def FindMarkersMetachecker(self, marker_database_name, version, fasta_file):
markers = markers_module.getAllMarkerSets()
filter_function = lambda x,y : (x == marker_database_name) and (y == version)
filtered_markers = dict([(x.name, x) for x in markers if filter_function(x.database, x.version)])
result_dir = tempfile.mkdtemp()
concatenated_hmm = tempfile.NamedTemporaryFile(delete=False)
for marker in filtered_markers.values():
with open(os.path.join(markers_module_path, marker.rel_path)) as fh:
for line in fh:
concatenated_hmm.write(line)
concatenated_hmm.close()
prefix = 'gtdb_'
dc = DataConstructor()
dc.buildData([fasta_file], result_dir, concatenated_hmm.name, prefix, quiet=True)
qa = QaParser(prefix=prefix)
qa.analyseResults(result_dir, concatenated_hmm.name)
aligner = HMMAligner(prefix=prefix, individualFile=True,
includeConsensus=False, outputFormat="Pfam")
aligner.makeAlignments(result_dir,
concatenated_hmm.name,prefix=prefix,bestHit=True)
result_dict = dict()
for folder in os.listdir(result_dir):
# returns the summary information in metachecka for addition into
# the DB
summary_info = qa.results[folder].calculateMarkers()
for marker_name in filtered_markers:
try:
with open(os.path.join(result_dir,folder,marker_name)+"_out.align") as fh:
fh.readline()
fh.readline()
seqline = fh.readline()
seq_start_pos = seqline.rfind(' ')
fh.readline()
fh.readline()
mask = fh.readline()
seqline = seqline[seq_start_pos:]
mask = mask[seq_start_pos:]
seqline = ''.join([seqline[x] for x in range(0, len(seqline)) if mask[x] == 'x'])
if (seqline.count('-') / float(len(seqline))) > 0.5: # Limit to less than half gaps
continue
result_dict[marker_name] = seqline
except IOError:
pass
#cleanup
os.remove(concatenated_hmm.name)
shutil.rmtree(result_dir)
return result_dict
def FindMarkersEmboss(self, marker_database_name, version, fasta_file):
markers = markers_module.getAllMarkerSets()
filter_function = lambda x,y : (x == marker_database_name) and (y == version)
filtered_markers = dict([(x.name, x) for x in markers if filter_function(x.database, x.version)])
result_dir = tempfile.mkdtemp()
# Lazy solution - split up into 10kb segments (offset by 5k) so that hmm_align only has to align 10kb max.
fh = open(os.path.join(result_dir, "segmented_fasta.fa"), "wb")
for (name, seq, qual) in readfq(open(fasta_file)):
pos = 10000
while True:
fh.write(">%i_%i_%s\n" % (pos - 10000, pos, name))
fh.write(seq[pos-10000:pos] + "\n")
if len(seq) <= pos:
break
pos += 5000
fh.close()
sequence_dict = dict()
hmmer = HMMERRunner()
subprocess.call(["transeq", '-sequence', os.path.join(result_dir, "segmented_fasta.fa"),
'-outseq', os.path.join(result_dir, marker_database_name + ".faa"),
'-table', '11',
'-frame', '6'])
for marker in filtered_markers.values():
hmmer.search(os.path.join(markers_module_path, marker.rel_path),
os.path.join(result_dir, marker_database_name + ".faa"),
os.path.join(result_dir, marker.name))
parser = HMMERParser(open(os.path.join(result_dir, marker.name, 'hmmer_out.txt')))
result = parser.next()
if result:
sequence_dict[marker.name] = result.target_name
target_seq_dict = dict()
count = 0
for (name, seq, qual) in readfq(open(os.path.join(result_dir, marker_database_name + ".faa"))):
count += 1
if name in sequence_dict.values():
target_seq_dict[name] = count
fh = open(os.path.join(result_dir, str(count) + ".faa"), 'wb')
fh.write(">" + name + "\n")
fh.write(seq)
fh.close()
result_dict = dict()
for (marker_name, target_name) in sequence_dict.items():
os.system("hmmalign --allcol --outformat Pfam -o %s %s %s" %
(os.path.join(result_dir, marker_name + ".aligned"),
os.path.join(markers_module_path, filtered_markers[marker_name].rel_path),
os.path.join(result_dir, str(target_seq_dict[target_name]) + ".faa")))
fh = open(os.path.join(result_dir, marker_name + ".aligned"))
fh.readline()
fh.readline()
seqline = fh.readline()
seq_start_pos = seqline.rfind(' ')
fh.readline()
fh.readline()
mask = fh.readline()
seqline = seqline[seq_start_pos:]
mask = mask[seq_start_pos:]
seqline = ''.join([seqline[x] for x in range(0, len(seqline)) if mask[x] == 'x'])
if (seqline.count('-') / float(len(seqline))) > 0.5: # Limit to less than half gaps
continue
result_dict[marker_name] = seqline
subprocess.call(["rm", "-rf", result_dir])
return result_dict