def test_clusters_from_uc_file(self):
""" clusters_from_uc_file functions as expected """
expected_clusters = {'s2':['s2','s3']}
expected_failures = ['s1']
expected_new_seeds = ['s2']
self.assertEqual(clusters_from_uc_file(self.uc_lines1),
(expected_clusters,expected_failures,expected_new_seeds))
def test_clusters_from_uc_file(self):
""" clusters_from_uc_file functions as expected """
expected_clusters = {"s2": ["s2", "s3"]}
expected_failures = ["s1"]
expected_new_seeds = ["s2"]
self.assertEqual(
clusters_from_uc_file(self.uc_lines1), (expected_clusters, expected_failures, expected_new_seeds)
)
def test_clusters_from_uc_file_multiple_hits(self):
""" clusters_from_uc_file handles error_on_multiple_hits correctly
"""
# when a query hits multiple hits and error_on_multiple_hits=True
# an error should be raised
self.assertRaises(
UclustParseError,
clusters_from_uc_file,
self.uc_lines_w_multiple_hits_per_query,
error_on_multiple_hits=True,
)
# when a query hits multiple hits and error_on_multiple_hits=False
# the query should show up in multiple clusters
actual = clusters_from_uc_file(self.uc_lines_w_multiple_hits_per_query, error_on_multiple_hits=False)
expected_clusters = {"s2": ["s2", "s3"], "s4": ["s4", "s3"]}
expected_failures = ["s1"]
expected_new_seeds = ["s2", "s4"]
self.assertEqual(actual, (expected_clusters, expected_failures, expected_new_seeds))
def test_clusters_from_uc_file_multiple_hits(self):
""" clusters_from_uc_file handles error_on_multiple_hits correctly
"""
# when a query hits multiple hits and error_on_multiple_hits=True
# an error should be raised
self.assertRaises(UclustParseError,
clusters_from_uc_file,
self.uc_lines_w_multiple_hits_per_query,
error_on_multiple_hits=True)
# when a query hits multiple hits and error_on_multiple_hits=False
# the query should show up in multiple clusters
actual = clusters_from_uc_file(self.uc_lines_w_multiple_hits_per_query,
error_on_multiple_hits=False)
expected_clusters = {'s2':['s2','s3'],
's4':['s4','s3']}
expected_failures = ['s1']
expected_new_seeds = ['s2','s4']
self.assertEqual(actual,
(expected_clusters,expected_failures,expected_new_seeds))
def cluster_seqs(seqspath, simm, folderout='/tmp', gapopen=None, gapext=None):
if folderout[-1] != "/":
folderout += "/"
params = {
'--usersort': True,
'--id': float(simm),
'--maxaccepts': 20,
'--maxrejects': 500,
'--stepwords': 20,
'--hsp': 0,
'--match': 1,
'--mismatch': -1
}
if gapopen is not None:
params['--gapopen'] = gapopen
if gapext is not None:
params['--gapext'] = gapext
uclust = Uclust(params, WorkingDir='/tmp')
input_data = {
'--input': seqspath,
'--uc': folderout + "clusters.uc",
'--log': folderout + "clusters.log"
}
result = uclust(input_data)
clusters, failures, newseeds = clusters_from_uc_file(result['ClusterFile'])
seqs = LoadSeqs(seqspath, aligned=False)
convheader = {}
clusterseqs = {}
#create dictinary to convert shortened headers to full headers
for header in seqs.getSeqNames():
convheader[header.split()[0]] = header
#match headers in each cluster to seqs to create cluster tuples list
for num, cluster in enumerate(clusters):
clusterseqs["cluster_" + str(num)] = []
for header in clusters[cluster]:
clusterseqs["cluster_" + str(num)].append((convheader[header],
seqs.getSeq(convheader[header])))
return clusterseqs
def cluster_seqs(seqspath, simm, folderout='/tmp', gapopen=None, gapext=None):
if folderout[-1] != "/":
folderout += "/"
params = {
'--usersort': True,
'--id': float(simm),
'--maxaccepts': 20,
'--maxrejects': 500,
'--stepwords': 20,
'--hsp': 0,
'--match': 1,
'--mismatch': -1
}
if gapopen is not None:
params['--gapopen'] = gapopen
if gapext is not None:
params['--gapext'] = gapext
uclust = Uclust(params, WorkingDir='/tmp')
input_data = {
'--input': seqspath,
'--uc': folderout + "clusters.uc",
'--log': folderout + "clusters.log"
}
result = uclust(input_data)
clusters, failures, newseeds = clusters_from_uc_file(result['ClusterFile'])
seqs = LoadSeqs(seqspath, aligned=False)
convheader = {}
clusterseqs = {}
#create dictinary to convert shortened headers to full headers
for header in seqs.getSeqNames():
convheader[header.split()[0]] = header
#match headers in each cluster to seqs to create cluster tuples list
for num, cluster in enumerate(clusters):
clusterseqs["cluster_" + str(num)] = []
for header in clusters[cluster]:
clusterseqs["cluster_" + str(num)].append(
(convheader[header], seqs.getSeq(convheader[header])))
return clusterseqs
def create_clusters(fastain, folderout, simmilarity, minseqs=0):
params = {
'--log': folderout + str(simmilarity) + "_clusters.log",
'--usersort': False,
'--id': float(simmilarity),
'--maxaccepts': 20,
'--maxrejects': 500,
'--stepwords': 20,
'--w': 12,
'--gapopen': '10.0',
'--gapext': '10.0',
}
uclust = Uclust(params, WorkingDir='/tmp')
input_data = {
'--input': fastain,
'--uc': folderout + str(simmilarity) + "_clusters.uc"
}
result = uclust(input_data)
clusters, failures, new_seeds = clusters_from_uc_file(result['ClusterFile'])
#read in headers to rebuild full headers
headers = {}
for header, seq in MinimalFastaParser(open(fastain)):
headers[header.split()[0]] = (header, seq)
otus = []
otusout = open(folderout + str(simmilarity) + "_clusters.txt", 'w')
for group, cluster in enumerate(clusters):
count = 0
for seq in cluster:
count += int(seq.split("_")[1])
if count >= minseqs:
otusout.write(str(group) + "\t")
otus.append([])
#map headers back to orignal ones with counts
for header in clusters[cluster]:
otus[group].append(headers[header])
otusout.write(headers[header][0] + "\t")
otusout.write("\n")
otusout.close()
return otus
'--maxaccepts': 20,
'--maxrejects': 500,
'--stepwords': 20,
'--w': 12,
'--gapopen': '1.0/*TI',
'--gapext': '1.0',
'--hsp': 0
}
uclust = Uclust(params, WorkingDir='/tmp')
input_data = {
'--input': argv[1],
'--uc': argv[2] + argv[3] + "_clusters.uc",
'--log': argv[2] + argv[3] + "_clusters.log"
}
result = uclust(input_data)
clusters, failures, new_seeds = clusters_from_uc_file(result['ClusterFile'])
print "RESULTS: ", len(clusters)
headers = {}
for header, seq in MinimalFastaParser(open(argv[1])):
headers[header.split()[0]] = header
otusout = open(argv[2] + argv[3] + "_clusters.txt", 'w')
for group, cluster in enumerate(clusters):
otusout.write(str(group) + "\t")
#map headers back to orignal ones with counts
for header in clusters[cluster]:
otusout.write(headers[header] + "\t")
otusout.write("\n")
otusout.close()
#log = open(argv[2] + argv[3] + "_clusters.log", 'w')
#log.write('\n'.join(input_data) + '\n'.join(params) +
# "clusters: " + str(len(clusters)) + "\n")
