def analyseFile(file, envPMIDs):
allpmids = defaultdict(set)
print("Starting file: ", file)
procDB = neo4jInterface(simulate=False, printQueries=True)
with open(file, 'r') as infile:
for line in infile:
aline = line.strip().split('\t')
pmid_cites = aline[0]
pmid_cited_by = aline[1]
if pmid_cites in envPMIDs and pmid_cited_by in envPMIDs:
allpmids[pmid_cites].add(pmid_cited_by)
if len(allpmids) > 0:
for pmid in allpmids:
for opmid in allpmids[pmid]:
if not opmid in envPMIDs:
continue
procDB.createRelationship('cpmid', ['PUBMED'], {'id': pmid},
'opmid', ['PUBMED'], {'id': opmid},
['PUBMED_CITED_BY'], None)
procDB.close()
def analyseFile(splitFileID, relPMIDs):
fileID = "{:>4}".format(splitFileID).replace(" ", "0")
diseaseHitsFile = resultBase + "/disease/medline17n" + fileID + ".index"
hitsFile = SyngrepHitFile(diseaseHitsFile, diseaseMap)
if len(hitsFile) == 0:
return
print("Document: " + str(fileID))
print("Start Document: " + str(fileID))
procDB = neo4jInterface(simulate=False, printQueries=False)
for docID in hitsFile:
if not docID in relPMIDs:
continue
synHits = hitsFile.getHitsForDocument(docID)
foundUniqueHits = set()
for hit in synHits:
if len(hit.foundSyn) < 5:
if not hit.perfectHit:
continue
hitSyn = hit.synonym
foundUniqueHits.add(hitSyn.id.replace('_', ':'))
for synonymID in foundUniqueHits:
pubmedExists = False
if addUnknownPubmeds:
procDB.createNodeIfNotExists(['EVIDENCE', 'PUBMED'],
{'id': docID})
pubmedExists = True
else:
if procDB.nodeExists(['PUBMED'], {'id': docID}):
pubmedExists = True
if pubmedExists:
res = procDB.createRelationship('disease', ['DISEASE'],
{'id': synonymID}, 'pubmed',
['PUBMED'], {'id': docID},
['DISEASE_MENTION'], None)
print("Add: ", fileID, docID, synonymID,
[x for x in res if res != None])
print("End Document: " + str(fileID))
procDB.close()
def __init__(self,
chemokines=list(),
db=neo4jInterface(simulate=False, printQueries=False)):
self.chemokines = list(chemokines)
self.db = db
self.all_nodes = dict()
self.all_relation_ids = set()
self.all_relations = dict()
self.simple_connections = set()
if self.chemokines != None and len(self.chemokines) > 0:
for chemokine in self.chemokines:
self._queryDB(chemokine)
def addMIRTs(mirtarbaseEvidences, mirtarEvs):
print("Starting MIRTS: ", len(mirtarbaseEvidences[0]))
dbcreatedMIRT2TAX = defaultdict(set)
dbcreatedMIRT2PUBMED = defaultdict(set)
dbcreatedMIRT2ExpTypes = defaultdict(set)
dbcreatedMIRT2SupportTypes = defaultdict(set)
dbcreatedMIRT2MIRNA = defaultdict(set)
dbcreatedMIRT2GENE = defaultdict(set)
dbcreatedMIRTIDs = set()
procDB = neo4jInterface(simulate=False, printQueries=False)
for mirnaEvidence in mirtarbaseEvidences[0]:
# Species (miRNA) Target Gene Target Gene (Entrez Gene ID) Species (Target Gene) Experiments Support Type References (PMID)
mirtarID = mirnaEvidence['miRTarBase ID']
mirtarMIRNA = mirnaEvidence['miRNA']
mirtarMIRNASpecies = mirnaEvidence['Species (miRNA)']
mirtarGENE = mirnaEvidence['Target Gene'].upper()
mirtarGENESpecies = mirnaEvidence['Species (Target Gene)']
mirtarRefs = mirnaEvidence['References (PMID)']
"""
EXPeriment and Functional Type
"""
mirtarExperiment = mirnaEvidence['Experiments']
mirtarExperiment = mirtarExperiment.split("/") if mirtarExperiment != None else []
mirtarExperimentNew = []
for x in [y.split(";") for y in mirtarExperiment if len(y) > 0]:
for elem in x:
mirtarExperimentNew.append(mirtarbase_exp_type(elem))
mirtarExperiment = mirtarExperimentNew
mirtarSupport = mirtarbase_function_label(mirnaEvidence['Support Type'])
mirnaSpeciesID = speciesName2TaxID.get(mirtarMIRNASpecies, None)
geneSpeciesID = speciesName2TaxID.get(mirtarGENESpecies, None)
if mirnaSpeciesID == None and geneSpeciesID == None:
continue
if mirnaSpeciesID == geneSpeciesID:
commonTaxID = mirnaSpeciesID
else:
commonTaxID = None
procDB.createNodeIfNotExists(['PUBMED', 'EVIDENCE'], {'id': mirtarRefs}, 'n', ['PUBMED'], {'id': mirtarRefs})
if not mirtarID in dbcreatedMIRTIDs:
dbcreatedMIRTIDs.add(mirtarID)
procDB.createNode(['MIRTARBASE', 'EVIDENCE'], {'id': mirtarID, 'tax_gene': geneSpeciesID, 'tax_mirna': mirnaSpeciesID})
if not mirtarSupport in dbcreatedMIRT2SupportTypes[mirtarID]:
dbcreatedMIRT2SupportTypes[mirtarID].add(mirtarSupport)
procDB.createRelationship('ms', ['MIRTARBASE_SUPPORT'], {'id': mirtarSupport}, 'mtb', ['MIRTARBASE'], {'id': mirtarID}, ['MIRTARBASE_FUNCTIONAL_SUPPORT'], None)
for expType in mirtarExperiment:
if not expType in dbcreatedMIRT2ExpTypes[mirtarID]:
dbcreatedMIRT2ExpTypes[mirtarID].add(expType)
procDB.createRelationship('me', ['MIRTARBASE_EXPERIMENT'], {'id': expType}, 'mtb', ['MIRTARBASE'], {'id': mirtarID}, ['MIRTARBASE_EXPERIMENT_SUPPORT'], None)
# TODO add relation props?
if not mirtarRefs in dbcreatedMIRT2PUBMED[mirtarID]:
dbcreatedMIRT2PUBMED[mirtarID].add(mirtarRefs)
procDB.createRelationship('pb', ['PUBMED'], {'id': mirtarRefs}, 'mtb', ['MIRTARBASE'], {'id': mirtarID}, ['MIRTARBASE_LITERATURE_SUPPORT'], {})
if commonTaxID != None:
if not mirnaSpeciesID in dbcreatedMIRT2TAX[mirtarID]:
dbcreatedMIRT2TAX[mirtarID].add(mirnaSpeciesID)
procDB.createRelationship('mtb', ['MIRTARBASE'], {'id': mirtarID}, 'taxid', ['TAX'], {'id': mirnaSpeciesID}, ['ORGANISM_SUPPORT'], {})
if not mirtarGENE in dbcreatedMIRT2GENE[mirtarID]:
dbcreatedMIRT2GENE[mirtarID].add(mirtarGENE)
procDB.createRelationship('gene', ['GENE'], {'id': mirtarGENE}, 'mtb', ['MIRTARBASE'], {'id': mirtarID}, ['GENE_MENTION'], {'tax': geneSpeciesID})
if not mirtarMIRNA in dbcreatedMIRT2MIRNA[mirtarID]:
dbcreatedMIRT2MIRNA[mirtarID].add(mirtarMIRNA)
procDB.createRelationship('mtb', ['MIRTARBASE'], {'id': mirtarID}, 'mirna', ['MIRNA'],
{'name': mirtarMIRNA}, ['MIRNA_MENTION'], {'tax': mirnaSpeciesID})
procDB.close()
from porestat.utils.DataFrame import DataFrame
from utils.idutils import ltype2label, makeDBGeneID, mirtarbase_exp_type, mirtarbase_function_label, speciesName2TaxID, \
dataDir
from database.Neo4JInterface import neo4jInterface
from utils.parallel import MapReduce
mirtarbaseEvidences = DataFrame.parseFromFile(dataDir + "/miRExplore/miRTarBase.csv", bConvertTextToNumber=False)
print(mirtarbaseEvidences.getHeader())
experimentTypes = Counter()
supportTypes = Counter()
referencesWithComma = Counter()
db = neo4jInterface(simulate=False, printQueries=False)
db.deleteRelationship('n', ['GENE'], None, 'm', ['MIRTARBASE'], None, ['GENE_MENTION'], None, 'r')
db.deleteRelationship('n', ['MIRTARBASE'], None, 'm', ['MIRNA'], None, ['MIRNA_MENTION'], None, 'r')
db.deleteRelationship('n', ['MIRTARBASE'], None, 'm', ['PUBMED'], None, ['MIRTARBASE_LITERATURE_SUPPORT'], None, 'r')
db.deleteRelationship('n', ['MIRTARBASE_SUPPORT'], None, 'm', ['MIRTARBASE'], None, ['MIRTARBASE_FUNCTIONAL_SUPPORT'], None, 'r')
db.deleteRelationship('n', ['MIRTARBASE_EXPERIMENT'], None, 'm', ['MIRTARBASE'], None, ['MIRTARBASE_EXPERIMENT_SUPPORT'], None, 'r')
db.deleteRelationship('n', ['MIRTARBASE'], None, 'm', ['TAX'], None, ['ORGANISM_SUPPORT'], None, 'r')
db.deleteNode(["MIRTARBASE"], None)
db.deleteNode(["MIRTARBASE_SUPPORT"], None)
db.deleteNode(["MIRTARBASE_EXPERIMENT"], None)
db.createUniquenessConstraint('MIRTARBASE', 'id')
if False:
db.close()
exit(0)
import os
from mjoppich.geneontology import GeneOntology
from porestat.utils.Parallel import MapReduce
from database.Neo4JInterface import neo4jInterface
from synonymes.SynfileMap import SynfileMap
from textmining.SyngrepHitFile import SyngrepHitFile
from utils.idutils import dataDir, eprint
resultBase = dataDir + "/miRExplore/textmine/results/"
diseaseMap = SynfileMap(resultBase + "/disease/synfile.map")
diseaseMap.loadSynFiles(('/home/users/joppich/ownCloud/data/', dataDir))
diseaseObo = GeneOntology(dataDir + "miRExplore/doid.obo")
db = neo4jInterface(simulate=False)
db.deleteRelationship('n', ['DISEASE'], None, 'm', ['PUBMED'], None,
['DISEASE_MENTION'], None)
allfiles = glob.glob(resultBase + "/hgnc/medline17n*.index")
allfileIDs = [
int(os.path.basename(x).replace('medline17n', '').replace('.index', ''))
for x in allfiles
]
allfileIDs = sorted(allfileIDs, reverse=True)
addUnknownPubmeds = False
retVal = db.matchNodes(['PUBMED'], None, nodename='n')
relevantPMIDs = set()
def analyseFile(splitFileID, relPMIDs):
fileID = "{:>4}".format(splitFileID).replace(" ", "0")
diseaseHitsFile = resultBase + "/cellline/medline17n" + fileID + ".index"
hitsFile = SyngrepHitFile(diseaseHitsFile, celllinesMap)
if len(hitsFile) == 0:
return
print("Start Document: " + str(fileID))
procDB = neo4jInterface(simulate=False, printQueries=False)
for docID in hitsFile:
if not docID in relPMIDs:
continue
synHits = hitsFile.getHitsForDocument(docID)
foundUniqueHits = set()
foundOrgs = set()
for hit in synHits:
if len(hit.foundSyn) < 5:
if not hit.perfectHit:
continue
hitSynFileID = hit.synonymID.synfile
foundOrgs.add(synfileID2tax[hitSynFileID])
hitSyn = hit.synonym
foundUniqueHits.add(hitSyn.id)
if len(foundUniqueHits) == 0:
continue
for celllineID in foundUniqueHits:
pubmedExists = False
if addUnknownPubmeds:
procDB.createNodeIfNotExists(['EVIDENCE', 'PUBMED'],
{'id': docID})
pubmedExists = True
else:
if procDB.nodeExists(['PUBMED'], {'id': docID}):
pubmedExists = True
if pubmedExists:
res = procDB.createRelationship('cellline', ['CELLLINE'],
{'id': celllineID}, 'pubmed',
['PUBMED'], {'id': docID},
['CELLLINE_MENTION'], None)
print("Add: ", fileID, docID, celllineID,
[x for x in res if res != None])
foundOrgs = foundOrgs.difference(allSet)
if len(foundOrgs) == 1:
pass
# create relation
# print('Associate: ' + str(foundOrgs))
elif len(foundOrgs) == 0:
pass
elif len(foundOrgs) > 1:
# print('Ambiguous pubmed: ' + docID)
pass
print("End Document: " + str(fileID))
procDB.close()