def search(self, fastaRecord, excludelist=[], usecache=True):
# Write the query to a tmp file:
tmpQueryFileIdent, tmpQueryFileName = tempfile.mkstemp()
writeFile(tmpQueryFileName, str(fastaRecord))
# File name used for blast cache
fileSuffix = ''
for name in excludelist:
l = re.split(r'\s+', name)
for n in l:
fileSuffix += n[0]
if fileSuffix:
fileSuffix = '_' + fileSuffix
blastFileName = os.path.join(
self.options.blastcache,
"%s.%d_%s%s.xml" % (fastaRecord.title, self.options.maxblasthits,
self.options.minsignificance, fileSuffix))
if usecache and os.path.exists(
blastFileName) and os.path.getsize(blastFileName) > 0:
# Use cached blast result
if excludelist:
print "\n\t\tUsing cached Blast results (excluding %s)..." % ', '.join(
excludelist),
else:
print "\n\t\tUsing cached Blast results...",
sys.stdout.flush()
blastFile = open(blastFileName, 'r')
else:
if excludelist:
print "\n\t\tSearching database (excluding %s)..." % ', '.join(
excludelist),
else:
print "\n\t\tSearching database...",
sys.stdout.flush()
if excludelist:
# Build a blast db of the relevant records:
excludeIDs = []
for taxon in excludelist:
excludeIDs.extend(self.index[taxon])
includeIDs = self.sequence_index.keys.difference(
set(excludeIDs))
if not includeIDs:
print "done (database exhausted)\n\t\t\t",
sys.stdout.flush()
return SearchResult(None)
blastDBfileName = os.path.join(self.options.project,
"tmpBlastDB.fasta")
tmpFastaFile = open(blastDBfileName, 'w')
for key in includeIDs:
tmpFastaFile.write(str(self.sequence_index.get_entry(key)))
tmpFastaFile.close()
cmd = "makeblastdb -dbtype nucl -title %s -in %s" % (
blastDBfileName, blastDBfileName)
#cmd = "xdformat -n -o %s %s" % (blastDBfileName, blastDBfileName)
cmd = self.escape(cmd)
systemCall(cmd, stdout='IGNORE', stderr='IGNORE')
else:
blastDBfileName = self.blastSequenceFileName
#blastDBfileName = self.blastDB
# Blast:
if self.options.blastwordsize:
wordSize = '-word_size %s' % self.options.blastwordsize
else:
wordSize = ''
if self.options.nolowcomplexfilter:
filterOption = '-dust no'
else:
filterOption = '-dust yes' # FIXME: Check that this is an ok default... It is not the defalut in blastn
cmd = 'blastn -db %s -outfmt 5 %s %s -evalue %s -max_target_seqs %s -query %s' \
% (blastDBfileName, wordSize, filterOption, self.options.minsignificance, self.options.maxblasthits,
tmpQueryFileName)
# cmd = "blastn %s -e %f -p blastn -d %s.fasta -i %s -m 7" % (wordSize, self.options.minsignificance, blastDBfileName, tmpQueryFileName)
cmd = self.escape(cmd)
STARTUPINFO = None
if os.name == 'nt':
STARTUPINFO = subprocess.STARTUPINFO()
STARTUPINFO.dwFlags |= subprocess.STARTF_USESHOWWINDOW
proc = subprocess.Popen(cmd,
shell=True,
startupinfo=STARTUPINFO,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
stdout_value, stderr_value = proc.communicate()
blastContent = str(stdout_value)
# stdin, stdout, stderr = os.popen3(cmd)
for f in glob.glob(
os.path.join(self.options.project, 'tmpBlastDB.*')):
os.remove(f)
# blastContent = stdout.read()
#
# stdout.close()
# stdin.close()
# stderr.close()
# This is a hack to remove an xml tag that just confuses things with blastn:
blastContent = re.sub(r'<Hit_id>.*?</Hit_id>', '', blastContent)
writeFile(blastFileName, blastContent)
#blastFile = StringIO.StringIO(blastContent)
blastFile = open(blastFileName, 'r')
#blastRecord = NCBIXML.read(blastFile)
try:
blastRecord = NCBIXML.read(blastFile)
except:
blastRecord = None
blastFile.close()
os.close(tmpQueryFileIdent)
os.unlink(tmpQueryFileName)
print "done.\n\t\t\t",
sys.stdout.flush()
return SearchResult(blastRecord)
def search(self, fastaRecord, excludelist=[], usecache=True):
# Write the query to a tmp file:
tmpQueryFileIdent, tmpQueryFileName = tempfile.mkstemp()
writeFile(tmpQueryFileName, str(fastaRecord))
# File name used for blast cache
fileSuffix = ''
for name in excludelist:
l = re.split(r'\s+', name)
for n in l:
fileSuffix += n[0]
if fileSuffix:
fileSuffix = '_' + fileSuffix
blastFileName = os.path.join(self.options.blastcache, "%s.%d_%s%s.xml" % (fastaRecord.title,
self.options.maxblasthits, self.options.minsignificance, fileSuffix))
if usecache and os.path.exists(blastFileName) and os.path.getsize(blastFileName)>0:
# Use cached blast result
if excludelist:
print "\n\t\tUsing cached Blast results (excluding %s)..." % ', '.join(excludelist),
else:
print "\n\t\tUsing cached Blast results...",
sys.stdout.flush()
blastFile = open(blastFileName, 'r')
else:
if excludelist:
print "\n\t\tSearching database (excluding %s)..." % ', '.join(excludelist),
else:
print "\n\t\tSearching database...",
sys.stdout.flush()
if excludelist:
# Build a blast db of the relevant records:
excludeIDs = []
for taxon in excludelist:
excludeIDs.extend(self.index[taxon])
includeIDs = self.sequence_index.keys.difference(set(excludeIDs))
if not includeIDs:
print "done (database exhausted)\n\t\t\t",
sys.stdout.flush()
return SearchResult(None)
blastDBfileName = os.path.join(self.options.project, "tmpBlastDB.fasta")
tmpFastaFile = open(blastDBfileName, 'w')
for key in includeIDs:
tmpFastaFile.write(str(self.sequence_index.get_entry(key)))
tmpFastaFile.close()
cmd = "makeblastdb -dbtype nucl -title %s -in %s" % (blastDBfileName, blastDBfileName)
#cmd = "xdformat -n -o %s %s" % (blastDBfileName, blastDBfileName)
cmd = self.escape(cmd)
systemCall(cmd, stdout='IGNORE', stderr='IGNORE')
else:
blastDBfileName = self.blastSequenceFileName
#blastDBfileName = self.blastDB
# Blast:
if self.options.blastwordsize:
wordSize = '-word_size %s' % self.options.blastwordsize
else:
wordSize = ''
if self.options.nolowcomplexfilter:
filterOption = '-dust no'
else:
filterOption = '-dust yes' # FIXME: Check that this is an ok default... It is not the defalut in blastn
cmd = 'blastn -db %s -outfmt 5 %s %s -evalue %s -max_target_seqs %s -query %s' \
% (blastDBfileName, wordSize, filterOption, self.options.minsignificance, self.options.maxblasthits,
tmpQueryFileName)
# cmd = "blastn %s -e %f -p blastn -d %s.fasta -i %s -m 7" % (wordSize, self.options.minsignificance, blastDBfileName, tmpQueryFileName)
cmd = self.escape(cmd)
STARTUPINFO = None
if os.name == 'nt':
STARTUPINFO = subprocess.STARTUPINFO()
STARTUPINFO.dwFlags |= subprocess.STARTF_USESHOWWINDOW
proc = subprocess.Popen(cmd, shell=True, startupinfo=STARTUPINFO,
stdout=subprocess.PIPE, stderr=subprocess.PIPE)
stdout_value, stderr_value = proc.communicate()
blastContent = str(stdout_value)
# stdin, stdout, stderr = os.popen3(cmd)
for f in glob.glob(os.path.join(self.options.project, 'tmpBlastDB.*')):
os.remove(f)
# blastContent = stdout.read()
#
# stdout.close()
# stdin.close()
# stderr.close()
# This is a hack to remove an xml tag that just confuses things with blastn:
blastContent = re.sub(r'<Hit_id>.*?</Hit_id>', '', blastContent)
writeFile(blastFileName, blastContent)
#blastFile = StringIO.StringIO(blastContent)
blastFile = open(blastFileName, 'r')
#blastRecord = NCBIXML.read(blastFile)
try:
blastRecord = NCBIXML.read(blastFile)
except:
blastRecord = None
blastFile.close()
os.close(tmpQueryFileIdent)
os.unlink(tmpQueryFileName)
print "done.\n\t\t\t",
sys.stdout.flush()
return SearchResult(blastRecord)
def _netblast_search(self, fastaRecord, excludelist=[], usecache=True):
"""
Blast against genbank over web
"""
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
if self.options.ONEMISSING:
orig_limit_query = self.options.limitquery
genus, species = re.search(r'_([^_]+)_([^_]+)$',
fastaRecord.title).groups()
self.options.limitquery = "barcode[keyword] NOT %s %s[ORGN]" % (
genus, species)
print "REFORMATTING LIMIT QUERY TO", self.options.limitquery
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
# Make a query to filter the returned results:
if excludelist:
entrezQuery = '(' + self.options.limitquery + ') NOT (uncultured[WORD] OR ' + '[ORGN] OR '.join(
excludelist) + '[ORGN])'
else:
entrezQuery = '(' + self.options.limitquery + ') NOT uncultured[WORD]'
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
if self.options.ONEMISSING:
self.options.limitquery = orig_limit_query
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
fileSuffix = ''
for name in excludelist:
l = re.split(r'\s+', name)
for n in l:
fileSuffix += n[0]
if fileSuffix:
fileSuffix = '_' + fileSuffix
# File name used for blast cache
blastFileName = os.path.join(
self.options.blastcache,
"%s.%d_%s%s.xml" % (fastaRecord.title, self.options.maxblasthits,
self.options.minsignificance, fileSuffix))
if usecache and os.path.exists(
blastFileName) and os.path.getsize(blastFileName) > 0:
# Use cached blast result
if excludelist:
print "\n\t\tUsing cached Blast results (excluding %s)..." % ', '.join(
excludelist),
else:
print "\n\t\tUsing cached Blast results...",
sys.stdout.flush()
else:
# Make a query to filter the returned results:
if excludelist:
print "\n\t\tSearching database (excluding %s)..." % ', '.join(
excludelist),
else:
print "\n\t\tSearching database...",
sys.stdout.flush()
fastaRecordFileName = os.path.join(self.options.project,
utils.randomString(8))
fastaRecordFile = open(fastaRecordFileName, 'w')
fastaRecordFile.write(str(fastaRecord))
fastaRecordFile.close()
resultHandle = None
if self.options.nolowcomplexfilter:
filterOption = '-dust no'
else:
filterOption = '-dust yes' # FIXME: Check that this is an ok default... It is not the defalut in blastn
if self.options.blastwordsize:
wordSize = '-word_size %s' % self.options.blastwordsize
else:
wordSize = ''
blastCmd = 'blastn -remote -outfmt 5 -db nt %s %s -evalue %s -max_target_seqs %s -entrez_query "%s" -query %s -out %s' \
% (wordSize, filterOption, self.options.minsignificance, self.options.maxblasthits, \
entrezQuery, fastaRecordFileName, blastFileName)
for i in range(20):
time.sleep(2 * i)
error = utils.systemCall(blastCmd,
stdout='IGNORE',
stderr='IGNORE')
try:
# retval = os.system(blastCmd)
# if retval != 0:
# print "Netblast failed with return value %d. Trying again..." % retval
error = utils.systemCall(blastCmd)
if error or not os.path.exists(
blastFileName) or os.path.getsize(
blastFileName) == 0:
print "Netblast failed. Trying again..."
continue
break
except KeyboardInterrupt:
sys.exit()
except:
print "Netblast failed. Trying again..."
pass
os.remove(fastaRecordFileName)
# Read file from cache
blastHandle = open(blastFileName, 'r')
# Parse the result:
try:
blastRecord = NCBIXML.read(blastHandle)
print "done.\n\t\t\t",
sys.stdout.flush()
except:
blastRecord = None
blastHandle.close()
return blastRecord
def _netblast_search(self, fastaRecord, excludelist=[], usecache=True):
"""
Blast against genbank over web
"""
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
if self.options.ONEMISSING:
orig_limit_query = self.options.limitquery
genus, species = re.search(r'_([^_]+)_([^_]+)$', fastaRecord.title).groups()
self.options.limitquery = "barcode[keyword] NOT %s %s[ORGN]" % (genus, species)
print "REFORMATTING LIMIT QUERY TO", self.options.limitquery
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
# Make a query to filter the returned results:
if excludelist:
entrezQuery = '(' + self.options.limitquery + ') NOT (uncultured[WORD] OR ' + '[ORGN] OR '.join(excludelist) + '[ORGN])'
else:
entrezQuery = '(' + self.options.limitquery + ') NOT uncultured[WORD]'
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
if self.options.ONEMISSING:
self.options.limitquery = orig_limit_query
### THIS IS A HACK FOR TESTING PURPOSES TO REMOVED AGAIN ##########################################
fileSuffix = ''
for name in excludelist:
l = re.split(r'\s+', name)
for n in l:
fileSuffix += n[0]
if fileSuffix:
fileSuffix = '_' + fileSuffix
# File name used for blast cache
blastFileName = os.path.join(self.options.blastcache, "%s.%d_%s%s.xml" % (fastaRecord.title,
self.options.maxblasthits, self.options.minsignificance, fileSuffix))
if usecache and os.path.exists(blastFileName) and os.path.getsize(blastFileName)>0:
# Use cached blast result
if excludelist:
print "\n\t\tUsing cached Blast results (excluding %s)..." % ', '.join(excludelist),
else:
print "\n\t\tUsing cached Blast results...",
sys.stdout.flush()
else:
# Make a query to filter the returned results:
if excludelist:
print "\n\t\tSearching database (excluding %s)..." % ', '.join(excludelist),
else:
print "\n\t\tSearching database...",
sys.stdout.flush()
fastaRecordFileName = os.path.join(self.options.project, utils.randomString(8))
fastaRecordFile = open(fastaRecordFileName, 'w')
fastaRecordFile.write(str(fastaRecord))
fastaRecordFile.close()
resultHandle = None
if self.options.nolowcomplexfilter:
filterOption = '-dust no'
else:
filterOption = '-dust yes' # FIXME: Check that this is an ok default... It is not the defalut in blastn
if self.options.blastwordsize:
wordSize = '-word_size %s' % self.options.blastwordsize
else:
wordSize = ''
blastCmd = 'blastn -remote -outfmt 5 -db nt %s %s -evalue %s -max_target_seqs %s -entrez_query "%s" -query %s -out %s' \
% (wordSize, filterOption, self.options.minsignificance, self.options.maxblasthits, \
entrezQuery, fastaRecordFileName, blastFileName)
for i in range(20):
time.sleep(2 * i)
error = utils.systemCall(blastCmd, stdout='IGNORE', stderr='IGNORE')
try:
# retval = os.system(blastCmd)
# if retval != 0:
# print "Netblast failed with return value %d. Trying again..." % retval
error = utils.systemCall(blastCmd)
if error or not os.path.exists(blastFileName) or os.path.getsize(blastFileName) == 0:
print "Netblast failed. Trying again..."
continue
break
except KeyboardInterrupt:
sys.exit()
except:
print "Netblast failed. Trying again..."
pass
os.remove(fastaRecordFileName)
# Read file from cache
blastHandle = open(blastFileName, 'r')
# Parse the result:
try:
blastRecord = NCBIXML.read(blastHandle)
print "done.\n\t\t\t",
sys.stdout.flush()
except:
blastRecord = None
blastHandle.close()
return blastRecord
