def __init__(self, dataDir):
super(FileSystemBackend, self).__init__()
self._dataDir = dataDir
# TODO this code is very ugly and should be regarded as a temporary
# stop-gap until we deal with iterating over the data tree properly.
# References
referencesDirName = "references"
referenceSetDir = os.path.join(self._dataDir, referencesDirName)
for referenceSetName in os.listdir(referenceSetDir):
relativePath = os.path.join(referenceSetDir, referenceSetName)
if os.path.isdir(relativePath):
referenceSet = references.HtslibReferenceSet(
referenceSetName, relativePath, self)
self.addReferenceSet(referenceSet)
# Datasets
datasetDirs = [
os.path.join(self._dataDir, directory)
for directory in os.listdir(self._dataDir)
if os.path.isdir(os.path.join(self._dataDir, directory))
and directory != referencesDirName
]
for datasetDir in datasetDirs:
dataset = datasets.FileSystemDataset(datasetDir, self)
self.addDataset(dataset)
def __init__(self, dataDir):
super(FileSystemBackend, self).__init__()
self._dataDir = dataDir
# TODO this code is very ugly and should be regarded as a temporary
# stop-gap until we deal with iterating over the data tree properly.
# References
referencesDirName = "references"
referenceSetDir = os.path.join(self._dataDir, referencesDirName)
for referenceSetId in os.listdir(referenceSetDir):
relativePath = os.path.join(referenceSetDir, referenceSetId)
if os.path.isdir(relativePath):
referenceSet = references.HtslibReferenceSet(
referenceSetId, relativePath)
self._referenceSetIdMap[referenceSetId] = referenceSet
for reference in referenceSet.getReferences():
referenceId = reference.getId()
self._referenceIdMap[referenceId] = reference
self._referenceSetIds = sorted(self._referenceSetIdMap.keys())
self._referenceIds = sorted(self._referenceIdMap.keys())
# Datasets
datasetDirs = [
os.path.join(self._dataDir, directory)
for directory in os.listdir(self._dataDir)
if os.path.isdir(os.path.join(self._dataDir, directory)) and
directory != referencesDirName]
for datasetDir in datasetDirs:
dataset = datasets.FileSystemDataset(datasetDir)
self._datasetIdMap[dataset.getId()] = dataset
self._datasetIds = sorted(self._datasetIdMap.keys())
def _readReferenceSetTable(self, cursor):
cursor.row_factory = sqlite3.Row
cursor.execute("SELECT * FROM ReferenceSet;")
for row in cursor:
referenceSet = references.HtslibReferenceSet(row[b'name'])
referenceSet.populateFromRow(row)
assert referenceSet.getId() == row[b"id"]
# Insert the referenceSet into the memory-based object model.
self.addReferenceSet(referenceSet)
def createRepo(self):
"""
Creates the repository for all the data we've just downloaded.
"""
repo = datarepo.SqlDataRepository(self.repoPath)
repo.open("w")
repo.initialise()
referenceSet = references.HtslibReferenceSet("GRCh37-subset")
referenceSet.populateFromFile(self.fastaFilePath)
referenceSet.setDescription("Subset of GRCh37 used for demonstration")
referenceSet.setNcbiTaxonId(9606)
for reference in referenceSet.getReferences():
reference.setNcbiTaxonId(9606)
reference.setSourceAccessions(
self.accessions[reference.getName()] + ".subset")
repo.insertReferenceSet(referenceSet)
dataset = datasets.Dataset("1kg-p3-subset")
dataset.setDescription("Sample data from 1000 Genomes phase 3")
repo.insertDataset(dataset)
variantSet = variants.HtslibVariantSet(dataset, "mvncall")
variantSet.setReferenceSet(referenceSet)
dataUrls = [vcfFile for vcfFile, _ in self.vcfFilePaths]
indexFiles = [indexFile for _, indexFile in self.vcfFilePaths]
variantSet.populateFromFile(dataUrls, indexFiles)
variantSet.checkConsistency()
repo.insertVariantSet(variantSet)
for sample, (bamFile, indexFile) in zip(self.samples,
self.bamFilePaths):
readGroupSet = reads.HtslibReadGroupSet(dataset, sample)
readGroupSet.populateFromFile(bamFile, indexFile)
readGroupSet.setReferenceSet(referenceSet)
repo.insertReadGroupSet(readGroupSet)
repo.commit()
repo.close()
self.log("Finished creating the repository; summary:\n")
repo.open("r")
repo.printSummary()
def addReferenceSet(self):
"""
Adds a new reference set into this repo.
"""
self._openRepo()
name = self._args.name
filePath = self._getFilePath(self._args.filePath,
self._args.relativePath)
if name is None:
name = getNameFromPath(self._args.filePath)
referenceSet = references.HtslibReferenceSet(name)
referenceSet.populateFromFile(filePath)
referenceSet.setDescription(self._args.description)
referenceSet.setNcbiTaxonId(self._args.ncbiTaxonId)
referenceSet.setIsDerived(self._args.isDerived)
referenceSet.setAssemblyId(self._args.assemblyId)
sourceAccessions = []
if self._args.sourceAccessions is not None:
sourceAccessions = self._args.sourceAccessions.split(",")
referenceSet.setSourceAccessions(sourceAccessions)
referenceSet.setSourceUri(self._args.sourceUri)
self._updateRepo(self._repo.insertReferenceSet, referenceSet)
def testMissingReferenceSetMetadata(self):
localId = "invalid_refset_meta"
path = self.getFullPath(localId)
with self.assertRaises(ValueError):
references.HtslibReferenceSet(localId, path, None)
def testMissingReferenceSetMetadata(self):
localId = "missing_refset_meta"
path = self.getFullPath(localId)
with self.assertRaises(exceptions.MissingReferenceSetMetadata):
references.HtslibReferenceSet(localId, path, None)
def testNoReferenceSetMetadata(self):
localId = "no_refset_meta"
path = self.getFullPath(localId)
with self.assertRaises(IOError):
references.HtslibReferenceSet(localId, path, None)
def run(self):
if not os.path.exists(self.outputDirectory):
os.makedirs(self.outputDirectory)
self.repo.open("w")
self.repo.initialise()
referenceFileName = "ref_brca1.fa"
inputRef = os.path.join(
self.inputDirectory, referenceFileName)
outputRef = os.path.join(
self.outputDirectory, referenceFileName)
shutil.copy(inputRef, outputRef)
fastaFilePath = os.path.join(
self.outputDirectory,
referenceFileName + '.gz')
pysam.tabix_compress(
outputRef, fastaFilePath)
with open(
os.path.join(
self.inputDirectory, "ref_brca1.json")) as refMetadataFile:
refMetadata = json.load(refMetadataFile)
with open(
os.path.join(
self.inputDirectory,
"referenceset_hg37.json")) as refMetadataFile:
refSetMetadata = json.load(refMetadataFile)
referenceSet = references.HtslibReferenceSet(
refSetMetadata['assemblyId'])
referenceSet.populateFromFile(fastaFilePath)
referenceSet.setAssemblyId(refSetMetadata['assemblyId'])
referenceSet.setDescription(refSetMetadata['description'])
referenceSet.setNcbiTaxonId(refSetMetadata['ncbiTaxonId'])
referenceSet.setIsDerived(refSetMetadata['isDerived'])
referenceSet.setSourceUri(refSetMetadata['sourceUri'])
referenceSet.setSourceAccessions(refSetMetadata['sourceAccessions'])
for reference in referenceSet.getReferences():
reference.setNcbiTaxonId(refMetadata['ncbiTaxonId'])
reference.setSourceAccessions(
refMetadata['sourceAccessions'])
self.repo.insertReferenceSet(referenceSet)
dataset = datasets.Dataset("brca1")
self.repo.insertDataset(dataset)
hg00096Individual = biodata.Individual(dataset, "HG00096")
with open(
os.path.join(
self.inputDirectory,
"individual_HG00096.json")) as jsonString:
hg00096Individual.populateFromJson(jsonString.read())
self.repo.insertIndividual(hg00096Individual)
hg00096BioSample = biodata.BioSample(dataset, "HG00096")
with open(
os.path.join(
self.inputDirectory,
"bioSample_HG00096.json")) as jsonString:
hg00096BioSample.populateFromJson(jsonString.read())
hg00096BioSample.setIndividualId(hg00096Individual.getId())
self.repo.insertBioSample(hg00096BioSample)
hg00099Individual = biodata.Individual(dataset, "HG00099")
with open(
os.path.join(
self.inputDirectory,
"individual_HG00099.json")) as jsonString:
hg00099Individual.populateFromJson(jsonString.read())
self.repo.insertIndividual(hg00099Individual)
hg00099BioSample = biodata.BioSample(dataset, "HG00099")
with open(
os.path.join(
self.inputDirectory,
"bioSample_HG00099.json")) as jsonString:
hg00099BioSample.populateFromJson(jsonString.read())
hg00099BioSample.setIndividualId(hg00099Individual.getId())
self.repo.insertBioSample(hg00099BioSample)
hg00101Individual = biodata.Individual(dataset, "HG00101")
with open(
os.path.join(
self.inputDirectory,
"individual_HG00101.json")) as jsonString:
hg00101Individual.populateFromJson(jsonString.read())
self.repo.insertIndividual(hg00101Individual)
hg00101BioSample = biodata.BioSample(dataset, "HG00101")
with open(
os.path.join(
self.inputDirectory,
"bioSample_HG00101.json")) as jsonString:
hg00101BioSample.populateFromJson(jsonString.read())
hg00101BioSample.setIndividualId(hg00101Individual.getId())
self.repo.insertBioSample(hg00101BioSample)
readFiles = [
"brca1_HG00096.sam",
"brca1_HG00099.sam",
"brca1_HG00101.sam"]
for readFile in readFiles:
name = readFile.split('_')[1].split('.')[0]
readSrc = pysam.AlignmentFile(
os.path.join(self.inputDirectory, readFile), "r")
readDest = pysam.AlignmentFile(
os.path.join(
self.outputDirectory,
name + ".bam"),
"wb", header=readSrc.header)
destFilePath = readDest.filename
for readData in readSrc:
readDest.write(readData)
readDest.close()
readSrc.close()
pysam.index(destFilePath)
readGroupSet = reads.HtslibReadGroupSet(dataset, name)
readGroupSet.populateFromFile(destFilePath, destFilePath + ".bai")
readGroupSet.setReferenceSet(referenceSet)
bioSamples = [hg00096BioSample, hg00099BioSample, hg00101BioSample]
for readGroup in readGroupSet.getReadGroups():
for bioSample in bioSamples:
if bioSample.getLocalId() == readGroup.getSampleName():
readGroup.setBioSampleId(bioSample.getId())
self.repo.insertReadGroupSet(readGroupSet)
ontologyMapFileName = "so-xp-simple.obo"
inputOntologyMap = os.path.join(
self.inputDirectory, ontologyMapFileName)
outputOntologyMap = os.path.join(
self.outputDirectory, ontologyMapFileName)
shutil.copy(inputOntologyMap, outputOntologyMap)
sequenceOntology = ontologies.Ontology("so-xp-simple")
sequenceOntology.populateFromFile(outputOntologyMap)
sequenceOntology._id = "so-xp-simple"
self.repo.insertOntology(sequenceOntology)
self.repo.addOntology(sequenceOntology)
vcfFiles = [
"brca1_1kgPhase3_variants.vcf",
"brca1_WASH7P_annotation.vcf",
"brca1_OR4F_annotation.vcf"]
for vcfFile in vcfFiles:
self.addVariantSet(
vcfFile,
dataset,
referenceSet,
sequenceOntology,
bioSamples)
seqAnnFile = "brca1_gencodev19.gff3"
seqAnnSrc = os.path.join(self.inputDirectory, seqAnnFile)
seqAnnDest = os.path.join(self.outputDirectory, "gencodev19.db")
dbgen = generate_gff3_db.Gff32Db(seqAnnSrc, seqAnnDest)
dbgen.run()
gencode = sequenceAnnotations.Gff3DbFeatureSet(dataset, "gencodev19")
gencode.setOntology(sequenceOntology)
gencode.populateFromFile(seqAnnDest)
gencode.setReferenceSet(referenceSet)
self.repo.insertFeatureSet(gencode)
self.repo.commit()
print("Done converting compliance data.", file=sys.stderr)
def testInstantiation(self):
for referenceSetId in self.setIds:
path = self.getFullPath(referenceSetId)
with self.assertRaises(exceptions.NotExactlyOneReferenceException):
references.HtslibReferenceSet(referenceSetId, path)
def getDataModelInstance(self, localId, dataPath):
return references.HtslibReferenceSet(localId, dataPath, None)
def run(self):
if not os.path.exists(self.outputDirectory):
os.makedirs(self.outputDirectory)
self.repo.open("w")
self.repo.initialise()
referenceFileName = "ref_brca1.fa"
inputRef = os.path.join(self.inputDirectory, referenceFileName)
outputRef = os.path.join(self.outputDirectory, referenceFileName)
shutil.copy(inputRef, outputRef)
fastaFilePath = os.path.join(self.outputDirectory,
referenceFileName + '.gz')
pysam.tabix_compress(outputRef, fastaFilePath)
with open(os.path.join(self.inputDirectory,
"ref_brca1.json")) as refMetadataFile:
refMetadata = json.load(refMetadataFile)
with open(os.path.join(self.inputDirectory,
"referenceset_hg37.json")) as refMetadataFile:
refSetMetadata = json.load(refMetadataFile)
referenceSet = references.HtslibReferenceSet(
refSetMetadata['assemblyId'])
referenceSet.populateFromFile(fastaFilePath)
referenceSet.setAssemblyId(refSetMetadata['assemblyId'])
referenceSet.setDescription(refSetMetadata['description'])
referenceSet.setNcbiTaxonId(refSetMetadata['ncbiTaxonId'])
referenceSet.setIsDerived(refSetMetadata['isDerived'])
referenceSet.setSourceUri(refSetMetadata['sourceUri'])
referenceSet.setSourceAccessions(refSetMetadata['sourceAccessions'])
for reference in referenceSet.getReferences():
reference.setNcbiTaxonId(refMetadata['ncbiTaxonId'])
reference.setSourceAccessions(refMetadata['sourceAccessions'])
self.repo.insertReferenceSet(referenceSet)
dataset = datasets.Dataset("brca1")
# Some info is set, it isn't important what
dataset.setInfo({"version": ga4gh.__version__})
self.repo.insertDataset(dataset)
hg00096Individual = biodata.Individual(dataset, "HG00096")
with open(os.path.join(self.inputDirectory,
"individual_HG00096.json")) as jsonString:
hg00096Individual.populateFromJson(jsonString.read())
self.repo.insertIndividual(hg00096Individual)
hg00096BioSample = biodata.BioSample(dataset, "HG00096")
with open(os.path.join(self.inputDirectory,
"bioSample_HG00096.json")) as jsonString:
hg00096BioSample.populateFromJson(jsonString.read())
hg00096BioSample.setIndividualId(hg00096Individual.getId())
self.repo.insertBioSample(hg00096BioSample)
hg00099Individual = biodata.Individual(dataset, "HG00099")
with open(os.path.join(self.inputDirectory,
"individual_HG00099.json")) as jsonString:
hg00099Individual.populateFromJson(jsonString.read())
self.repo.insertIndividual(hg00099Individual)
hg00099BioSample = biodata.BioSample(dataset, "HG00099")
with open(os.path.join(self.inputDirectory,
"bioSample_HG00099.json")) as jsonString:
hg00099BioSample.populateFromJson(jsonString.read())
hg00099BioSample.setIndividualId(hg00099Individual.getId())
self.repo.insertBioSample(hg00099BioSample)
hg00101Individual = biodata.Individual(dataset, "HG00101")
with open(os.path.join(self.inputDirectory,
"individual_HG00101.json")) as jsonString:
hg00101Individual.populateFromJson(jsonString.read())
self.repo.insertIndividual(hg00101Individual)
hg00101BioSample = biodata.BioSample(dataset, "HG00101")
with open(os.path.join(self.inputDirectory,
"bioSample_HG00101.json")) as jsonString:
hg00101BioSample.populateFromJson(jsonString.read())
hg00101BioSample.setIndividualId(hg00101Individual.getId())
self.repo.insertBioSample(hg00101BioSample)
readFiles = [
"brca1_HG00096.sam", "brca1_HG00099.sam", "brca1_HG00101.sam"
]
for readFile in readFiles:
name = readFile.split('_')[1].split('.')[0]
readSrc = pysam.AlignmentFile(
os.path.join(self.inputDirectory, readFile), "r")
readDest = pysam.AlignmentFile(os.path.join(
self.outputDirectory, name + ".bam"),
"wb",
header=readSrc.header)
destFilePath = readDest.filename
for readData in readSrc:
readDest.write(readData)
readDest.close()
readSrc.close()
pysam.index(destFilePath)
readGroupSet = reads.HtslibReadGroupSet(dataset, name)
readGroupSet.populateFromFile(destFilePath, destFilePath + ".bai")
readGroupSet.setReferenceSet(referenceSet)
dataset.addReadGroupSet(readGroupSet)
bioSamples = [hg00096BioSample, hg00099BioSample, hg00101BioSample]
for readGroup in readGroupSet.getReadGroups():
for bioSample in bioSamples:
if bioSample.getLocalId() == readGroup.getSampleName():
readGroup.setBioSampleId(bioSample.getId())
self.repo.insertReadGroupSet(readGroupSet)
ontologyMapFileName = "so-xp-simple.obo"
inputOntologyMap = os.path.join(self.inputDirectory,
ontologyMapFileName)
outputOntologyMap = os.path.join(self.outputDirectory,
ontologyMapFileName)
shutil.copy(inputOntologyMap, outputOntologyMap)
sequenceOntology = ontologies.Ontology("so-xp-simple")
sequenceOntology.populateFromFile(outputOntologyMap)
sequenceOntology._id = "so-xp-simple"
self.repo.insertOntology(sequenceOntology)
self.repo.addOntology(sequenceOntology)
vcfFiles = [
"brca1_1kgPhase3_variants.vcf", "brca1_WASH7P_annotation.vcf",
"brca1_OR4F_annotation.vcf"
]
for vcfFile in vcfFiles:
self.addVariantSet(vcfFile, dataset, referenceSet,
sequenceOntology, bioSamples)
# Sequence annotations
seqAnnFile = "brca1_gencodev19.gff3"
seqAnnSrc = os.path.join(self.inputDirectory, seqAnnFile)
seqAnnDest = os.path.join(self.outputDirectory, "gencodev19.db")
dbgen = generate_gff3_db.Gff32Db(seqAnnSrc, seqAnnDest)
dbgen.run()
gencode = sequence_annotations.Gff3DbFeatureSet(dataset, "gencodev19")
gencode.setOntology(sequenceOntology)
gencode.populateFromFile(seqAnnDest)
gencode.setReferenceSet(referenceSet)
self.repo.insertFeatureSet(gencode)
# add g2p featureSet
g2pPath = os.path.join(self.inputDirectory, "cgd")
# copy all files input directory to output path
outputG2PPath = os.path.join(self.outputDirectory, "cgd")
os.makedirs(outputG2PPath)
for filename in glob.glob(os.path.join(g2pPath, '*.*')):
shutil.copy(filename, outputG2PPath)
featuresetG2P = g2p_featureset.PhenotypeAssociationFeatureSet(
dataset, outputG2PPath)
featuresetG2P.setOntology(sequenceOntology)
featuresetG2P.setReferenceSet(referenceSet)
featuresetG2P.populateFromFile(outputG2PPath)
self.repo.insertFeatureSet(featuresetG2P)
# add g2p phenotypeAssociationSet
phenotypeAssociationSet = g2p_associationset\
.RdfPhenotypeAssociationSet(dataset, "cgd", outputG2PPath)
self.repo.insertPhenotypeAssociationSet(phenotypeAssociationSet)
self.repo.commit()
dataset.addFeatureSet(gencode)
# RNA Quantification
rnaDbName = os.path.join(self.outputDirectory, "rnaseq.db")
store = rnaseq2ga.RnaSqliteStore(rnaDbName)
store.createTables()
rnaseq2ga.rnaseq2ga(self.inputDirectory + "/rna_brca1.tsv",
rnaDbName,
"rna_brca1.tsv",
"rsem",
featureType="transcript",
readGroupSetNames="HG00096",
featureSetNames="gencodev19",
dataset=dataset)
rnaQuantificationSet = rna_quantification.SqliteRnaQuantificationSet(
dataset, "rnaseq")
rnaQuantificationSet.setReferenceSet(referenceSet)
rnaQuantificationSet.populateFromFile(rnaDbName)
self.repo.insertRnaQuantificationSet(rnaQuantificationSet)
self.repo.commit()
def getDataModelInstance(self, localId, dataPath):
referenceSet = references.HtslibReferenceSet(localId)
referenceSet.populateFromFile(dataPath)
return referenceSet
