iplMatrix = None
global verbose, globalList, globalPathway
for o, a in opts:
if o == "-q":
verbose = False
elif o == "-i":
iplMatrix = a
## build sourceList
typeFile = "/".join(re.split("/", sifFile)[:-2] + ["TYPE.NA"])
sourceList = re.split("/", sifFile)[-2] + ".list_t"
h = mData.rList(featureFile)
(n, i) = mPathway.rSIF(sifFile, typef=typeFile)
(gn, gi) = mPathway.rPathway(globalPathway)
p = mPathway.Pathway(n, i)
s = mPathway.sortConnected(p)
f = open(sourceList, "w")
c = 1
for i in s:
u = list(set(i) & set(h))
if len(u) >= 5:
f.write("component_%s\t%s\n" % (c, "\t".join(u)))
c += 1
break
f.write("component_all\t%s\n" % ("\t".join(list(set(n.keys()) & set(h)))))
f.write("all\t%s\n" % ("\t".join(list(set(gn.keys()) & set(h)))))
f.close()
## build overlapList
try:
opts, args = getopt.getopt(args, "o:q")
except getopt.GetoptError, err:
print str(err)
usage(2)
if len(args) != 1:
log("ERROR: incorrect number of arguments", die=True)
inf = args[0]
outf = None
global verbose
for o, a in opts:
if o == "-o":
outf = a
elif o == "-q":
verbose = False
## execute
(n, i) = mPathway.rPathway(inf)
p = mPathway.Pathway(n, i)
p.selfTest()
if outf != None:
mPathway.wPathway(outf, p.nodes, p.interactions)
if __name__ == "__main__":
main(sys.argv[1:])
parser.add_option(
"--flattened",
type="string",
dest="flattened",
action="store",
help=
"Join genes with all pathway links for complexes and families they belong to. Print that network with just proteins"
)
(options, args) = parser.parse_args()
# nodes:
# name -> type
# interactions:
# name -> interacting nodes
nodes, Interactions, Proteins = mPathway.rPathway(options.pathway_file,
reverse=False,
retProteins=True)
rev_nodes, revInteractions = mPathway.rPathway(options.pathway_file,
reverse=True,
retProteins=False)
# maps complex strings to the components in each
componentMap = mPathway.getComponentMap(rev_nodes, revInteractions)
complexRE = re.compile(".*\((complex|family)\).*")
abstractRE = re.compile(".*\(abstract\).*")
# print out a 2-column interactions file of simple PPIs
# protein -> protein
ppi_edges = {}
tf_edges = {}
## parse arguments
try:
opts, args = getopt.getopt(args, "o:q")
except getopt.GetoptError, err:
print str(err)
usage(2)
if len(args) != 1:
log("ERROR: incorrect number of arguments", die = True)
inf = args[0]
outf = None
global verbose
for o, a in opts:
if o == "-o":
outf = a
elif o == "-q":
verbose = False
## execute
(n, i) = mPathway.rPathway(inf)
p = mPathway.Pathway(n, i)
p.selfTest()
if outf != None:
mPathway.wPathway(outf, p.nodes, p.interactions)
if __name__ == "__main__":
main(sys.argv[1:])
# Date: 1-29-12
# takes a sets file from expand.pl and follows interactions on non-leaf nodes
# to get an expanded set. Must specify a regex key for non-leaf nodes on input
# (default is "(abstract)")
#
# The current use for this is to find the transitive neighbors of abstract
# concepts in the superpathway
import re, sys, mPathway
from optparse import OptionParser
parser = OptionParser()
parser.add_option("-r","--non_leaf",type="string",dest="non_leaf_node", action="store", help="Non leaf regex string key", default=None)
parser.add_option("-p","--pathway_file",type="string",dest="pathway_file", action="store", help="superpathway file")
(options, args) = parser.parse_args()
rev_nodes, revInteractions = mPathway.rPathway(options.pathway_file, reverse = True, retProteins = False)
# maps complex strings to the components in each
componentMap = mPathway.getComponentMap(rev_nodes, revInteractions)
space2under = re.compile(' ')
under2space = re.compile('_')
abstractRE = re.compile(".*\(abstract\).*")
complexRE = re.compile(".*\((complex|family)\).*")
if options.non_leaf_node is None:
nonLeafRE = abstractRE
else:
nonLeafRE = re.compile(options.non_leaf_node)
# get the constituents of a complex
prefix = args[0]
inputArguments = args[1:]
global verbose
for o, a in opts:
if o == "-q":
verbose = False
## execute
inputPathways = []
for element in inputArguments:
if os.path.isdir(element):
for file in os.listdir(element):
if file.endswith("pathway.tab"):
inputPathways.append(file)
elif element.endswith("pathway.tab"):
inputPathways.append(element)
## append pathways
outPathway = mPathway.Pathway({}, {})
for file in inputPathways:
(nodes, interactions) = mPathway.rPathway(file)
appendPathway = mPathway.Pathway(nodes, interactions)
outPathway = mPathway.combinePathways(outPathway, appendPathway)
## write pathways
mPathway.wPathway(prefix, outPathway.nodes, outPathway.interactions)
if __name__ == "__main__":
main(sys.argv[1:])
def filterNet(files, phenotypes = [], statLine = None, outDir = None):
global filterBounds
filterString = "%s_%s" % (filterBounds[0], filterBounds[1])
## read global pathway
(gNodes, gInteractions) = mPathway.rPathway(globalPathway)
## read drugs
#drugData = mData.rSet(drugBank)
## write LABEL.NA, TYPE.NA
if outputAttributes:
typef = open("TYPE.NA", "w")
labelf = open("LABEL.NA", "w")
typef.write("TYPE (class=java.lang.String)\n")
labelf.write("LABEL (class=java.lang.String)\n")
for i in gNodes.keys():
typef.write("%s = %s\n" % (i, gNodes[i]))
if gNodes[i] == "protein":
labelf.write("%s = %s\n" % (i, i))
else:
labelf.write("%s = %s\n" % (i, ""))
#drugs here
typef.close()
labelf.close()
## read scores
uData = dict()
sData = dict()
for i in range(len(files)):
uData[i] = mData.rCRSData(files[i])
sData[i] = dict()
for j in uData[i].keys():
sData[i][j] = dict()
for k in uData[i][j].keys():
try:
sData[i][j][k] = abs(float(uData[i][j][k]))
except ValueError:
sData[i][j][k] = "NA"
## iterate phenotypes
for p in sData[0].keys():
if len(phenotypes) > 0:
if p not in phenotypes:
continue
pNodes = dict()
pInteractions = dict()
## write SCORE.NA
if outputAttributes:
scoref = open(p+"_SCORE.NA", "w")
scoref.write("SCORE (class=java.lang.Float)\n")
for i in gNodes.keys():
if i in uData[0][p]:
if uData[0][p][i] == "NA":
scoref.write("%s = %s\n" % (i, "0"))
else:
scoref.write("%s = %s\n" % (i, uData[0][p][i]))
else:
scoref.write("%s = %s\n" % (i, "0"))
scoref.close()
## compute thresholds
pStats = []
if statLine == None:
for i in range(len(sData.keys())):
pStats.append(mCalculate.mean_std(sData[i][p].values()))
else:
for i in re.split(",",statLine):
(v1, v2) = re.split(";",i)
pStats.append((float(v1), float(v2)))
log("%s\t%s;%s" % (p, pStats[0][0], pStats[0][1]))
for i in range(1, len(pStats)):
log(",%s;%s" % (pStats[i][0], pStats[i][1]))
log("\n")
## iterate links
for a in gInteractions.keys():
if a not in sData[0][p]:
continue
elif sData[0][p][a] == "NA":
continue
for b in gInteractions[a].keys():
if b not in sData[0][p]:
continue
elif sData[0][p][b] == "NA":
continue
## score nodes by threshold
aScore = []
bScore = []
linkScore = []
for i in range(len(sData.keys())):
linkScore.append([sData[i][p][a], sData[i][p][b]])
for i in range(len(sData.keys())):
if linkScore[i][0] > pStats[i][0]+filterBounds[1]*pStats[i][1]:
aScore.append(2)
elif linkScore[i][0] > pStats[i][0]+filterBounds[0]*pStats[i][1]:
aScore.append(1)
else:
aScore.append(0)
if linkScore[i][1] > pStats[i][0]+filterBounds[1]*pStats[i][1]:
bScore.append(2)
elif linkScore[i][1] > pStats[i][0]+filterBounds[0]*pStats[i][1]:
bScore.append(1)
else:
bScore.append(0)
## selection rule
if includeType == "OR":
if max(aScore)+max(bScore) >= 3:
(pNodes, pInteractions) = addLink(a, b, pNodes, pInteractions, gNodes, gInteractions)
elif includeType == "AND":
votes = 0
for i in range(len(sData.keys())):
if aScore[i]+bScore[i] >= 3:
votes += 0
if votes == len(sData.keys()):
(pNodes, pInteractions) = addLink(a, b, pNodes, pInteractions, gNodes, gInteractions)
elif includeType == "MAIN":
if aScore[0]+bScore[0] >= 3:
(pNodes, pInteractions) = addLink(a, b, pNodes, pInteractions, gNodes, gInteractions)
## connect top scoring disconnected nodes
sortedTop = []
for i in sData[0][p].keys():
if i not in gNodes:
continue
if gNodes[i] in ["protein"]:
sortedTop.append(i)
sortedTop.sort(lambda x, y: cmp(sData[0][p][y],sData[0][p][x]))
while (sData[0][p][sortedTop[0]] == "NA"):
sortedTop.pop(0)
if len(sortedTop) == 0:
break
for i in range(topDisconnected):
if i > len(sortedTop)-1:
break
if sData[0][p][sortedTop[i]] < pStats[0][0]+filterBounds[0]*pStats[0][1]:
break
if sortedTop[i] not in gNodes:
continue
if sortedTop[i] not in pNodes:
pNodes[sortedTop[i]] = gNodes[sortedTop[i]]
pInteractions[sortedTop[i]] = dict()
pInteractions[sortedTop[i]]["__DISCONNECTED__"] = "-disconnected-"
## output
if outDir == None:
wrtDir = p
else:
wrtDir = outDir
if not os.path.exists(wrtDir):
os.system("mkdir %s" % (wrtDir))
## output for pathway-predictor
if outputPARADIGM:
protSet = set()
for i in gNodes:
if gNodes[i] == "protein":
protSet.update([i])
netNodes = mPathway.sortConnected(pNodes, pInteractions, mPathway.revInteractions(pInteractions))
trainNodes = []
for i in netNodes:
if len((protSet) & set(i)) > featureReq:
trainNodes += i
if len(trainNodes) == 0:
log("ERROR: no nets contained enough data\n...trying again\n")
if filterBounds[0]+0.1 <= filterBounds[1]:
filterBounds[1] -= 0.1
else:
filterBounds[0] -= 0.1
filterBounds[1] -= 0.1
filterNet(files, phenotypes = phenotypes, statLine = statLine, outDir = outDir)
sys.exit(0)
(lNodes, lInteractions) = mPathway.constructInteractions(trainNodes, pNodes, pInteractions)
if outputAttributes:
mPathway.wSIF("%s/%s_%s_pp.sif" % (wrtDir, p, filterString), lInteractions)
## connect class node
classNode = "class"
lInteractions[classNode] = dict()
for i in lNodes.keys():
if i not in protSet:
continue
lInteractions[classNode][i] = "-cl>"
lNodes[classNode] = "active"
mPathway.wPathway("%s/%s_%s_pp.tab" % (wrtDir, p, filterString), lNodes, lInteractions)
## output nodrug pathway
else:
mPathway.wSIF("%s/%s_%s_nodrug.sif" % (wrtDir, p, filterString), pInteractions)
(cpNodes, cpInteractions) = mPathway.filterComplexesByGeneSupport(pNodes, pInteractions,
mPathway.revInteractions(pInteractions), gNodes,
mPathway.getComponentMap(gNodes, mPathway.revInteractions(gInteractions)))
mPathway.wSIF("%s/%s_%s_nodrug_cleaned.sif" % (wrtDir, p, filterString), cpInteractions)
iplMatrix = None
global verbose, globalList, globalPathway
for o, a in opts:
if o == "-q":
verbose = False
elif o == "-i":
iplMatrix = a
## build sourceList
typeFile = "/".join(re.split("/", sifFile)[:-2] + ["TYPE.NA"])
sourceList = re.split("/", sifFile)[-2] + ".list_t"
h = mData.rList(featureFile)
(n, i) = mPathway.rSIF(sifFile, typef=typeFile)
(gn, gi) = mPathway.rPathway(globalPathway)
p = mPathway.Pathway(n, i)
s = mPathway.sortConnected(p)
f = open(sourceList, "w")
c = 1
for i in s:
u = list(set(i) & set(h))
if len(u) >= 5:
f.write("component_%s\t%s\n" % (c, "\t".join(u)))
c += 1
break
f.write("component_all\t%s\n" % ("\t".join(list(set(n.keys()) & set(h)))))
f.write("all\t%s\n" % ("\t".join(list(set(gn.keys()) & set(h)))))
f.close()
## build overlapList
inputArguments = args[1:]
global verbose
for o, a in opts:
if o == "-q":
verbose = False
## execute
inputPathways = []
for element in inputArguments:
if os.path.isdir(element):
for file in os.listdir(element):
if file.endswith("pathway.tab"):
inputPathways.append(file)
elif element.endswith("pathway.tab"):
inputPathways.append(element)
## append pathways
outPathway = mPathway.Pathway({}, {})
for file in inputPathways:
(nodes, interactions) = mPathway.rPathway(file)
appendPathway = mPathway.Pathway(nodes, interactions)
outPathway = mPathway.combinePathways(outPathway, appendPathway)
## write pathways
mPathway.wPathway(prefix, outPathway.nodes, outPathway.interactions)
if __name__ == "__main__":
main(sys.argv[1:])
if len(args) != 3:
log("ERROR: incorrect number of arguments", die = True)
featureFile = args[0]
pathwayFile = args[1]
scoreFile = args[2]
global verbose
for o, a in opts:
if o == "-q":
verbose = False
## execute
featureList = mData.rList(featureFile)
(gNodes, gInteractions) = mPathway.rPathway(pathwayFile)
scoreMap = {}
scoreMap[sessionName] = mData.r2Col(scoreFile)
## find connected
connectList = set()
for source in featureList:
if source not in gNodes:
continue
for target in featureList:
if target not in gNodes:
continue
if source == target:
continue
paths = mPathway.shortestPath(source, target, gInteractions, maxDistance = maxDistance)
if len(paths) == 0: