def treeQA(self, options):
"""QA command"""
self.logger.info(
'[CheckM - tree_qa] Assessing phylogenetic markers found in each bin.'
)
checkDirExists(options.tree_dir)
# set HMM file for each bin
markerSetParser = MarkerSetParser()
hmmModelInfoFile = os.path.join(options.tree_dir, 'storage',
DefaultValues.PHYLO_HMM_MODEL_INFO)
binIdToModels = markerSetParser.loadBinModels(hmmModelInfoFile)
# calculate marker gene statistics
RP = ResultsParser(binIdToModels)
binStats = RP.analyseResults(options.tree_dir,
DefaultValues.BIN_STATS_PHYLO_OUT,
DefaultValues.HMMER_TABLE_PHYLO_OUT)
# determine taxonomy of each bin
treeParser = TreeParser()
treeParser.printSummary(options.out_format, options.tree_dir, RP,
options.bTabTable, options.file, binStats)
if options.file != '':
self.logger.info('QA information written to: ' + options.file)
self.timeKeeper.printTimeStamp()
def run(self):
# read all taxonomic-specific marker genes
print('Reading taxonomic-specific marker genes.')
taxonomicMarkers = set()
taxonParser = TaxonParser()
taxonMarkerSets = taxonParser.readMarkerSets()
for _, taxa in taxonMarkerSets.items():
for _, markerSet in taxa.items():
taxonomicMarkers = taxonomicMarkers.union(
markerSet.getMarkerGenes())
print(' Taxonomic-specific marker genes: %d' % len(taxonomicMarkers))
# read all lineage-specific marker genes
print('Reading lineage-specific marker genes.')
lineageMarkers = set()
treeParser = TreeParser()
uniqueIdToLineageStatistics = treeParser.readNodeMetadata()
for uniqueId, d in uniqueIdToLineageStatistics.items():
markerSet = MarkerSet(uniqueId, 'NA', int(d['# genomes']),
eval(d['marker set']))
lineageMarkers = lineageMarkers.union(markerSet.getMarkerGenes())
print(' Lineage-specific marker genes: %d' % len(lineageMarkers))
# gather all marker genes
markerGenes = taxonomicMarkers.union(lineageMarkers)
print(' Total marker genes: %d' % len(markerGenes))
# get genes from same clan as marker genes
print('Gathering HMMs from the same clan as marker genes.')
pfam = PFAM()
genesInSameClan = pfam.genesInSameClan(markerGenes)
allMarkers = markerGenes.union(genesInSameClan)
# create file with all model accession numbers
keyFile = os.path.join(tempfile.gettempdir(), str(uuid.uuid4()))
fout = open(keyFile, 'w')
for modelAcc in allMarkers:
fout.write(modelAcc + '\n')
fout.close()
# fetch specified models
HF = HMMERRunner(mode='fetch')
HF.fetch(self.hmms, keyFile, self.outputHMMs, bKeyFile=True)
# index the HMM file
if os.path.exists(self.outputHMMs + '.ssi'):
os.remove(self.outputHMMs + '.ssi')
HF.index(self.outputHMMs)
# remove key file
os.remove(keyFile)
def run(self):
# read all taxonomic-specific marker genes
print 'Reading taxonomic-specific marker genes.'
taxonomicMarkers = set()
taxonParser = TaxonParser()
taxonMarkerSets = taxonParser.readMarkerSets()
for _, taxa in taxonMarkerSets.iteritems():
for _, markerSet in taxa.iteritems():
taxonomicMarkers = taxonomicMarkers.union(markerSet.getMarkerGenes())
print ' Taxonomic-specific marker genes: %d' % len(taxonomicMarkers)
# read all lineage-specific marker genes
print 'Reading lineage-specific marker genes.'
lineageMarkers = set()
treeParser = TreeParser()
uniqueIdToLineageStatistics = treeParser.readNodeMetadata()
for uniqueId, d in uniqueIdToLineageStatistics.iteritems():
markerSet = MarkerSet(uniqueId, 'NA', int(d['# genomes']), eval(d['marker set']))
lineageMarkers = lineageMarkers.union(markerSet.getMarkerGenes())
print ' Lineage-specific marker genes: %d' % len(lineageMarkers)
# gather all marker genes
markerGenes = taxonomicMarkers.union(lineageMarkers)
print ' Total marker genes: %d' % len(markerGenes)
# get genes from same clan as marker genes
print 'Gathering HMMs from the same clan as marker genes.'
pfam = PFAM()
genesInSameClan = pfam.genesInSameClan(markerGenes)
allMarkers = markerGenes.union(genesInSameClan)
# create file with all model accession numbers
keyFile = os.path.join(tempfile.gettempdir(), str(uuid.uuid4()))
fout = open(keyFile, 'w')
for modelAcc in allMarkers:
fout.write(modelAcc + '\n')
fout.close()
# fetch specified models
HF = HMMERRunner(mode='fetch')
HF.fetch(self.hmms, keyFile, self.outputHMMs, bKeyFile=True)
# index the HMM file
if os.path.exists(self.outputHMMs + '.ssi'):
os.remove(self.outputHMMs + '.ssi')
HF.index(self.outputHMMs)
# remove key file
os.remove(keyFile)
def run(self):
# read internal nodes file
metadata = {}
for line in open('./experiments/classTree.internal_nodes.tsv'):
uid, label = [x.strip() for x in line.split('\t')]
metadata[uid] = label
# read all lineage-specific marker genes
treeParser = TreeParser()
uniqueIdToLineageStatistics = treeParser.readNodeMetadata()
for uid in metadata:
stats = uniqueIdToLineageStatistics[uid]
markerSet = MarkerSet(uid, 'NA', int(stats['# genomes']), eval(stats['marker set']))
metadata[uid] += ' [%d, %d, %d]' % (stats['# genomes'], markerSet.numMarkers(), markerSet.numSets())
# write out results
fout = open('./experiments/classTree.internal_nodes.metadata.tsv', 'w')
for uid, label in metadata.iteritems():
fout.write(uid + '\t' + label + '\n')
fout.close()
def lineageSet(self, options, db=None):
"""Lineage set command"""
self.logger.info(
'[CheckM - lineage_set] Inferring lineage-specific marker sets.')
checkDirExists(options.tree_dir)
# set HMM file for each bin
markerSetParser = MarkerSetParser()
hmmModelInfoFile = os.path.join(options.tree_dir, 'storage',
DefaultValues.PHYLO_HMM_MODEL_INFO)
binIdToModels = markerSetParser.loadBinModels(hmmModelInfoFile)
# calculate marker gene statistics
resultsParser = ResultsParser(binIdToModels)
resultsParser.analyseResults(options.tree_dir,
DefaultValues.BIN_STATS_PHYLO_OUT,
DefaultValues.HMMER_TABLE_PHYLO_OUT)
# These options are incompatible with how the lineage-specific marker set is selected, so
# the default values are currently hard-coded
options.num_genomes_markers = 2
options.bootstrap = 0
options.bRequireTaxonomy = False
treeParser = TreeParser()
treeParser.getBinMarkerSets(
options.tree_dir, options.marker_file, options.num_genomes_markers,
options.bootstrap, options.bNoLineageSpecificRefinement,
options.bForceDomain, options.bRequireTaxonomy, resultsParser,
options.unique, options.multi)
self.logger.info('Marker set written to: ' + options.marker_file)
self.timeKeeper.printTimeStamp()
def run(self):
# read internal nodes file
metadata = {}
for line in open('./experiments/classTree.internal_nodes.tsv'):
uid, label = [x.strip() for x in line.split('\t')]
metadata[uid] = label
# read all lineage-specific marker genes
treeParser = TreeParser()
uniqueIdToLineageStatistics = treeParser.readNodeMetadata()
for uid in metadata:
stats = uniqueIdToLineageStatistics[uid]
markerSet = MarkerSet(uid, 'NA', int(stats['# genomes']),
eval(stats['marker set']))
metadata[uid] += ' [%d, %d, %d]' % (stats['# genomes'],
markerSet.numMarkers(),
markerSet.numSets())
# write out results
fout = open('./experiments/classTree.internal_nodes.metadata.tsv', 'w')
for uid, label in metadata.items():
fout.write(uid + '\t' + label + '\n')
fout.close()
def run(self, ubiquityThreshold, minGenomes):
# Pre-compute gene count table
print 'Computing gene count table.'
start = time.time()
metadata = self.img.genomeMetadata()
self.markerSetBuilder.cachedGeneCountTable = self.img.geneCountTable(metadata.keys())
end = time.time()
print ' globalGeneCountTable: %.2f' % (end - start)
# read selected node for defining marker set
print 'Reading node defining marker set for each internal node.'
selectedMarkerNode = {}
for line in open('/srv/whitlam/bio/db/checkm/selected_marker_sets.tsv'):
lineSplit = line.split('\t')
selectedMarkerNode[lineSplit[0].strip()] = lineSplit[1].strip()
# read duplicate taxa
print 'Reading list of identical taxa in genome tree.'
duplicateTaxa = {}
for line in open('/srv/whitlam/bio/db/checkm/genome_tree/genome_tree.derep.txt'):
lineSplit = line.rstrip().split()
if len(lineSplit) > 1:
duplicateTaxa[lineSplit[0]] = lineSplit[1:]
# read in node metadata
print 'Reading node metadata.'
treeParser = TreeParser()
uniqueIdToLineageStatistics = treeParser.readNodeMetadata()
# read genome tree
print 'Reading in genome tree.'
treeFile = '/srv/whitlam/bio/db/checkm/genome_tree/genome_tree_prok.refpkg/genome_tree.final.tre'
tree = dendropy.Tree.get_from_path(treeFile, schema='newick', as_rooted=True, preserve_underscores=True)
# determine lineage-specific gene loss and duplication (relative to potential marker genes used by a node)
print 'Determining lineage-specific gene loss and duplication'
fout = open('/srv/whitlam/bio/db/checkm/genome_tree/missing_duplicate_genes_50.tsv', 'w')
processed = 0
numInternalNodes = len(tree.internal_nodes())
for node in tree.internal_nodes():
processed += 1
statusStr = ' Finished processing %d of %d (%.2f%%) internal nodes.' % (processed, numInternalNodes, float(processed)*100/numInternalNodes)
sys.stdout.write('%s\r' % statusStr)
sys.stdout.flush()
nodeId = node.label.split('|')[0]
missingGenes = []
duplicateGenes = []
nodeStats = uniqueIdToLineageStatistics[nodeId]
if nodeStats['# genomes'] >= minGenomes:
# get marker genes defined for current node along with all parental nodes
markerGenes = set()
parentNode = node
while parentNode != None:
parentNodeId = parentNode.label.split('|')[0]
stats = uniqueIdToLineageStatistics[parentNodeId]
markerSet = MarkerSet(parentNodeId, stats['taxonomy'], stats['# genomes'], eval(stats['marker set']))
markerGenes = markerGenes.union(markerSet.getMarkerGenes())
parentNode = parentNode.parent_node
# silly hack since PFAM ids are inconsistent between the PFAM data and IMG data
revisedMarkerGeneIds = set()
for mg in markerGenes:
if mg.startswith('PF'):
revisedMarkerGeneIds.add(mg[0:mg.rfind('.')].replace('PF', 'pfam'))
else:
revisedMarkerGeneIds.add(mg)
# get all genomes below the internal node (including genomes removed as duplicates)
genomeIds = []
for leaf in node.leaf_nodes():
genomeIds.append(leaf.taxon.label.replace('IMG_', ''))
if leaf.taxon.label in duplicateTaxa:
for genomeId in duplicateTaxa[leaf.taxon.label]:
genomeIds.append(genomeId.replace('IMG_', ''))
genomeIds.append(leaf.taxon.label.replace('IMG_', ''))
missingGenes = self.markerSetBuilder.missingGenes(genomeIds, revisedMarkerGeneIds, ubiquityThreshold)
duplicateGenes = self.markerSetBuilder.duplicateGenes(genomeIds, revisedMarkerGeneIds, ubiquityThreshold)
fout.write('%s\t%s\t%s\n' % (nodeId, str(missingGenes), str(duplicateGenes)))
sys.stdout.write('\n')
fout.close()
