-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathhic_qc.py
executable file
·1155 lines (974 loc) · 56.4 KB
/
hic_qc.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
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#!/usr/bin/env python
# takes a bam file and makes a histogram of distances between mate alignments to the
# reference assembly
# takes the first 1M read pairs by default
### USAGE:
# python hic_qc.py -b <BAM_FILE> -n <NUM_READS_TO_USE> -o <outfile_stub>
# creates files in the working directory with relevant plots, also text files of statistics.
# flip -r flag (assuming you have dependencies) to make a PDF report with everything together.
import argparse
from collections import Counter
import filecmp
import importlib.metadata
import logging
import markdown as md
import os
import re
import sys
import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import pdfkit
import pysam
from scipy import optimize
matplotlib.use('Agg')
HICQC_VERSION = importlib.metadata.version('hic_qc')
#try:
# FileNotFoundError
#except NameError:
# FileNotFoundError = IOError
# default QC thresholds if there is no thresholds file
DEFAULT_MIN_SAME_STRAND_HQ_PERCENTAGE = 0.015
DEFAULT_MIN_INFORMATIVE_READ_PAIRS_PERCENTAGE = 0.05
DEFAULT_MAX_NONINFORMATIVE_READ_PAIR_PERCENTAGE = 0.50
DEFAULT_MIN_LONG_CONTACT_PERCENTAGE = 0.03
DEFAULT_MIN_INTERCONTIG_CONTACT_PERCENTAGE = 0.025
DEFAULT_MIN_USABLE_READS_PER_CONTIG = 600
DEFAULT_MIN_USABLE_READS_PER_CONTIG_PER_MILLION = 100000
DEFAULT_MAX_DUPE_PERCENTAGE = 0.40
DEFAULT_MAX_ZERO_DIST_PERCENTAGE = 0.20
DEFAULT_MAX_ZERO_MAPQ0_PERCENTAGE = 0.20
DEFAULT_MAX_UNMAPPED_PERCENTAGE = 0.10
def calc_nxx(header, xx=50):
'''Calculate the NXX (typically N50) of an assembly given a pysam.AlignmentHeader object.
Args:
header (pysam.AlignmentHeader): the header from which to extract reference sequence lengths
xx (float): the XX of NXX. we assume it is N50 so the default is 50.
Returns:
contig_len (int): the NXX (probably N50) of the assembly
total (int): The total length of contigs in the assembly
'''
frac = xx / 100.0
lens = [contig['LN'] for contig in header['SQ']]
lens.sort()
total = sum(lens)
nxx_len = total * frac
cumsum = 0
contig_len = 0
for length in reversed(lens):
contig_len = int(length)
cumsum += length
if cumsum >= nxx_len:
break
return contig_len, total
class HiCQC(object):
'''Class for extracting QC metrics from bam files and using them
to create plots, print results, save tables, and create pdf reports.
'''
def __init__(self, outfile_prefix='Read_mate_dist', sample_type='genome', thresholds_file=None,
rp_stats=None, mq_stats=None, edist_stats=None, lib_enzyme=None):
'''Initialize metrics for later extraction and conversion.
'''
logging.basicConfig(format="[%(name)s - %(asctime)s] %(message)s", level=logging.INFO)
self.logger = logging.getLogger("hic_qc")
self.sample_type = sample_type.lower()
self.qc_purpose = 'Unknown'
self.lib_enzyme = lib_enzyme if lib_enzyme is not None else ['undefined']
self.ref_assembly = "reference assembly not found"
self.fwd_hic_reads = "forward Hi-C reads not found"
self.rev_hic_reads = "reverse Hi-C reads not found"
if self.sample_type == 'metagenome':
self.qc_purpose = 'Metagenome Deconvolution'
elif self.sample_type == 'genome':
self.qc_purpose = 'Genome Scaffolding'
if thresholds_file is None:
self.min_same_strand_hq_percentage = DEFAULT_MIN_SAME_STRAND_HQ_PERCENTAGE
self.min_informative_read_pairs_percentage = DEFAULT_MIN_INFORMATIVE_READ_PAIRS_PERCENTAGE
self.max_noninformative_read_pair_percentage = DEFAULT_MAX_NONINFORMATIVE_READ_PAIR_PERCENTAGE
self.min_long_contact_percentage = DEFAULT_MIN_LONG_CONTACT_PERCENTAGE
self.min_intercontig_contact_percentage = DEFAULT_MIN_INTERCONTIG_CONTACT_PERCENTAGE
self.min_usable_reads_per_contig = DEFAULT_MIN_USABLE_READS_PER_CONTIG
self.min_usable_reads_per_contig_per_million = DEFAULT_MIN_USABLE_READS_PER_CONTIG_PER_MILLION
self.max_dupe_percentage = DEFAULT_MAX_DUPE_PERCENTAGE
self.max_zero_dist_percentage = DEFAULT_MAX_ZERO_DIST_PERCENTAGE
self.max_zero_mapq0_percentage = DEFAULT_MAX_ZERO_MAPQ0_PERCENTAGE
self.max_unmapped_percentage = DEFAULT_MAX_UNMAPPED_PERCENTAGE
else:
import json
with open(thresholds_file) as f:
thresholds = json.load(f)
self.min_same_strand_hq_percentage = float(thresholds[sample_type]['MIN_SAME_STRAND_HQ_PERCENTAGE'])
self.min_informative_read_pairs_percentage = float(thresholds[sample_type]['MIN_INFORMATIVE_READ_PAIRS_PERCENTAGE'])
self.max_noninformative_read_pair_percentage = float(thresholds[sample_type]['MAX_NONINFORMATIVE_READ_PAIR_PERCENTAGE'])
self.min_long_contact_percentage = float(thresholds[sample_type]['MIN_LONG_CONTACT_PERCENTAGE'])
self.min_intercontig_contact_percentage = float(thresholds[sample_type]['MIN_INTERCONTIG_CONTACT_PERCENTAGE'])
self.min_usable_reads_per_contig = float(thresholds[sample_type]['MIN_USABLE_READS_PER_CONTIG'])
self.min_usable_reads_per_contig_per_million = float(thresholds[sample_type]['MIN_USABLE_READS_PER_CONTIG_PER_MILLION'])
self.max_dupe_percentage = float(thresholds[sample_type]['MAX_DUPE_PERCENTAGE'])
self.max_zero_dist_percentage = float(thresholds[sample_type]['MAX_ZERO_DIST_PERCENTAGE'])
self.max_zero_mapq0_percentage = float(thresholds[sample_type]['MAX_ZERO_MAPQ0_PERCENTAGE'])
self.max_unmapped_percentage = float(thresholds[sample_type]['MAX_UNMAPPED_PERCENTAGE'])
self.per_read_metrics = set(['duplicate_reads', 'mapq0_reads', 'split_reads', 'total_reads', 'unmapped_reads',])
self.per_pair_metrics = set(['different_ref_stub_pairs',
'informative_pairs',
'intercontig_pairs',
'intercontig_pairs_hq',
'pairs_greater_10k',
'pairs_greater_10k_on_contigs_greater_10k',
'pairs_intracontig_hq',
'pairs_intracontig_hq_gt10kbp',
'pairs_on_contigs_greater_10k',
'pairs_on_same_strand',
'pairs_on_same_strand_hq',
'proximo_usable_rp',
'proximo_usable_rp_hq',
'total_pairs_on_same_contig',
'total_read_pairs',
'total_read_pairs_hq',
'zero_dist_pairs',
])
# Dictionary of key --> numerator, denominator pairs for stringify_stats
self.to_percents = {
# Top Table
'perc_pairs_on_same_strand_hq': ('pairs_on_same_strand_hq', 'pairs_intracontig_hq'),
'perc_informative_read_pairs': ('informative_pairs', 'total_read_pairs'),
# Other Good Metrics
'perc_pairs_intra_hq_gt10kbp': ('pairs_intracontig_hq_gt10kbp', 'pairs_intracontig_hq'),
'perc_intercontig_pairs_hq_gt10kbp': ('pairs_intercontig_hq_gt10kbp', 'total_read_pairs_hq'),
# Usable reads number is in self.to_round
# Noninformative Read Breakdown
'perc_noninformative_read_pairs': ('noninformative_read_pairs', 'total_read_pairs'),
'perc_duplicate_reads': ('duplicate_reads', 'total_reads'),
'perc_zero_dist_pairs': ('zero_dist_pairs', 'total_read_pairs'),
'perc_unmapped_reads': ('unmapped_reads', 'total_reads'),
'perc_mapq0_reads': ('mapq0_reads', 'total_reads'),
# Extended Metrics
'perc_pairs_greater_10k': ('pairs_greater_10k', 'total_read_pairs'),
'perc_pairs_greater_10k_on_contigs_greater_10k': ('pairs_greater_10k_on_contigs_greater_10k', 'pairs_on_contigs_greater_10k'),
'perc_pairs_on_same_strand': ('pairs_on_same_strand', 'total_pairs_on_same_contig'),
'perc_intercontig_pairs': ('intercontig_pairs', 'total_read_pairs'),
'perc_intercontig_pairs_hq': ('intercontig_pairs_hq', 'total_read_pairs_hq'),
'perc_split_reads': ('split_reads', 'total_reads'),
'perc_hq_rp': ('total_read_pairs_hq', 'total_read_pairs'),
'perc_different_ref_stub_pairs': ('different_ref_stub_pairs', 'total_read_pairs')
}
self.to_round = set([
'proximo_usable_rp_per_ctg_gt_5k',
'proximo_usable_rp_hq_per_ctg_gt_5k',
'proximo_usable_rp_hq_per_ctg_gt_5k_per_million'
])
self.convert_to_pairs = set(['unmapped_reads', 'split_reads', 'duplicate_reads', 'mapq0_reads'])
self.paths = {'script_dir': os.path.dirname(__file__), 'outfile_prefix': outfile_prefix}
self.paths['pg_logo'] = os.path.join(self.paths['script_dir'], 'collateral', 'PGBlueLogoHorSmall.png')
self.N50 = None
self.stats = Counter()
self.dists = Counter()
self.total_array = []
self.non_dup_array = []
if rp_stats is not None and mq_stats is not None and edist_stats is not None:
rps = []
mqs = []
edists = []
self.mapping_dict = {}
rp_stats = sorted(map(lambda x: int(1000 * x), rp_stats))
mq_stats = sorted(map(int, mq_stats))
edist_stats = list(reversed(sorted(map(int, edist_stats))))
for rp in rp_stats:
self.mapping_dict[rp] = {}
for mq in mq_stats:
self.mapping_dict[rp][mq] = {}
for ed in edist_stats:
self.mapping_dict[rp][mq][ed] = 0
self.rp_stats = rp_stats
self.mq_stats = mq_stats
self.edist_stats = edist_stats
else:
self.mapping_dict = None
self.rp_array = None
def parse_bam(self, bamfile, max_read_pairs=-1):
'''Extract QC metrics from a specified bam file. It requires a read name sorted bam file.
By default, it will parse all reads in the bam file, but a limit can be specified by max_read_pairs.
This method performs read pairing.
'''
if max_read_pairs != -1:
self.logger.info('parsing the first {} read pairs in bam file {} '\
'to QC Hi-C library quality'.format(max_read_pairs, bamfile))
else:
self.logger.info('parsing all read pairs in bam file {} '\
'to QC Hi-C library quality'.format(bamfile))
self.paths['bamfile'] = bamfile
self.paths['bamname'] = os.path.basename(bamfile)
a = None
b = None
i = 0
with pysam.AlignmentFile(self.paths['bamfile']) as bam_fh:
self.extract_header_info(bam_fh.header)
for read in bam_fh:
if read.is_secondary or read.is_supplementary:
continue
if a is None:
a = read
continue
if max_read_pairs != -1 and i / 2 > max_read_pairs:
break
if read.query_name == a.query_name:
b = read
self.process_pair(a, b)
a = None
b = None
else:
a = read
if i % 1000 == 0:
self.update_dup_stats()
i += 1
self.finalize_stats()
def extract_header_info(self, header, xx=50):
'''Extract reference names, calculate N50, get total assembly length, and get set of contigs > 10kbp from a pysam header.
Also checks if input bamfile is labeled as coordinate sorted and throws a ValueError if True.
Args:
header (pysam.AlignmentFile.header): the header from which to extract reference sequence lengths
xx (float): the XX of NXX. we assume it is N50 so the default is 50
Uses:
self.mapping_dict (dict(int-->dict(int-->dict(int-->int)))): Nested dict with min_size, mapq, and edist keys
and read pair counts as the inner values.
Sets:
self.refs dict(int-->str): ref_id --> ref_name mappings for the assembly
self.contig_len (int): the NXX (probably N50) of the assembly based on the header
self.total (int): the total length of the assembly
self.contigs_greater_10k (set(str)): The set of names of contigs with length > 10Kbp
self.contigs_greater_5k (set(str)): The set of names of contigs with length > 5Kbp
self.contigs_greater (dict(int-->set(str))): Dictionary with minimum lengths as keys and sets of contigs as values
self.command_line(str): Full command-line argument used for alignment
self.bwa_command(str): Subset of self.command_line containing only the BWA options used
self.samblaster(str): Command used by samblaster
Raises:
ValueError if header labels bamfile as coordinate sorted
'''
if 'HD' in header and 'SO' in header['HD'] and header['HD']['SO'].lower().strip() == 'coordinate':
raise ValueError('Error: hic_qc.py requires read name sorted input, but bamfile {} is coordinate sorted'.format(self.paths['bamfile']))
self.refs = header.references
self.stats['num_refs'] = len(self.refs)
self.N50, self.total_length = calc_nxx(header)
self.contigs_greater_10k = set([contig['SN'] for contig in header['SQ'] if contig['LN'] > 10000])
self.contigs_greater_5k = set([contig['SN'] for contig in header['SQ'] if contig['LN'] > 5000])
self.contigs_greater = {}
if self.mapping_dict is not None:
for min_size in self.mapping_dict.keys():
self.contigs_greater[min_size] = set([contig['SN'] for contig in header['SQ'] if contig['LN'] > min_size])
# TODO: add more robust logic to different BAM headers, and/or comment the assumptions made in this code
if 'PG' in header and 'bwa' in header['PG'][0]['CL']:
self.bwa_command_line = header['PG'][0]['CL']
if 'bwa-mem2' in self.bwa_command_line:
self.bwa_command = re.search(r'(bwa-mem2 )[^//]*', self.bwa_command_line).group()
else:
self.bwa_command = re.search(r'(bwa )[^//]*', self.bwa_command_line).group()
bwa_command_elements = self.bwa_command_line.split()
full_fwd_reads = None
full_rev_reads = None
for token in bwa_command_elements:
token_proc = token.strip().lower()
if token_proc.endswith('.fasta') or token_proc.endswith('.fa') or token_proc.endswith('.fna') \
or token_proc.endswith('.fasta.gz') or token_proc.endswith('.fa.gz') or token_proc.endswith('fna.gz'):
self.ref_assembly = os.path.basename(token)
elif token_proc.endswith('_r1.fastq') or token_proc.endswith('_r1.fq') \
or token_proc.endswith('_r1.fastq.gz') or token_proc.endswith('_r1.fq.gz'):
self.fwd_hic_reads = os.path.basename(token)
full_fwd_reads = token.strip()
elif token_proc.endswith('_r2.fastq') or token_proc.endswith('_r2.fq') \
or token_proc.endswith('_r2.fastq.gz') or token_proc.endswith('_r2.fq.gz'):
self.rev_hic_reads = os.path.basename(token)
full_rev_reads = token.strip()
files_matched = False
if full_fwd_reads is not None and os.path.exists(full_fwd_reads) and \
full_rev_reads is not None and os.path.exists(full_rev_reads):
if filecmp.cmp(full_fwd_reads, full_rev_reads, shallow=True):
files_matched = True
if os.path.exists(self.fwd_hic_reads) and os.path.exists(self.rev_hic_reads):
#do a shallow check first - if the file stats don't match, the files don't match, so don't do a full comparison
if filecmp.cmp(self.fwd_hic_reads, self.rev_hic_reads, shallow=True):
files_matched = True
if files_matched or (self.fwd_hic_reads is not None and self.fwd_hic_reads == self.rev_hic_reads) or (full_fwd_reads is not None and full_fwd_reads == full_rev_reads) \
or (self.fwd_hic_reads == "forward Hi-C reads not found") or (self.rev_hic_reads == "reverse Hi-C reads not found"):
self.fwd_hic_reads = '<span class="mixed-results">{0}</span>'.format(self.fwd_hic_reads)
self.rev_hic_reads = '<span class="mixed-results">{0}</span>'.format(self.rev_hic_reads)
else:
self.bwa_command = 'BWA command not found'
if 'PG' in header and len(header['PG']) > 1 and 'samblaster ' in header['PG'][1]['CL']:
self.samblaster = header['PG'][1]['CL']
else:
self.samblaster = 'samblaster command not found'
def process_pair(self, a, b):
'''Extract stats from a pair of reads.
Updates per read and per mapped pair stats separately.
Args:
a (pysam.AlignedSegment) One read in pair
b (pysam.AlignedSegment) Other read in pair
'''
self.stats['total_read_pairs'] += 1
self.update_read_stats(a)
self.update_read_stats(b)
if self.is_noninformative_read_pair(a, b):
self.stats['noninformative_read_pairs'] += 1
elif self.is_informative_pair(a, b):
self.stats['informative_pairs'] += 1
if not a.is_unmapped and not b.is_unmapped:
self.update_mapped_pair_stats(a, b)
def update_read_stats(self, read):
'''Update per read stats based on given read.
Args:
read (pysam.AlignedSegment): read to extract stats from
'''
self.stats['total_reads'] += 1
if read.is_unmapped:
self.stats['unmapped_reads'] += 1
elif read.mapping_quality == 0:
self.stats['mapq0_reads'] += 1
if read.has_tag('SA'):
self.stats['split_reads'] += 1
if read.is_duplicate:
self.stats['duplicate_reads'] += 1
def update_mapped_pair_stats(self, a, b):
'''Update mapped pair stats given a pair of reads.
Args:
a (pysam.AlignedSegment): One read
b (pysam.AlignedSegment): Other read
'''
is_high_qual_pair = self.is_high_qual_pair(a, b)
if is_high_qual_pair:
self.stats['total_read_pairs_hq'] += 1
if a.reference_name != b.reference_name:
self.stats['intercontig_pairs'] += 1
if is_high_qual_pair:
self.stats['intercontig_pairs_hq'] += 1
if a.reference_name in self.contigs_greater_10k and b.reference_name in self.contigs_greater_10k:
self.stats['pairs_intercontig_hq_gt10kbp'] += 1
refa_stub = a.reference_name.split('.')[0]
refb_stub = b.reference_name.split('.')[0]
if refa_stub != refb_stub:
self.stats['different_ref_stub_pairs'] += 1
if a.reference_name in self.contigs_greater_5k and b.reference_name in self.contigs_greater_5k:
if min(a.mapping_quality, b.mapping_quality) > 0 and not any([a.is_duplicate, b.is_duplicate]):
self.stats['proximo_usable_rp'] += 1
if is_high_qual_pair:
self.stats['proximo_usable_rp_hq'] += 1
else:
self.stats['total_pairs_on_same_contig'] += 1
if (a.is_reverse and b.is_reverse) or (not a.is_reverse and not b.is_reverse):
self.stats['pairs_on_same_strand'] += 1
dist = abs(a.reference_start - b.reference_start)
self.dists[dist] += 1
if is_high_qual_pair:
self.stats['pairs_intracontig_hq'] += 1
if dist > 10000:
self.stats['pairs_intracontig_hq_gt10kbp'] += 1
if (a.is_reverse and b.is_reverse) or (not a.is_reverse and not b.is_reverse):
self.stats['pairs_on_same_strand_hq'] += 1
if dist > 10000:
self.stats['pairs_greater_10k'] += 1
if dist == 0:
self.stats['zero_dist_pairs'] += 1
## Additional stats for dynamo table.
if dist < 0:
raise ValueError('Error: Distance between reads is less than zero.')
elif dist < 1e3:
self.stats['reads_spanning_up_to_1k'] += 1
elif dist < 1e4:
self.stats['reads_spanning_1k_to_10k'] += 1
elif dist < 1e5:
self.stats['reads_spanning_10k_to_100k'] += 1
elif dist < 1e6:
self.stats['reads_spanning_100k_to_1000k'] += 1
else:
self.stats['reads_spanning_greater_than_1000k'] += 1
if a.reference_name in self.contigs_greater_10k:
self.stats['pairs_on_contigs_greater_10k'] += 1
if is_high_qual_pair:
self.stats['pairs_on_contigs_greater_10k_hq'] += 1
if dist > 10000:
self.stats['pairs_greater_10k_on_contigs_greater_10k_hq'] += 1
if dist > 10000:
self.stats['pairs_greater_10k_on_contigs_greater_10k'] += 1
if self.mapping_dict is not None and \
not (a.is_duplicate or b.is_duplicate):
self.update_rp_array(a, b)
def update_rp_array(self, a, b):
'''Update nested dict of mapping stats based on current read pair a, b.
Uses:
self.rp_stats (list(int)): List of minimum insert sizes sorted from low to high
self.mq_stats (list(int)): List of minimum mapq values sorted from low to high
self.edist_stats (list(int)): List of maximum edit distances sorted from high to low
Modifies:
self.mapping_dict (dict(int-->dict(int-->dict(int-->int)))): Nested dict with min_size, mapq, and edist keys
and read pair counts as the inner values.
'''
mq = min(a.mapping_quality, b.mapping_quality)
isize = abs(a.reference_start - b.reference_start)
edist = max(a.get_tag('NM'), b.get_tag('NM'))
for min_size in self.rp_stats:
if isize >= min_size:
for mapq in self.mq_stats:
if mq >= mapq:
for ed in self.edist_stats:
if edist <= ed:
self.mapping_dict[min_size][mapq][ed] += 1
def is_noninformative_read_pair(self, a, b):
is_noninformative = False
if a.is_unmapped or b.is_unmapped:
is_noninformative = True
elif a.is_duplicate or b.is_duplicate:
is_noninformative = True
elif a.mapping_quality == 0 or b.mapping_quality == 0:
is_noninformative = True
elif a.reference_name == b.reference_name:
if abs(a.reference_start - b.reference_start) == 0:
is_noninformative = True
return is_noninformative
def is_informative_pair(self, a, b):
is_informative = True
if a.is_unmapped or b.is_unmapped:
is_informative = False
elif a.is_duplicate or b.is_duplicate:
is_informative = False
elif a.mapping_quality == 0 or b.mapping_quality == 0:
is_informative = False
elif a.reference_name == b.reference_name and \
abs(a.reference_start - b.reference_start) < 10000:
is_informative = False
return is_informative
def is_high_qual_pair(self, a, b):
return min(a.mapping_quality, b.mapping_quality) >= 20 and \
max(a.get_tag('NM'), b.get_tag('NM')) <= 5 and \
not a.is_duplicate and not b.is_duplicate
def update_dup_stats(self):
'''Update lists of duplication statistics.
'''
self.total_array.append(self.stats['total_reads'])
self.non_dup_array.append(self.stats['total_reads'] - self.stats['duplicate_reads'])
def finalize_stats(self):
'''Finalize stats from a bam file after reads are processed.
'''
self.total_array = np.array(self.total_array)
self.non_dup_array = np.array(self.non_dup_array)
if self.stats['pairs_on_contigs_greater_10k'] > 0:
self.stats['proportion_pairs_greater_10k_on_contigs_greater_10k'] = self.stats['pairs_greater_10k_on_contigs_greater_10k'] / \
self.stats['pairs_on_contigs_greater_10k']
else:
self.stats['proportion_pairs_greater_10k_on_contigs_greater_10k'] = 0
if len(self.contigs_greater_5k) > 0:
self.stats['proximo_usable_rp_per_ctg_gt_5k'] = self.stats['proximo_usable_rp'] / len(self.contigs_greater_5k)
self.stats['proximo_usable_rp_hq_per_ctg_gt_5k'] = self.stats['proximo_usable_rp_hq'] / len(self.contigs_greater_5k)
self.stats['proximo_usable_rp_hq_per_ctg_gt_5k_per_million'] = self.stats['proximo_usable_rp_hq_per_ctg_gt_5k'] / (self.stats['total_read_pairs'] / 1e6)
else:
self.stats['proximo_usable_rp_per_ctg_gt_5k'] = 0
self.stats['proximo_usable_rp_hq_per_ctg_gt_5k'] = 0
self.stats['proximo_usable_rp_hq_per_ctg_gt_5k_per_million'] = 0
if self.mapping_dict is not None:
self.write_mapping_stats()
# We are stricter on wanting a low number of dupes when it looks like we are only looking at a QC amount of sequencing (<10M read pairs)
self.allowed_dupe_percentage = 1.0 if self.stats['total_read_pairs'] > 1e7 else 0.5
def write_mapping_stats(self):
with open('{}.mapping_stats.tsv'.format(self.paths['outfile_prefix']), 'w') as outfile:
print('edist', 'mapq', 'min_size', 'count', sep='\t', file=outfile)
for min_size in self.rp_stats:
for mapq in self.mq_stats:
for ed in self.edist_stats:
count = self.mapping_dict[min_size][mapq][ed]
print(ed, mapq, min_size, count, sep='\t', file=outfile)
if self.mapping_dict is not None:
self.write_mapping_stats()
def write_mapping_stats(self):
with open('{}.mapping_stats.tsv'.format(self.paths['outfile_prefix']), 'w') as outfile:
print('edist', 'mapq', 'min_size', 'count', sep='\t', file=outfile)
for min_size in self.rp_stats:
for mapq in self.mq_stats:
for ed in self.edist_stats:
count = self.mapping_dict[min_size][mapq][ed]
print(ed, mapq, min_size, count, sep='\t', file=outfile)
def plot_histograms(self):
'''Make the read distance long, short, and log_log histograms using matplotlib and write them to disk.
Args:
self.dists (dict(int, int) of mate distances and counts): Distances to plot in histogram.
self.num_pairs (int): number of read pairs analyzed
Uses:
self.paths['outfile_prefix'] (str): Path prefix for output files.
'''
num_dists = sum(self.dists.values())
num_pairs = self.stats['total_read_pairs']
title_string = '\nMate distance distribution for first {} read pairs for sample\n{}'.format(num_pairs,
self.paths['bamname'])
key_len = len(self.dists)
long_hist_path = self.paths['outfile_prefix'] + '_long.png'
fig1, ax = plt.subplots(1)
if key_len > 0:
plt.hist(list(self.dists.keys()), weights=list(self.dists.values()), bins=50, edgecolor='black', color='red')
ax.set_ylim(0.5, max(num_dists * 2, 1))
plt.yscale('log', nonpositive='clip')
plt.title(title_string)
plt.xlabel('Distance between read pair mates in Hi-C mapping (same contig)')
plt.ylabel('Number of reads')
else:
plt.title('Warning: No read pair distribution to plot')
fig1.savefig(long_hist_path)
plt.close(fig1)
self.paths['long_hist'] = long_hist_path
fig2, ax = plt.subplots(1)
short_hist_path = self.paths['outfile_prefix'] + '_short.png'
if key_len > 0:
plt.hist(list(self.dists.keys()), weights=list(self.dists.values()), bins=range(0, 20000, 500), edgecolor='black', color='red')
ax.set_xlim(0, 20000)
ax.set_ylim(0.5, num_pairs * 2)
plt.yscale('log', nonpositive='clip')
plt.title(title_string)
plt.xlabel('Distance between read pair mates in Hi-C mapping (same contig)')
plt.ylabel('Number of reads')
else:
plt.title('Warning: No read pair distribution to plot')
fig2.savefig(short_hist_path)
plt.close(fig2)
self.paths['short_hist'] = short_hist_path
fig3, ax = plt.subplots(1)
log_log_hist_path = self.paths['outfile_prefix'] + '_log_log.png'
offset_dists = {}
for key, value in self.dists.items():
offset_dists[key+1] = value
if key_len > 0:
min_dist = min(self.dists.keys())
max_dist = max(self.dists.keys())
plt.hist(list(offset_dists.keys()),
weights=list(offset_dists.values()),
bins=np.logspace(np.log10(min_dist+1),
np.log10(max_dist),
50),
log=True, edgecolor='black', color='red')
ax.set_ylim(0.5, max(num_pairs * 2, 1))
plt.yscale('log', nonpositive='clip')
plt.xscale('log')
plt.xlim(left=1)
plt.title(title_string)
plt.xlabel('Distance between read pair mates in Hi-C mapping (same contig, log scale)')
plt.ylabel('Number of reads (log scale)')
plt.tight_layout()
else:
plt.title('Warning: No read pair distribution to plot')
fig3.savefig(log_log_hist_path)
plt.close(fig3)
self.paths['log_log_hist'] = log_log_hist_path
fig4, ax = plt.subplots(1)
log_log_norm_hist_path = self.paths['outfile_prefix'] + '_log_log_norm.png'
offset_dists = {}
for key, value in self.dists.items():
offset_dists[key+1] = value
if key_len > 0:
min_dist = min(self.dists.keys())
max_dist = max(self.dists.keys())
plt.hist(list(offset_dists.keys()),
weights=list(offset_dists.values()),
bins=np.logspace(np.log10(min_dist+1),
np.log10(max_dist),
50),
log=True, edgecolor='black', color='red', density=True)
ax.set_ylim(0.0001, 1.0)
plt.yscale('log', nonpositive='clip')
plt.xscale('log')
plt.xlim(left=1)
plt.title(title_string)
plt.xlabel('Distance between read pair mates in Hi-C mapping (same contig, log scale)')
plt.ylabel('Density of reads (density, log scale)')
plt.tight_layout()
else:
plt.title('Warning: No read pair distribution to plot')
fig4.savefig(log_log_norm_hist_path)
plt.close(fig4)
self.paths['log_log_norm_hist'] = log_log_norm_hist_path
def html_from_judgement(self):
'''Set a formatted HTML string based on two judgment bool values
Uses:
self.judge_good (bool): does the hi-c library show characteristics of 'goodness', e.g. many long-distance contacts etc.
self.judge_bad (bool): does the hi-c library show 'bad' characteristics, e.g. zero-distance reads or too many duplicates.
Sets:
self.judge_html (str): an HTML string to be substituted into the report to subjectively grade the assembly. 4 possibilities.
Raises:
ValueError: if impossible logical situations occur given two bools.
'''
if self.judge_good and not self.judge_bad:
self.judge_html = '<span class="pass">SUFFICIENT</span>'
elif not self.judge_good and self.judge_bad:
self.judge_html = '<span class="fail">INSUFFICIENT</span>'
elif self.judge_good and self.judge_bad:
self.judge_html = '<span class="mixed-results">MIXED RESULTS</span>'
elif not self.judge_good and not self.judge_bad:
self.judge_html = '<span class="low-signal">LOW SIGNAL</span>'
else:
raise ValueError('logical impossibility!')
# driving metrics
if self.good_same_strand:
if float(self.stats['pairs_on_same_strand_hq']) / max(self.stats['pairs_intracontig_hq'], 1) > 0.999:
self.same_strand_hq_html = '<span class="mixed-results">{0}</span>'
else:
self.same_strand_hq_html = '<span class="pass">{0}</span>'
else:
self.same_strand_hq_html = '<span class="fail">{0}</span>'
if self.good_informative_read_pairs:
self.informative_read_pairs_html = '<span class="pass">{0}</span>'
else:
self.informative_read_pairs_html = '<span class="fail">{0}</span>'
# other good metrics
if self.good_long_contacts:
self.long_contacts_html = '<span class="pass">{0}</span>'
else:
self.long_contacts_html = '<span class="fail">{0}</span>'
if self.good_intercontig_contacts:
self.intercontig_hq_contacts_html = '<span class="pass">{0}</span>'
else:
self.intercontig_hq_contacts_html = '<span class="fail">{0}</span>'
if self.good_usable_reads:
self.usable_hq_gt_5k_html = '<span>{0}</span>'
else:
self.usable_hq_gt_5k_html = '<span>{0}</span>'
if self.good_usable_reads_per_million:
self.usable_hq_gt_5k__per_million_html = '<span class="pass">{0}</span>'
else:
self.usable_hq_gt_5k__per_million_html = '<span class="fail">{0}</span>'
# noninformative reads breakdown
if not self.bad_noninformative_read_pairs:
self.noninformative_read_pairs_html = '<span class="pass">{0}</span>'
else:
self.noninformative_read_pairs_html = '<span class="fail">{0}</span>'
if not self.high_dupe:
self.high_dupe_html = '<span class="pass">{0}</span>'
else:
self.high_dupe_html = '<span class="fail">{0}</span>'
if not self.many_zero_dist_pairs:
self.many_zero_dist_pairs_html = '<span class="pass">{0}</span>'
else:
self.many_zero_dist_pairs_html = '<span class="fail">{0}</span>'
if not self.many_unmapped_reads:
self.many_unmapped_reads_html = '<span class="pass">{0}</span>'
else:
self.many_unmapped_reads_html = '<span class="fail">{0}</span>'
if not self.many_mapq_zero_reads:
self.many_zero_mapq_reads_html = '<span class="pass">{0}</span>'
else:
self.many_zero_mapq_reads_html = '<span class="fail">{0}</span>'
def pass_judgement(self):
'''Pass judgement on the library according to certain mostly subjective ideas about what is good
Uses:
self.stats ({str: float/str}): mapping of lib characteristics to their values
Sets:
self.judge_good (bool): does the hi-c library show characteristics of 'goodness', e.g. many same strand contacts.
self.judge_bad (bool): does the hi-c library show 'bad' characteristics, e.g. zero-mapq reads or too many duplicates.
self.judge_html (str): an HTML string to put into pass/fail box
'''
# these metrics drive the subjective quality judgement
self.good_same_strand = float(self.stats['pairs_on_same_strand_hq']) / \
max(self.stats['pairs_intracontig_hq'], 1) > \
self.min_same_strand_hq_percentage
self.good_informative_read_pairs = float(self.stats['informative_pairs']) / \
max(self.stats['total_read_pairs'], 1) > \
self.min_informative_read_pairs_percentage
# other good metrics
self.good_long_contacts = float(self.stats['pairs_greater_10k_on_contigs_greater_10k_hq']) / \
max(self.stats['pairs_on_contigs_greater_10k_hq'], 1) > \
self.min_long_contact_percentage
self.good_intercontig_contacts = float(self.stats['pairs_intercontig_hq_gt10kbp']) / \
max(self.stats['total_read_pairs_hq'], 1) > \
self.min_intercontig_contact_percentage
self.good_usable_reads = float(self.stats['proximo_usable_rp_hq_per_ctg_gt_5k']) > \
self.min_usable_reads_per_contig
self.good_usable_reads_per_million = float(self.stats['proximo_usable_rp_hq_per_ctg_gt_5k_per_million']) > \
self.min_usable_reads_per_contig_per_million
# noninformative read breakdown
# We are stricter on wanting a low number of dupes when it looks like we are only looking at a QC amount of sequencing (<10M read pairs)
self.bad_noninformative_read_pairs = float(self.stats['noninformative_read_pairs']) / \
max(self.stats['total_read_pairs'], 1) > \
self.max_noninformative_read_pair_percentage
self.high_dupe = float(self.stats['duplicate_reads']) / \
max(self.stats['total_reads'], 1) > \
self.max_dupe_percentage * \
self.allowed_dupe_percentage
self.many_zero_dist_pairs = float(self.stats['zero_dist_pairs']) / \
max(self.stats['total_read_pairs'], 1) > \
self.max_zero_dist_percentage
self.many_unmapped_reads = float(self.stats['unmapped_reads']) / \
max(self.stats['total_reads'], 1) > \
self.max_unmapped_percentage
self.many_mapq_zero_reads = float(self.stats['mapq0_reads']) / \
max(self.stats['total_reads'], 1) > \
self.max_zero_mapq0_percentage
self.judge_good = self.good_same_strand
self.judge_bad = not self.good_informative_read_pairs
self.html_from_judgement()
def stringify_stats(self):
'''Convert stats to output dictionary with pretty strings and percents.
Uses:
self.to_percents ({str: (int, int)}): Mapping of keys to numerator, denominator pair for conversion to percents.
self.to_round (set(str)): Set of keys from stats dict to round.
self.convert_to_pairs (set(str)): Set of keys from stats dict that represent per read statistics.
self.per_pair_metrics (set(str)): Set of keys from stats dict that represent per read pair statistics.
self.paths ({str: str}): Mapping of names to paths.
Sets:
self.other_stats ({str: (float, str)}): Mapping of keys to value, format pair.
self.out_stats ({str: str}): Mapping of stat keys to formatted strings.
'''
# Dict of key --> (value, fmt) pairs for items that aren't counts
self.other_stats = {
'N50': (self.N50, '{:,}'),
'contigs': (len(self.refs), '{:,}'),
'contigs_greater_10k': (len(self.contigs_greater_10k), '{:,}'),
'contigs_greater_5k': (len(self.contigs_greater_5k), '{:,}'),
'total_length': (self.total_length, '{:,}'),
'total_reads': (self.stats['total_reads'], '{:,}'),
'judgment': (self.judge_html, '{}'),
'qc_purpose': (self.qc_purpose, '{}'),
'same_strand_threshold': (100.0 * self.min_same_strand_hq_percentage, '{}'),
'informative_read_pairs_threshold': (100.0 * self.min_informative_read_pairs_percentage, '{}'),
'noninformative_read_pairs_threshold': (100.0 * self.max_noninformative_read_pair_percentage, '{}'),
'long_contacts_threshold': (100.0 * self.min_long_contact_percentage, '{}'),
'intercontig_hq_contacts_threshold': (100.0 * self.min_intercontig_contact_percentage, '{}'),
'usable_hq_gt_5k_threshold': (self.min_usable_reads_per_contig, '{}'),
'min_usable_reads_per_contig_per_million_threshold': (self.min_usable_reads_per_contig_per_million, '{}'),
'high_dupe_threshold': (100.0 * self.max_dupe_percentage * self.allowed_dupe_percentage, '{}'),
'many_zero_dist_threshold': (100.0 * self.max_zero_dist_percentage, '{}'),
'many_zero_mapq_threshold': (100.0 * self.max_zero_mapq0_percentage, '{}'),
'many_unmapped_threshold': (100.0 * self.max_unmapped_percentage, '{}'),
'alignment_command_line': (self.bwa_command, '{}'),
'samblaster': (self.samblaster, '{}'),
'lib_enzyme': (', '.join(self.lib_enzyme), '{}'),
'ref_assembly': (self.ref_assembly, '{}'),
'fwd_hic_reads': (self.fwd_hic_reads, '{}'),
'rev_hic_reads': (self.rev_hic_reads, '{}'),
}
self.out_stats = {}
for key, (num, denom) in self.to_percents.items():
try:
self.out_stats[key] = '{:.2f}%'.format((self.stats[num] / self.stats[denom]) * 100)
except ZeroDivisionError as e:
self.out_stats[key] = 'NaN'
for key in self.to_round:
self.out_stats[key] = '{:.2f}'.format(self.stats[key])
for item in self.convert_to_pairs:
self.out_stats[item] = '{:,}'.format(self.stats[item] // 2)
for item in self.per_pair_metrics:
self.out_stats[item] = '{:,}'.format(self.stats[item])
for item in self.paths:
if item == "bamname":
self.out_stats[item] = self.paths[item]
else:
self.out_stats[item] = os.path.abspath(self.paths[item])
for key, value in self.contigs_greater.items():
key_str = 'contigs_greater_{:.0f}k'.format(key / 1000)
self.out_stats[key_str] = '{:,}'.format(len(value))
for key, (value, fmt) in self.other_stats.items():
self.out_stats[key] = fmt.format(value)
# driving stats
self.out_stats['same_strand_hq_html'] = self.same_strand_hq_html.format(self.out_stats['perc_pairs_on_same_strand_hq'])
self.out_stats['informative_read_pairs_html'] = self.informative_read_pairs_html.format(self.out_stats['perc_informative_read_pairs'])
self.out_stats['noninformative_read_pairs_html'] = self.noninformative_read_pairs_html.format(self.out_stats['perc_noninformative_read_pairs'])
# other good metrics
self.out_stats['long_contacts_html'] = self.long_contacts_html.format(self.out_stats['perc_pairs_intra_hq_gt10kbp'])
self.out_stats['intercontig_hq_contacts_html'] = self.intercontig_hq_contacts_html.format(self.out_stats['perc_intercontig_pairs_hq_gt10kbp'])
self.out_stats['usable_hq_gt_5k_html'] = self.usable_hq_gt_5k_html.format(self.out_stats['proximo_usable_rp_hq_per_ctg_gt_5k'])
self.out_stats['usable_hq_gt_5k_per_million_html'] = self.usable_hq_gt_5k__per_million_html.format(self.out_stats['proximo_usable_rp_hq_per_ctg_gt_5k_per_million'])
# noninformative breakdown
self.out_stats['high_dupe_html'] = self.high_dupe_html.format(self.out_stats['perc_duplicate_reads'])
self.out_stats['many_zero_dist_pairs_html'] = self.many_zero_dist_pairs_html.format(self.out_stats['perc_zero_dist_pairs'])
self.out_stats['many_unmapped_reads_html'] = self.many_unmapped_reads_html.format(self.out_stats['perc_unmapped_reads'])
self.out_stats['many_zero_mapq_reads_html'] = self.many_zero_mapq_reads_html.format(self.out_stats['perc_mapq0_reads'])
self.out_stats['version'] = HICQC_VERSION
def log_stats(self, count_diff_refname_stub=False):
'''Log statistical summary.
Uses:
self.paths ({str: str}): Mapping of names to paths.
self.out_stats ({str: str}): Mapping of stat keys to formatted strings.
count_diff_refname_stub (bool): Whether we are counting the contig name stub differences.
'''
('Histograms written to:', self.paths['long_hist'], self.paths['short_hist'], self.paths['log_log_hist'])
self.logger.info('Number of contigs (more is harder):')
self.logger.info(self.out_stats['contigs'])
self.logger.info('Number of contigs greater than 10KB (longer contigs are better):')
self.logger.info(self.out_stats['contigs_greater_10k'])
self.logger.info('N50 of input assembly (longer contigs are better):')
self.logger.info(self.out_stats['N50'])
self.logger.info('Length of input assembly (bigger is harder):')
self.logger.info(self.out_stats['total_length'])
self.logger.info('Counts of zero distances (many is a sign of bad prep):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['zero_dist_pairs'],
self.out_stats['total_read_pairs'],
self.out_stats['perc_zero_dist_pairs'])
)
self.logger.info('Count of same-contig read pairs with distance > 10KB (many is a sign of good prep):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['pairs_greater_10k'],
self.out_stats['total_read_pairs'],
self.out_stats['perc_pairs_greater_10k'])
)
self.logger.info('Proportion of reads mapping to contigs > 10 Kbp with inserts > 10 Kbp:')
self.logger.info('{} of total {} {}'.format(
self.out_stats['pairs_greater_10k_on_contigs_greater_10k'],
self.out_stats['pairs_on_contigs_greater_10k'],
self.out_stats['perc_pairs_greater_10k_on_contigs_greater_10k'])
)
self.logger.info('Count of read pairs with mates mapping to different chromosomes/contigs ' \
'(sign of good prep IF same genome):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['intercontig_pairs'],
self.out_stats['total_read_pairs'],
self.out_stats['perc_intercontig_pairs'])
)
self.logger.info('Count of split reads (more is usually good, as indicates presence of Hi-C junction in read):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['split_reads'],
self.stats['total_reads'],
self.out_stats['perc_split_reads'])
)
self.logger.info('Count of MAPQ zero reads (bad, ambiguously mapped):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['mapq0_reads'],
self.out_stats['total_reads'],
self.out_stats['perc_mapq0_reads'])
)
self.logger.info('Count of duplicate reads (-1 if insufficient to estimate; duplicates are bad; ' \
'WILL ALWAYS BE ZERO UNLESS BAM FILE IS PREPROCESSED TO SET THE DUPLICATES FLAG):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['duplicate_reads'],
self.out_stats['total_reads'],
self.out_stats['perc_duplicate_reads'])
)
if count_diff_refname_stub:
self.logger.info('Count of read pairs with mates mapping to different reference groupings, ' \
'e.g. genomes (sign of bad prep potentially):')
self.logger.info('{} of total {} {}'.format(
self.out_stats['different_ref_stub_pairs'],
self.out_stats['total_read_pairs'],
self.out_stats['perc_different_ref_stub_pairs'])
)