1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
|
import os.path as op
from progressbar import ProgressBar
import pysam
import pybedtools
from seqcluster.libs import pysen
import numpy as np
from seqcluster.libs.utils import file_exists
import seqcluster.libs.logger as mylog
from seqcluster.libs.classes import *
from seqcluster.detect.metacluster import _get_seqs_from_cluster
from seqcluster.libs.do import run
logger = mylog.getLogger(__name__)
def detect_complexity(bam_in, genome, out):
"""
genome coverage of small RNA
"""
if not genome:
logger.info("No genome given. skipping.")
return None
out_file = op.join(out, op.basename(bam_in) + "_cov.tsv")
if file_exists(out_file):
return None
fai = genome + ".fai"
cov = pybedtools.BedTool(bam_in).genome_coverage(max=1)
cov.saveas(out_file)
total = 0
for region in cov:
if region[0] == "genome" and int(region[1]) != 0:
total += float(region[4])
logger.info("Total genome with sequences: %s " % total)
def clean_bam_file(bam_in, mask=None):
"""
Remove from alignment reads with low counts and highly # of hits
"""
seq_obj = defaultdict(int)
if mask:
mask_file = op.splitext(bam_in)[0] + "_mask.bam"
if not file_exists(mask_file):
pybedtools.BedTool(bam_in).intersect(b=mask, v=True).saveas(mask_file)
bam_in = mask_file
out_file = op.splitext(bam_in)[0] + "_rmlw.bam"
# bam.index(bam_in, {'algorithm':{}})
run("samtools index %s" % bam_in)
if not file_exists(bam_in + ".bai"):
raise IOError("Failed to created bam index of %s. Try to do it manually" % bam_in)
bam_handle = pysam.AlignmentFile(bam_in, "rb")
with pysam.AlignmentFile(out_file, "wb", template=bam_handle) as out_handle:
for read in bam_handle.fetch():
seq_name = int(read.query_name.replace('seq_', ''))
match_size = [nts for oper, nts in read.cigartuples if oper == 0]
subs_size = [nts for oper, nts in read.cigartuples if oper == 4]
if match_size[0] < 17:
continue
if subs_size:
if subs_size[0] > 3:
continue
try:
nh = read.get_tag('NH')
except KeyError:
nh = 1
seq_obj[seq_name] = sequence(seq_name)
seq_obj[seq_name].align = nh
out_handle.write(read)
return out_file, seq_obj
def detect_clusters(c, current_seq, MIN_SEQ, non_un_gl=False):
"""
Parse the merge file of sequences position to create clusters that will have all
sequences that shared any position on the genome
:param c: file from bedtools with merge sequence positions
:param current_seq: list of sequences
:param MIN_SEQ: int cutoff to keep the cluster or not. 10 as default
:return: object with information about:
* cluster
* dict with sequences (as keys) and cluster_id (as value)
* sequences
* loci
"""
current_loci = {}
current_clus = {}
# sequence2clusters = [set()] * (max(current_seq.keys()) + 2)
sequence2clusters = defaultdict(set)
lindex = 0
eindex = 0
previous_id = 0
for line in c.features():
c, start, end, name, score, strand, c_id = line
c_id = int(c_id)
name = int(name.replace('seq_', ''))
pos = int(start) if strand == "+" else int(end)
if name not in current_seq:
continue
if c.find('Un_gl') > -1 and non_un_gl:
continue
if c_id != previous_id:
if previous_id > 0:
if len(current_clus[eindex].idmembers) < MIN_SEQ:
for s in current_clus[eindex].idmembers:
sequence2clusters[s] = sequence2clusters[s] - set([eindex])
del current_clus[eindex]
logger.debug("detect_cluster: %s %s %s" % (c_id, previous_id, name))
lindex += 1
eindex += 1
current_clus[eindex] = cluster(eindex)
newpos = position(lindex, c, start, end, strand)
current_loci[lindex] = newpos
# update locus, sequences in each line
current_loci[lindex].end = int(end)
current_loci[lindex].coverage[pos] += 1
size = range(pos, pos + current_seq[name].len)
current_loci[lindex].counts.update(dict(zip(size, [current_seq[name].total()] * current_seq[name].len)))
current_clus[eindex].idmembers[name] = 1
current_clus[eindex].add_id_member([name], lindex)
current_seq[name].add_pos(lindex, pos)
# current_seq[name].align = 1
previous_id = c_id
sequence2clusters[name].add(eindex)
logger.info("%s Clusters read" % eindex)
# merge cluster with shared sequences
metacluster_obj, cluster_id = _find_metaclusters(current_clus, sequence2clusters, current_seq, MIN_SEQ)
return cluster_info_obj(current_clus, metacluster_obj, current_loci, current_seq)
def _common(items, seen):
intersect = [e for e in map(seen.get, items)]
return list(filter(None, intersect))
def _update(clusters, idx, hash):
return hash.update(dict(zip(clusters, [idx] * len(clusters))))
def _find_metaclusters(clus_obj, sequence2clusters, current_seq, min_seqs):
"""
Mask under same id all clusters that share sequences
:param clus_obj: cluster object coming from detect_cluster
:param min_seqs: int cutoff to keep the cluster or not. 10 as default
:return: updated clus_obj and dict with seq_id: cluster_id
"""
seen = defaultdict(int)
metacluster = defaultdict(set)
c_index = len(sequence2clusters)
logger.info("Creating meta-clusters based on shared sequences: %s" % c_index)
meta_idx = 1
bar = ProgressBar(maxval=c_index).start()
bar.update()
for itern, name in enumerate(sequence2clusters):
clusters = sequence2clusters[name]
if len(clusters) == 0:
c_index -= 1
continue
current_seq[name].align = 1
meta_idx += 1
bar.update(itern)
# import pdb; pdb.set_trace()
already_in = _common(clusters, seen)
_update(clusters, meta_idx, seen)
metacluster[meta_idx] = metacluster[meta_idx].union(clusters)
if already_in:
for seen_metacluster in already_in:
clusters2merge = metacluster[seen_metacluster]
metacluster[meta_idx] = metacluster[meta_idx].union(clusters2merge)
_update(clusters2merge, meta_idx, seen)
# metacluster[seen_metacluster] = 0
del metacluster[seen_metacluster]
logger.info("%s metaclusters from %s sequences" % (len(metacluster), c_index))
return metacluster, seen
def _find_families_deprecated(clus_obj, min_seqs):
"""
Mask under same id all clusters that share sequences
:param clus_obj: cluster object coming from detect_cluster
:param min_seqs: int cutoff to keep the cluster or not. 10 as default
:return: updated clus_obj and dict with seq_id: cluster_id
"""
logger.info("Creating meta-clusters based on shared sequences.")
seen = defaultdict()
metacluster = defaultdict(list)
c_index = clus_obj.keys()
meta_idx = 0
p = ProgressBar(maxval=len(c_index), redirect_stdout=True).start()
for itern, c in enumerate(c_index):
p.update(itern)
clus = clus_obj[c]
if len(clus.idmembers.keys()) < min_seqs:
del clus_obj[c]
continue
logger.debug("reading cluster %s" % c)
logger.debug("loci2seq %s" % clus.loci2seq)
already_in, not_in = _get_seqs_from_cluster(clus.idmembers.keys(), seen)
logger.debug("seen %s news %s" % (already_in, not_in))
meta_idx += 1
metacluster[meta_idx].append(c)
seen.update(dict(zip(not_in, [meta_idx] * len(not_in))))
if len(already_in) > 0:
logger.debug("seen in %s" % already_in)
for eindex in already_in:
for cluster in metacluster[eindex]:
metacluster[meta_idx].append(cluster)
prev_clus = clus_obj[cluster]
logger.debug("_find_families: prev %s current %s" % (eindex, clus.id))
# add current seqs to seen cluster
seqs_in = prev_clus.idmembers.keys()
seen.update(dict(zip(seqs_in, [meta_idx] * len(seqs_in))))
# for s_in_clus in prev_clus.idmembers:
# seen[s_in_clus] = meta_idx
# clus.idmembers[s_in_clus] = 1
# add current locus to seen cluster
# for loci in prev_clus.loci2seq:
# logger.debug("adding %s" % loci)
# if not loci_old in current_clus[eindex].loci2seq:
# clus.add_id_member(list(prev_clus.loci2seq[loci]), loci)
# logger.debug("loci %s" % clus.loci2seq.keys())
del metacluster[eindex]
# clus_obj[c] = clus
# logger.debug("num cluster %s" % len(clus_obj.keys()))
logger.info("%s clusters merged" % len(metacluster))
return metacluster, seen
def peak_calling(clus_obj):
"""
Run peak calling inside each cluster
"""
new_cluster = {}
for cid in clus_obj.clus:
cluster = clus_obj.clus[cid]
cluster.update()
logger.debug("peak calling for %s" % cid)
bigger = cluster.locimaxid
if bigger in clus_obj.loci:
s, e = min(clus_obj.loci[bigger].counts.keys()), max(clus_obj.loci[bigger].counts.keys())
scale = s
if clus_obj.loci[bigger].strand == "-":
scale = e
logger.debug("bigger %s at %s-%s" % (bigger, s, e))
dt = np.array([0] * (abs(e - s) + 12))
for pos in clus_obj.loci[bigger].counts:
ss = abs(int(pos) - scale) + 5
dt[ss] += clus_obj.loci[bigger].counts[pos]
x = np.array(range(0, len(dt)))
logger.debug("x %s and y %s" % (x, dt))
# tab = pd.DataFrame({'x': x, 'y': dt})
# tab.to_csv( str(cid) + "peaks.csv", mode='w', header=False, index=False)
if len(x) > 35 + 12:
peaks = list(np.array(pysen.pysenMMean(x, dt)) - 5)
logger.debug(peaks)
else:
peaks = ['short']
cluster.peaks = peaks
new_cluster[cid] = cluster
clus_obj.clus = new_cluster
return clus_obj
|
