def enrichmentIndex(self, pAC50All, FP=1):
self.pAC50All = pAC50All
if FP == 1:
prenrich = self.prout + self.cdesc.pFP.split("/")[-1] + "_" + str(self.aggType) + "_enrich-index" + "/"
else:
prenrich = self.prout + str(self.clusterMeth) + "_" + str(self.distmeth) + "_" + str(self.aggType) + "_enrich-index/"
pathFolder.createFolder(prenrich)
self.prenrich = prenrich
lfileenrich = listdir(prenrich)
if FP == 0:
if not "pdesclean" in self.__dict__:
self.pdesclean = prenrich + "descClean.csv"
if not path.exists(self.pdesclean):
runExternalSoft.dataManager(self.cdesc.pdesc1D2D, 0, self.corval, self.maxquantile, prenrich)
if len(lfileenrich) > 2: return 0
runExternalSoft.enrichmentIndex(self.pdesclean, prenrich, pAC50All, self.clusterMeth, self.distmeth, self.aggType)
else:
if len(lfileenrich) > 2: return 0
runExternalSoft.enrichmentIndex(self.cdesc.pFP, prenrich, pAC50All, self.clusterMeth, self.distmeth, self.aggType)
return 0
def setConstantPreproc(self, pAC50, corval, maxQuantile, nbNA, prAnalysis):
self.corval = corval
self.maxQauntile = maxQuantile
self.pAC50 = pAC50
self.prAnalysis = prAnalysis
# output
paffclean = self.prAnalysis + "AC50Clean.csv"
pdesc1D2Dclean = self.prAnalysis + "descClean.csv"
if path.exists(paffclean):
self.pAC50clean = paffclean
if path.exists(pdesc1D2Dclean):
self.pdesc1D2Dclean = pdesc1D2Dclean
return 0
elif path.exists(self.pdesc1D2D) and path.getsize(self.pdesc1D2D) > 10:
# preproc
runExternalSoft.dataManager(self.pdesc1D2D, self.pAC50,
self.corval, self.maxQauntile, nbNA,
self.prAnalysis)
if path.exists(paffclean):
self.pAC50clean = paffclean
if path.exists(pdesc1D2Dclean):
self.pdesc1D2Dclean = pdesc1D2Dclean
return 0
def createMDS(self, pdesc1D2D, pAC50, corval, maxQuantile, prMDS):
# output
pdesc1D2Dclean = prMDS + "descClean.csv"
if not path.exists(pdesc1D2Dclean):
if path.exists(pdesc1D2D) and path.getsize(pdesc1D2D) > 10:
# preproc
runExternalSoft.dataManager(pdesc1D2D, 0, corval, maxQuantile,
prMDS)
else:
print "Error ->", pdesc1D2D
runExternalSoft.drawMDS(pdesc1D2Dclean, pAC50, prMDS)
def createSOM(pdesc1D2D, pAC50, corval, maxQuantile, pModel, nbNA, prSOM):
# output
pdesc1D2Dclean = prSOM + "descClean.csv"
if not path.exists(pdesc1D2Dclean):
if path.exists(pdesc1D2D) and path.getsize(pdesc1D2D) > 10:
# preproc
runExternalSoft.dataManager(pdesc1D2D, 0, corval, maxQuantile, nbNA, prSOM)
else:
print "Error ->", pdesc1D2D
runExternalSoft.drawEnrichSOM(pdesc1D2Dclean, pAC50, pModel, prSOM)
def PCACross(pdesc, pAC50_hepg2, pAC50_hek293, nbNA, corval, maxQuantile,
prCrossPCA):
# output
pdesc1D2Dclean = prCrossPCA + "descClean.csv"
if not path.exists(pdesc1D2Dclean):
if path.exists(pdesc) and path.getsize(pdesc) > 10:
# preproc
runExternalSoft.dataManager(pdesc, 0, corval, maxQuantile, nbNA,
prCrossPCA)
else:
print "Error ->", pdesc
runExternalSoft.drawPCACross(pdesc1D2Dclean, pAC50_hepg2, pAC50_hek293,
prCrossPCA)
def loadClassForPred(self, corval, maxQuantile, lCASID=[]):
#Desc 1D and 2D
self.computeDesc()
lpdescClean = runExternalSoft.dataManager(self.pdesc1D2D, "0", corval,
maxQuantile, self.prout)
self.pdesc1D2Dclean = lpdescClean[0]
if lCASID != []:
self.reduceMatrixDesc(self.pdesc1D2Dclean, lCASID)
self.computeFP("All")
if lCASID != []:
self.reduceMatrixFP(self.dFP, lCASID)
def createMainClustering(self, doublecluster = 0, lcas = []):
if self.distmeth == None: # case of fp
self.prCluster = self.prout + self.cdesc.pFP.split("/")[-1] + "-" + str(self.clusterMeth) + "_" + str(self.distmeth) \
+ "_" + str(self.aggType.replace(".", "")) + "_" + str(self.optimalNBclustMeth) + "/"
else:
self.prCluster = self.prout + str(self.clusterMeth) + "_" + str(self.distmeth) + "_" + str(
self.aggType.replace(".", "")) + "_" + str(self.optimalNBclustMeth) + "/"
pathFolder.createFolder(self.prCluster)
# data preparation
self.pdesclean = self.prCluster + "descClean.csv"
if not path.exists(self.pdesclean):
if path.exists(self.cdesc.pdesc1D2D) and path.getsize(self.cdesc.pdesc1D2D) > 10:
# preproc
if self.distmeth == None:
if lcas != []:
self.cdesc.reduceMatrixFP(lcas, self.pdesclean)
else:
copyfile(self.cdesc.pFP, self.pdesclean)
else:
runExternalSoft.dataManager(self.cdesc.pdesc1D2D, 0, self.corval, self.maxquantile, self.prCluster)
else:
print "Error ->", self.cdesc.pdesc1D2D
pcluster = self.prCluster + "cluster.csv"
if not path.exists(pcluster):
if self.distmeth == None:
pcluster = runExternalSoft.clusteringSimilarityMatrix(self.pdesclean, self.prCluster, self.aggType, self.clusterMeth, self.optimalNBclustMeth)
else:
#clustering -> first level
pcluster = runExternalSoft.clustering(self.pdesclean, "0", self.prCluster, self.distmeth, self.aggType, self.clusterMeth, self.optimalNBclustMeth)
if doublecluster == 1:
#Clustering second level
if pcluster != 0:
self.createSecondaryClustering(pcluster)
# create main cluster file
pclustersFinal = self.prCluster + "clusterMain.csv"
if not path.exists(pclustersFinal):
fclustersFinal = open(pclustersFinal, "w")
fclustersFinal.write("ID,Cluster1,Cluster2\n")
fcluster1 = open(pcluster, "r")
lchemCluster1 = fcluster1.readlines()
fcluster1.close()
dclust = {}
for chemCluster1 in lchemCluster1[1:]:
chemCluster1 = chemCluster1.strip().replace("\"", "").split(",")
chemID = chemCluster1[0]
clust = chemCluster1[1]
dclust[chemID] = [clust]
for fileCluster in listdir(self.prCluster):
if search("Clust", fileCluster):
pclust2 = self.prCluster + fileCluster + "/cluster.csv"
if path.exists(pclust2):
fclust2 = open(pclust2, "r")
lchemCluster2 = fclust2.readlines()
fclust2.close()
for chemCluster2 in lchemCluster2[1:]:
chemCluster2 = chemCluster2.strip().replace("\"", "").split(",")
chemID = chemCluster2[0]
clust2 = chemCluster2[1]
dclust[chemID].append(clust2)
#write main cluster
for chemID in dclust.keys():
if len(dclust[chemID]) == 1:
dclust[chemID].append("1")
fclustersFinal.write(str(chemID) + "," + ",".join(dclust[chemID]) + "\n")
fclustersFinal.close()
self.pclusters = pclustersFinal
else:
self.pclusters = pcluster
def prepareActiveMatrix(self,
corval,
maxQuantile,
NBNA,
pAC50All,
prout,
luciferase=0):
self.corval = corval
self.maxQuantile = maxQuantile
pdescAct = prout + "descActive"
pAC50Act = prout + "AC50Active"
if path.exists(
pdescAct) and path.getsize(pdescAct) > 10 and path.exists(
pAC50Act) and path.getsize(pAC50Act) > 10:
lpdescActClean = runExternalSoft.dataManager(
pdescAct, pAC50Act, corval, maxQuantile, NBNA, prout)
self.pdescCleanActive = lpdescActClean[0]
self.pAC50AllActive = lpdescActClean[1]
return [self.pdescCleanActive, self.pAC50AllActive]
ddesc = toolbox.loadMatrix(self.pdesc1D2D)
dAC50All = toolbox.loadMatrix(pAC50All)
if luciferase == 0:
i = 0
imax = len(ddesc.keys())
while i < imax:
casID = dAC50All.keys()[i]
nbNA = 0
for kAC50 in dAC50All[casID].keys():
if kAC50 == "CASID" or kAC50 == "Luc_IC50": # not considered luciferase
continue
else:
if dAC50All[casID][kAC50] == "NA":
nbNA += 1
#print nbNA, len(dAC50All[casID].keys())
if nbNA == (len(dAC50All[casID].keys()) - 2):
del dAC50All[casID]
try:
del ddesc[casID]
except:
pass
imax = imax - 1
else:
i += 1
toolbox.writeMatrix(ddesc, pdescAct)
toolbox.writeMatrix(dAC50All, pAC50Act)
lpdescActClean = runExternalSoft.dataManager(
pdescAct, pAC50Act, corval, maxQuantile, NBNA, prout)
self.pdescCleanActive = lpdescActClean[0]
self.pAC50AllActive = lpdescActClean[1]
return [self.pdescCleanActive, self.pAC50AllActive]
else:
i = 0
imax = len(dAC50All.keys())
while i < imax:
casID = dAC50All.keys()[i]
if not casID in ddesc.keys():
del dAC50All[casID]
imax = imax - 1
i = i - 1
continue
for kAC50 in dAC50All[casID].keys():
if kAC50 != "Luc_IC50" and kAC50 != "CASID": # not considered luciferase
del dAC50All[casID][kAC50]
else:
if dAC50All[casID][kAC50] == "NA":
del dAC50All[casID]
try:
del ddesc[casID]
except:
pass
imax = imax - 1
i = i - 1
break
i += 1
toolbox.writeMatrix(ddesc, pdescAct)
toolbox.writeMatrix(dAC50All, pAC50Act)
lpdescActClean = runExternalSoft.dataManager(
pdescAct, pAC50Act, corval, maxQuantile, NBNA, prout)
self.pdescCleanActive = lpdescActClean[0]
self.pAC50AllActive = lpdescActClean[1]
return [self.pdescCleanActive, self.pAC50AllActive]