network, 'TF', 'target', create_using=nx.Graph() )
pos = nx.spring_layout(G, scale=10, dim=2)
#pos = nx.circular_layout(G)
#pos = nx.shell_layout(G)
#pos = nx.spectral_layout(G)
# Custom the nodes:
#nx.draw(G, with_labels=True, node_color='blue', node_size=1500, edge_color=network['value'].head(100), width=10.0, edge_cmap=plt.cm.Blues)
#nx.draw(G, with_labels=False, node_color='r', alpha=0.5, node_size=500, edge_color=network['value'].head(len(G.edges(data=True))), width=10.0, edge_cmap=plt.cm.Blues)
pos = nx.nx_pydot.pydot_layout(G)
#pos = nx.nx_pydot.pydot_layout(G, prog='dot')
#pos = nx.nx_pydot.pydot_layout(G, prog='neato')
# labels
nx.draw_networkx_labels(G, pos, font_size=8, font_family='sans-serif')
# nodes
nx.draw_networkx_nodes(G, pos, node_size=200, node_color='r', alpha=0.3)
# edges
nx.draw_networkx_edges(G,
pos,
dge_color=network['value'].head(
len(G.edges(data=True))),
width=3.0,
edge_cmap=plt.cm.Blues,
alpha=0.3)
plt.axis('off')
plt.show()
ts = time.time()
st = datetime.fromtimestamp(ts).strftime('%Y-%m-%d_%H%M%S')
#save figure with hetmap
figpath = path + "\\figures\\network_" + st + ".png"
#network.to_csv(path + "\\figures\\network_" + st + ".csv")
plt.savefig(figpath)
return figpath
#from pypanda import Panda
#from pypanda import AnalyzePanda
#from pypanda import Lioness
#import pandas as pd
#
#p = Panda('ToyExpressionData.txt', 'ToyMotifData.txt', 'ToyPPIData.txt', remove_missing=True)
#p.save_panda_results(file = 'Toy_Panda.pairs')
#plot = AnalyzePanda(p)
#plot.top_network_plot(top=100, file='top_100_genes.png')
from arboretum.algo import grnboost2, genie3
from arboretum.utils import load_tf_names
netdata = dfz.T # rotate matrix
network = grnboost2(expression_data=netdata,
tf_names=list(netdata)) # generate network
############################ PLOT 3D NETWORK ############################################
import networkx as nx
import matplotlib.pyplot as plt
# Build a dataframe with your connections
df = pd.DataFrame({
'from': ['A', 'B', 'C', 'A'],
'to': ['D', 'A', 'E', 'C'],
'value': [1, 10, 5, 5]
})
df.rename(columns={'importance': 'value'}, inplace=True)
df
def GeneRegulationNetwork(self, netthreshold, config, netconfig):
# Transpose the dataframe to get correct format to create the network
dfT = self.dfz.transpose()
# Get all the TF Gene names
tf_names = list(dfT)
# Create a Dask Client, just in case we want parellalize the algorithm
client = Client(processes=False)
# create dataframe network with columns --> TF, target Gene, Importance
if netconfig == 1:
network = grnboost2(expression_data=dfT,
tf_names=tf_names,
client_or_address=client)
print("grnboost2")
else:
network = genie3(expression_data=dfT,
tf_names=tf_names,
client_or_address=client)
# We put a threshold because we have a lot of conections and we want to obtain a clear graph with the most representatives conected genes
limit = network.index.size * netthreshold
G = nx.from_pandas_edgelist(network.head(int(limit)),
'TF',
'target', ['importance'],
create_using=nx.Graph(directed=False))
N = len(list(G.node())) # number of genes nodes
V = list(G.node()) # list of genes nodes
Edges = list(G.edges())
layt = {
1: nx.fruchterman_reingold_layout(G, dim=3),
2: nx.circular_layout(G, dim=3)
}.get(config, nx.circular_layout(G, dim=3))
laytN = list(layt.values())
Xn = [laytN[k][0] for k in range(N)] # x-coordinates of nodes
Yn = [laytN[k][1] for k in range(N)] # y-coordinates
Zn = [laytN[k][2] for k in range(N)] # z-coordinates
Xe = []
Ye = []
Ze = []
for e in Edges:
Xe += [layt[e[0]][0], layt[e[1]][0],
None] # x-coordinates of edge ends
Ye += [layt[e[0]][1], layt[e[1]][1], None]
Ze += [layt[e[0]][2], layt[e[1]][2], None]
trace1 = Scatter3d(x=Xe,
y=Ye,
z=Ze,
mode='lines',
line=Line(color='rgb(125,125,125)', width=1),
hoverinfo='none')
trace2 = Scatter3d(x=Xn,
y=Yn,
z=Zn,
mode='markers+text',
textposition='top center',
name='genes',
marker=Marker(symbol='circle',
size=3,
color='#6959CD',
colorscale='Viridis',
line=Line(color='rgb(50,50,50)',
width=1)),
text=V,
hoverinfo='text')
axis = dict(showbackground=False,
showline=False,
zeroline=False,
showgrid=False,
showticklabels=False,
title='')
fig = Figure(data=Data([trace1, trace2]),
layout=Layout(
title="Gene Regulatory Network",
width=1000,
height=1000,
showlegend=False,
scene=Scene(
xaxis=XAxis(axis),
yaxis=YAxis(axis),
zaxis=ZAxis(axis),
),
margin=Margin(t=100),
hovermode='closest',
annotations=Annotations([
Annotation(showarrow=False,
text="Khaos Research Group",
xref='paper',
yref='paper',
x=0,
y=0.1,
xanchor='left',
yanchor='bottom',
font=Font(size=20))
]),
))
plotly.offline.plot(fig, filename='3DNetworkx_.html', auto_open=True)
script = plot(fig,
output_type='div',
include_plotlyjs=False,
show_link=True)
#print(script)
return script