-
Notifications
You must be signed in to change notification settings - Fork 2
/
count_junctions.py
198 lines (158 loc) · 7.83 KB
/
count_junctions.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
#!/usr/bin/env python
'''
Counts reads for each exon-exon, exon-intron and intron-exon junctions for all introns in a BED file
Usage: count_junctions.py bamfile introns_bedfile [--min_overlap, --expand_size, --reportOverlapLength]
For this way of counting its easiest to provide a bed file for the introns.
The bam file needs to be sorted and have an index file. The bed file does not need to be sorted
writes simply the 6-column bed file adding columns EE_count, SD_count, SA_count, intronic, ambigous counts as additional columns
where
. SD is splice-donor ie 5'SS (EI on plus strand and IE on minus strand)
. SA is splice-acceptor ie 3'SS (IE on plus strand and EI on minus strand).
if --reportOverlapLength is TRUE:
adds additional columns consisting of comma-separated lists for length of EE_up EE_dn SD_up SD_dn SA_up SA_dn overlaps
'''
__author__ = 'schmidm'
import sys
import argparse
import HTSeq
parser = argparse.ArgumentParser(usage=__doc__)
parser.add_argument('bam_file')
parser.add_argument('bed_file')
parser.add_argument('--min_overlap', default=2, type=int, help="minimum base pair overlap required upstream and downstream of junction")
parser.add_argument('--expand_size', default=100, type=int, help="expand intron by this size on each size in first pass filtering of reads and only consider those reads for classification")
parser.add_argument('--reportOverlapLength', default=False, action="store_true", help="report length of upstream and downstream overlaps for each overlap type")
args = parser.parse_args()
bam_file = sys.argv[1]
bed_file = sys.argv[2]
MIN_OVERLAP = args.min_overlap #minimum amount of bases on each side of a junction that need to be present for a call
WINDOW_SIZE = args.expand_size # only consider reads with start or end within intron +/- WINDOW_SIZE usually read-length
class INTERVAL:
def __init__(self, line):
self.chr = line[0]
self.start = int(line[1])
self.end = int(line[2])
self.id = line[3]
self.score = line[4]
self.strand = line[5]
self.EE = [] #exon-exon splice junction reads
self.SA = [] #splice acceptor ie 5'SS exon-intron junction reads
self.SD = [] #splice donor ie 5'SS exon-intron junction reads
self.ambigous = 0
self.intronic = 0
def add_read(self, iv):
if iv.start < self.end:
# upstream or overlapping
if (self.start - WINDOW_SIZE) < iv.start < (self.start - MIN_OVERLAP) and iv.end > self.start:
# alignment starts upstream and ends internal or downstream
if (self.start + MIN_OVERLAP) < iv.end < self.end:
self.add_EI(iv)
elif (self.end + MIN_OVERLAP) < iv.end < (self.end + WINDOW_SIZE) :
self_found = False
# make sure there is no internal alignment to intron
for cigop in aligned_read.cigar:
if cigop.type == "M" and self.overlaps_internal(cigop.ref_iv):
self_found = True
break
if not self_found:
self.add_EE(iv)
else:
self.ambigous += 1
else:
self.ambigous += 1
elif self.start < iv.start < (self.end - MIN_OVERLAP):
# alignment starts inside
if iv.end < self.end:
# alignments ends internal --> intronic
self.intronic += 1
elif (self.end + MIN_OVERLAP) < iv.end < (self.end + WINDOW_SIZE) :
self.add_IE(iv)
else:
self.ambigous += 1
def add_EE(self, iv):
self.EE.append(iv)
#self.EE += 1
#self.EEdists.append((self.start - iv.start, iv.end - self.end))
def add_IE(self, iv):
if self.strand == '+':
self.SA.append(iv)
#self.SA += 1
#self.SAdists.append((self.end - iv.start, iv.end - self.end))
else:
self.SD.append(iv)
#self.SD += 1
#self.SDdists.append((self.start - iv.start, iv.end - self.start))
def add_EI(self, iv):
if self.strand == '+':
self.SD.append(iv)
#self.SA += 1
#self.SAdists.append((self.end - iv.start, iv.end - self.end))
else:
self.SA.append(iv)
#self.SD += 1
#self.SDdists.append((self.start - iv.start, iv.end - self.start))
def overlaps_internal(self, iv):
''' checks whether interval has any overlap with the bedinterval,
ie whether there are intronic parts in the interval '''
if iv.start < self.start and iv.end > self.start:
return True
elif iv.start < self.end and iv.end > self.end:
return True
elif iv.start > self.start and iv.end < self.end:
return True
return False
def summary_str(self):
'''
:return: summary of object, ie counts per junction as string
'''
return '\t'.join([self.chr, str(self.start), str(self.end), self.id, self.score, self.strand, str(len(self.EE)), str(len(self.SD)), str(len(self.SA)), str(self.intronic), str(self.ambigous)])
def overlaps_as_str(self):
'''
:return: tab separate list of comma-separated lists for upstream and downstream overlaps for EE, SD, SA
'''
if self.strand == '+':
ee_up = ','.join(str(self.start - ee.start) for ee in self.EE)
ee_dn = ','.join(str(ee.end - self.end) for ee in self.EE)
sd_up = ','.join(str(self.start - ee.start) for ee in self.SD)
sd_dn = ','.join(str(ee.end - self.start) for ee in self.SD)
sa_up = ','.join(str(self.end - ee.start) for ee in self.SA)
sa_dn = ','.join(str(ee.end - self.end) for ee in self.SA)
else:
sa_dn = ','.join(str(self.start - ee.start) for ee in self.SA)
sa_up = ','.join(str(ee.end - self.start) for ee in self.SA)
sd_dn = ','.join(str(self.end - ee.start) for ee in self.SD)
sd_up = ','.join(str(ee.end - self.end) for ee in self.SD)
ee_dn = ','.join(str(self.start - ee.start) for ee in self.EE)
ee_up = ','.join(str(ee.end - self.end) for ee in self.EE)
return ee_up + '\t' + ee_dn + '\t' + sd_up + '\t' + sd_dn + '\t' + sa_up + '\t' + sa_dn
##parse the bed file
bed_lines = [ INTERVAL(line.rstrip().split( "\t" )) for line in open( bed_file ) ]
##load the bam file
almnt_file = HTSeq.BAM_Reader( bam_file )
unaligned_count = 0
outside_range_count = 0
i = 0
if not args.reportOverlapLength:
print 'chr', '\t', 'start', '\t', 'end', '\t', 'id', '\t', 'score', '\t', 'strand', '\t', 'EE', '\t', 'SD', '\t', 'SA', '\t', 'intronic', '\t','ambigous'
else:
print 'chr', '\t', 'start', '\t', 'end', '\t', 'id', '\t', 'score', '\t', 'strand', '\t', 'EE', '\t', 'SD', '\t', 'SA', '\t', 'intronic', '\t', 'ambigous', '\t', 'EE_up', '\t', 'EE_dn',' \t', 'SD_up', '\t', 'SD_dn', '\t', 'SA_up', '\t', 'SA_dn'
for intron in bed_lines:
if intron.strand == '+':
read_strand = '-'
else:
read_strand = '+'
intron_window = HTSeq.GenomicInterval( intron.chr, intron.start - WINDOW_SIZE, intron.end + WINDOW_SIZE, read_strand )
for aligned_read in almnt_file[intron_window]:
if not aligned_read.aligned:
unaligned_count += 1
continue
iv = aligned_read.iv
if not aligned_read.optional_field('NH') == 1:
continue
if not iv.strand == read_strand:
#not: almnt_file[intron_window] ignores the strand info even though its there
continue
intron.add_read(iv)
if not args.reportOverlapLength:
print intron.summary_str()
else:
print intron.summary_str() + '\t' + intron.overlaps_as_str()