-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbamtest.py
288 lines (221 loc) · 8.7 KB
/
bamtest.py
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
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
import snppy.gene
import snppy.gtf
import snppy.rtree
import snppy.range
import collections
import pysam
import sys
GTF_FILE = 'Homo_sapiens.GRCh37_slim.64.gtf'
# GTF_FILE = 'empty.gtf'
def cigarToRefRange(pos, cigar):
out = []
for op,l in cigar:
if op == 0:
out.append(snppy.range.Range((pos, pos+l)))
if op == 0 or op == 2 or op == 3:
pos += l
return out
class RTreeCollection(object):
def __init__(self, keyed_objects):
objects_by_key = collections.defaultdict(list)
for k, ob in keyed_objects:
objects_by_key[k].append(ob)
self.trees = {}
for k, v in objects_by_key.iteritems():
self.trees[k] = snppy.rtree.RTreeNode.construct(v, 20, 20)
def lookup(self, k, *spans):
rt = self.trees.get(k)
if rt is None: return frozenset()
hits = set()
for span in spans:
hits.update(rt.search(span))
return frozenset(hits)
class CountUniqueGeneFragments(object):
def __init__(self, gtf_input):
print >>sys.stderr, 'reading GTF file'
chr_genes = []
for gene in snppy.gene.genesFromGTF(snppy.gtf.GTFFile.read(open(GTF_FILE, 'rbU'))):
chr_genes.append((gene.chromosome, gene))
self.genes = RTreeCollection(chr_genes)
print >>sys.stderr, 'done'
self.gene_counts = collections.defaultdict(int)
self.reads_by_name = {}
self.C = 0
self.S = 0
def processReadSingleton(self, r):
r_ch, r_ex = samfile.getrname(r.tid), cigarToRefRange(r.pos, r.cigar)
r_hits = self.genes.lookup(r_ch, *r_ex)
if len(r_hits) == 1:
self.gene_counts[iter(r_hits).next()] += 1
def processReadPair(self, r1, r2):
r1_ch, r1_ex = samfile.getrname(r1.tid), cigarToRefRange(r1.pos, r1.cigar)
r2_ch, r2_ex = samfile.getrname(r2.tid), cigarToRefRange(r2.pos, r2.cigar)
r1_hits = self.genes.lookup(r1_ch, *r1_ex)
r2_hits = self.genes.lookup(r2_ch, *r2_ex)
if len(r1_hits) == 1 and len(r2_hits) == 1 and r1_hits == r2_hits:
self.gene_counts[iter(r1_hits).next()] += 1
else:
print r1_hits, r2_hits
self.S += 1
def addRead(self, read):
self.C = self.C + 1
if read.qname in self.reads_by_name:
r1 = self.reads_by_name.pop(read.qname)
r2 = read
self.processReadPair(r1, r2)
else:
self.reads_by_name[read.qname] = read
if self.C % 100000 == 0:
print >>sys.stderr, self.C, len(self.reads_by_name), self.S, sum(self.gene_counts.values())
def done(self):
for r in self.reads_by_name.itervalues():
self.processReadSingleton(r)
def dump(self, out):
out.write('Symbol\tCount\n')
for i in self.gene_counts.iteritems():
out.write('%s\t%d\n' % i)
# def transcriptAlignmentOverlap(transcript, alignment):
# return sum([ r.size[0] for r in snppy.range.intersection(sorted(transcript.exons), alignment) ])
# def processIntergenicHits(hits):
# for (lo, hi), extents, read in hits:
# print 'XXX', 0.0, read[0].qname, read[1].qname
# gene_counts = collections.defaultdict(int)
# gene_intron_counts = collections.defaultdict(int)
# transcript_counts = collections.defaultdict(int)
# def processGroupedHits(hit_genes, hits):
# if len(hit_genes) > 1:
# print '* reads hit multiple genes:', [g.name for g in hit_genes], len(hits)
# elif len(hit_genes) == 1:
# print '* reads hit single gene:', [g.name for g in hit_genes], len(hits)
# hits.sort()
# for (lo, hi), extents, read in hits:
# aln_bases = sum([ e.size[0] for e in extents ])
# best = []
# for g in hit_genes:
# for t in g.transcripts:
# overlap = transcriptAlignmentOverlap(t, extents)
# if len(best) == 0 or best[0][0] < overlap:
# best = []
# best.append((overlap, read, g, t))
# frac = 0.0
# if len(best):
# if best[0][0] == 0:
# best = []
# else:
# frac = best[0][0] / float(aln_bases)
# if len(best) == 1:
# if frac == 1.0:
# gene_counts[g.name] += 1
# transcript_counts[t.name] += 1
# else:
# # hits only one transcript of one gene, but although it
# # overlaps one or more exons, it has some sequence that aligns
# # to an intron, or outside the gene.
# print 'TRN', frac, read[0].qname, read[1].qname, '->', g.name, t.name
# print ' ', extents, [ x.size[0] for x in extents ]
# print ' ', read[0]
# print ' ', read[1]
# elif len(best) > 1:
# g = list(set([ g for o, r, g, t in best ]))
# if len(g) == 1:
# if frac == 1.0:
# gene_counts[g[0].name] += 1
# else:
# # hits many transcripts of one gene equally well, but
# # although it overlaps one or more exons, it has some
# # sequence that aligns to an intron, or outside the gene.
# print 'GEN', frac, read[0].qname, read[1].qname, '->', g[0].name, '!' if len(hit_genes) > 1 else '', [ t.name for o, r, g, t in best ]
# print ' ', extents, [ x.size[0] for x in extents ]
# print ' ', read[0]
# print ' ', read[1]
# else:
# # hits multiple genes.
# # assign to one based upon the unambiguous transcription
# # evidence for the genes/transcripts to which the read pair
# # matched.
# print 'AMB', frac, read[0].qname, read[1].qname, '->', [ x.name for x in hit_genes ]
# print ' ', extents, [ x.size[0] for x in extents ]
# print ' ', read[0]
# print ' ', read[1]
# elif len(best) == 0:
# if len (hit_genes) == 1:
# # hits one gene in an intron.
# gene_intron_counts[hit_genes[0].name] += 1
# else:
# # hits more than one gene in an intron.
# print 'INT', frac, read[0].qname, read[1].qname, '->', [ x.name for x in hit_genes ]
# print ' ', extents, [ x.size[0] for x in extents ]
# print ' ', read[0]
# print ' ', read[1]
# samfile = pysam.Samfile("sorted.bam", "rb")
# N = 100000
# C = 0
# gap_counts = collections.defaultdict(int)
# while 1:
# grp = collections.defaultdict(list)
# read_name = {}
# def processSingleton(r):
# # we can only do this once we've fully processed a bam file, and found the unpaired reads.
# r_hits = readOverlappingGenes(r)
# def processPair(r1, r2):
# r1_ch, r1_ex = samfile.getrname(r1.tid), cigarToRefRange(r1.pos, r1.cigar)
# r2_ch, r2_ex = samfile.getrname(r2.tid), cigarToRefRange(r2.pos, r2.cigar)
# if r1_ch != r2_ch:
# print '*', r1.qname, 'and', r2.qname, 'map to different chromosomes'
# else:
# r1_hits = readOverlappingGenes(r1_ch, r1_ex)
# r2_hits = readOverlappingGenes(r2_ch, r2_ex)
# if len(r1_hits) and len(r2_hits):
# # if there's an intersection between the genes that r1 and r2 hits, take that as the hit gene set.
# rboth_hits = tuple(sorted(r1_hits & r2_hits))
# if len(rboth_hits):
# extents = snppy.range.union(r1_ex, r2_ex)
# grp[rboth_hits].append(((extents[0].extents[0][0], extents[-1].extents[0][1]), extents, (r1, r2)))
# else:
# print '*', r1.qname, 'and', r2.qname, 'map to distinct gene sets', [ g.name for g in r1_hits ], 'and', [ g.name for g in r2_hits ]
# else:
# r_hits = tuple(sorted(r1_hits | r2_hits))
# extents = snppy.range.union(r1_ex, r2_ex)
# grp[r_hits].append(((extents[0].extents[0][0], extents[-1].extents[0][1]), extents, (r1, r2)))
# for n in xrange(N):
# read = samfile.next()
# C = C + 1
# if read.qname in read_name:
# ra = read_name[read.qname]
# rb = read
# processPair(ra, rb)
# del read_name[read.qname]
# else:
# read_name[read.qname] = read
# ex = cigarToRefRange(read.pos, read.cigar)
# for j in range(1, len(ex)):
# junction = samfile.getrname(read.tid), ex[j-1].extents[0][1], ex[j].extents[0][0]
# gap_counts[junction] += 1
# print '* reads processed:', C
# print '* reads remaining unpaired:', len(read_name)
# if frozenset() in grp:
# print '* intergenic read pairs', len(grp[frozenset()])
# processIntergenicHits(grp.pop(frozenset()))
# n_unique = 0
# n_multiple = 0
# for k, v in grp.iteritems():
# if len(k) > 1:
# n_multiple += len(v)
# else:
# n_unique += len(v)
# processGroupedHits(k, v)
# print '* read pairs mapping to a single gene', n_unique
# print '* read pairs mapping to multiple genes', n_multiple
# for k, v in sorted(gap_counts.items(), key = lambda k: -k[1]):
# print '%d\t%s' % (v, k)
# samfile.close()
if __name__ == '__main__':
samfile = pysam.Samfile("sorted.bam", "rb")
counter = CountUniqueGeneFragments(GTF_FILE)
for read in samfile:
counter.addRead(read)
if counter.C == 100000:
break
counter.done()
counter.dump(sys.stdout)
samfile.close()