forked from AstraZeneca-NGS/VarDict
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathvardict.pl
executable file
·2768 lines (2692 loc) · 113 KB
/
vardict.pl
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 perl
# Parse a list of refseq and check CDS coverage
use warnings;
use Getopt::Std;
use strict;
our ($opt_h, $opt_H, $opt_b, $opt_D, $opt_d, $opt_s, $opt_c, $opt_S, $opt_E, $opt_n, $opt_N, $opt_e, $opt_g, $opt_x, $opt_f, $opt_r, $opt_B, $opt_z, $opt_v, $opt_p, $opt_F, $opt_C, $opt_m, $opt_Q, $opt_T, $opt_q, $opt_Z, $opt_X, $opt_P, $opt_3, $opt_k, $opt_R, $opt_G, $opt_a, $opt_o, $opt_O, $opt_V, $opt_t, $opt_y, $opt_I);
unless( getopts( 'hHvtzypDC3F:d:b:s:e:S:E:n:c:g:x:f:r:B:N:Q:m:T:q:Z:X:P:k:R:G:a:o:O:V:I:' )) {
USAGE();
}
USAGE() if ( $opt_H );
USAGE() unless ( $opt_b );
my $BAM = $opt_b; # the bam file
my $sample = $1 if ( $BAM =~ /([^\/\._]+).sorted[^\/]*.bam/ );
my $samplem = "";
if ( $opt_n ) {
$sample = $1 if ( $BAM =~ /$opt_n/ );
}
$sample = $opt_N if ( $opt_N );
unless($sample) { $sample = $1 if ( $BAM =~ /([^\/]+)[_\.][^\/]*bam/ ); }
$opt_d = "\t" unless( $opt_d );
my $c_col = $opt_c ? $opt_c - 1 : 2;
my $S_col = $opt_S ? $opt_S - 1 : 6;
my $E_col = $opt_E ? $opt_E - 1 : 7;
my $s_col = $opt_s ? $opt_s - 1 : 9;
my $e_col = $opt_e ? $opt_e - 1 : 10;
my $g_col = $opt_g ? $opt_g - 1 : 12;
$opt_m = $opt_m ? $opt_m : 8;
$s_col = $S_col if ( $opt_S && (!$opt_s) );
$e_col = $E_col if ( $opt_E && (!$opt_e) );
$opt_F = defined($opt_F) ? $opt_F : "0x500";
my $VEXT = defined($opt_X) ? $opt_X : 3; # the extension of deletion and insertion for complex variants
$opt_P = defined($opt_P) ? $opt_P : 5;
$opt_k = $opt_k ? $opt_k : 1; # Whether to perform local realignment. Set to 0 to disable.
my $fasta = $opt_G ? $opt_G : "/ngs/reference_data/genomes/Hsapiens/hg19/seq/hg19.fa";
my %CHRS; # Key: chr Value: chr_len
my $EXT = defined($opt_x) ? $opt_x : 0;
my $FREQ = $opt_f ? $opt_f : 0.05;
my $QRATIO = $opt_o ? $opt_o : 1.5; # The Qratio
my $BIAS = 0.05; # The cutoff to decide whether a positin has read strand bias
my $MINB = $opt_B ? $opt_B : 2; # The minimum reads for bias calculation
my $MINR = $opt_r ? $opt_r : 2; # The minimum reads for variance allele
my $GOODQ = defined($opt_q) ? $opt_q : 25; # The phred score in fastq to be considered as good base call
$opt_O = defined($opt_O) ? $opt_O : 0; # The minimun mean mapping quality to be considered
$opt_V = defined($opt_V) ? $opt_V : 0.05; # The minimun alelle frequency allowed in normal for a somatic mutation
my $RLEN = 0; # The read length
my $BUFFER = 200;
my $LOWQUAL = 10;
my $INDELSIZE = $opt_I ? $opt_I : 120;
my %SPLICE;
if ( $opt_p ) {
$FREQ = -1;
$MINR = 0;
}
if ( $opt_h ) {
print join("\t", qw(Sample Gene Chr Start End Ref Alt Depth AltDepth RefFwdReads RefRevReads AltFwdReads AltRevReads Genotype AF Bias PMean PStd QMean QStd 5pFlankSeq 3pFlankSeq)), "\n";
}
my ($BAM1, $BAM2) = split(/\|/, $BAM);
my @BAMS = split(/:/, $BAM1);
open(BAMH, "samtools view -H $BAMS[0] |");
while(<BAMH>) {
if ( /^\@SQ/ ) {
my $chr = $1 if ( /\s+SN:(\S+)/ );
$CHRS{ $chr } = $1 if( /\sLN:(\d+)/ );
}
}
close(BAMH);
$fasta = $CHRS{ 1 } ? "/ngs/reference_data/genomes/Hsapiens/GRCh37/seq/GRCh37.fa" : "/ngs/reference_data/genomes/Hsapiens/hg19/seq/hg19.fa" unless( $opt_G );
my @SEGS = ();
if ( $opt_R ) {
my ($chr, $reg, $gene) = split(/:/, $opt_R);
$chr =~ /^chr/ ? ( $chr =~ s/^chr// ) : ($chr = "chr$chr") unless( $CHRS{ $chr } );
$gene = $chr unless( $gene );
my ($start, $end) = split(/-/, $reg);
$end = $start unless( $end );
$start =~ s/,//g;
$end =~ s/,//g;
$start -= $EXT;
$end += $EXT;
$start++ if ( $opt_z && $start < $end );
$start = $end if ( $start > $end );
push(@SEGS, [[$chr, $start, $end, $gene]]);
} else {
my @segraw = ();
while( <> ) {
chomp;
next if ( /^#/ || /^browser/ || /^track/ );
unless($opt_a) {
my @a = split(/$opt_d/);
if ((! $opt_a) && @a == 8 && $a[6] =~ /^\d+$/ && $a[7] =~ /^\d+$/ && $a[6] >= $a[1] && $a[7] <= $a[2]) {
$opt_a = "10:0.95";
$opt_z = 1 unless( defined($opt_z) );
}
}
push(@segraw, $_);
}
if ($opt_a) {
my ($pchr, $pend) = (0, 0);
my $SI = 0;
my %tsegs = ();
foreach(@segraw) {
my ($chr, $start, $end, $gene, $score, $strand, $istart, $iend) = split(/$opt_d/);
$chr =~ /^chr/ ? ( $chr =~ s/^chr// ) : ($chr = "chr$chr") unless( $CHRS{ $chr } );
$start++ && $istart++ if ( $opt_z && $start < $end );
push( @{ $tsegs{ $chr } }, [$chr, $start, $end, $gene, $istart, $iend] );
}
while(my ($chr, $sv) = each %tsegs ) {
my @tmp = sort { $a->[4] <=> $b->[4]; } @$sv;
foreach my $tv (@tmp) {
my ($chr, $start, $end, $gene, $istart, $iend) = @$tv;
$SI++ if ( $pend && ($chr ne $pchr || $istart > $pend) );
push(@{$SEGS[$SI]}, [$chr, $start, $end, $gene, $istart, $iend]);
$pchr = $chr;
$pend = $iend;
}
}
ampVardict();
exit(0);
} else {
foreach(@segraw) {
my @A = split(/$opt_d/);
if ( (! $opt_c) && @A == 4 && $A[1] =~ /^\d+$/ && $A[2] =~ /^\d+$/ && $A[1] <= $A[2] ) {
($c_col, $S_col, $E_col, $g_col, $s_col, $e_col) = (0, 1, 2, 3, 1, 2);
$opt_z = 1 unless( defined($opt_z) ); # only set if it's a 4 column BED file
}
my ($chr, $cdss, $cdse, $gene) = @A[$c_col, $S_col, $E_col, $g_col];
$chr =~ /^chr/ ? ( $chr =~ s/^chr// ) : ($chr = "chr$chr") unless( $CHRS{ $chr } );
my @starts = split(/,/, $A[$s_col]);
my @ends = split(/,/, $A[$e_col]);
my @CDS = ();
$gene = $chr unless( $gene );
for(my $i = 0; $i < @starts; $i++) {
my ($s, $e) = ($starts[$i], $ends[$i]);
next if ( $cdss > $e ); # not a coding exon
last if ( $cdse < $s ); # No more coding exon
$s = $cdss if ( $s < $cdss );
$e = $cdse if ( $e > $cdse );
$s -= $EXT; # unless ( $s == $cdss );
$e += $EXT; # unless ( $e == $cdse );
$s++ if ( $opt_z && $s < $e );
push(@CDS, [$chr, $s, $e, $gene]);
}
push(@SEGS, \@CDS);
}
}
}
for(my $i = 0; $i < @SEGS; $i++) {
for(my $j = 0; $j < @{ $SEGS[$i] }; $j++) {
my $REF = getREF(@{ $SEGS[$i]->[$j] }[0..2]);
if ( $BAM2 ) { # pair mode
if ( $opt_n ) {
$sample = $1 if ( $BAM1 =~ /$opt_n/ );
$samplem = $1 if ( $BAM2 =~ /$opt_n/ );
}
($sample, $samplem) = split(/\|/, $opt_N) if ( $opt_N );
$samplem = $sample . "_match" unless( $samplem );
somdict($SEGS[$i]->[$j], toVars(@{ $SEGS[$i]->[$j] }[0..2], $BAM1, $REF), toVars(@{ $SEGS[$i]->[$j] }[0..2], $BAM2, $REF));
} else {
vardict($SEGS[$i]->[$j], toVars(@{ $SEGS[$i]->[$j] }[0..2], $BAM1, $REF));
}
}
}
sub ampVardict {
for(my $i = 0; $i < @SEGS; $i++) {
my @vars = ();
my %pos = ();
my ($chr, $start, $end, $gene, $istart, $iend);
for(my $j = 0; $j < @{$SEGS[$i]}; $j++) {
($chr, $start, $end, $gene, $istart, $iend) = @{$SEGS[$i]->[$j]};
my $REF = getREF($chr, $start, $end);
push(@vars, toVars($chr, $start, $end, $BAM1, $REF));
for(my $p = $istart; $p <= $iend; $p++) {
push(@{ $pos{ $p } }, [$j, $chr, $start, $end, $istart, $iend]);
}
}
while(my ($p, $v) = each %pos) {
my @gvs = (); # Good variants
my @ref = (); # reference
my $nt;
my $maxaf = 0;
my $vartype = "SNV";
my $flag = 0;
my $vref;
my $nocov=0; #
my $maxcov=0;
my %goodamp;
my @vcovs = ();
foreach my $amps (@$v) {
my ($amp, $chr, $S, $E) = @$amps;
if ( $vars[$amp]->{ $p }->{ VAR }->[0] ) {
my $tv = $vars[$amp]->{ $p }->{ VAR }->[0];
push(@vcovs, $tv->{ tcov });
$vartype = varType($tv->{ refallele }, $tv->{ varallele });
if ( isGoodVar($tv, $vars[$amp]->{ $p }->{ REF }, $vartype) ) {
push(@gvs, [$tv, "$chr:$S-$E"]);
if ( $nt && $tv->{ n } ne $nt ) {
$flag = 1;
}
if ($tv->{ freq } > $maxaf ) {
($maxaf, $nt, $vref) = ($tv->{ freq }, $tv->{ n }, $tv);
}
$goodamp{ "$amp-$tv->{ refallele }-$tv->{ varallele }" } = 1;
$maxcov = $tv->{ tcov } if ( $tv->{ tcov } > $maxcov );
}
} elsif ( $vars[$amp]->{ $p }->{ REF } ) {
push(@vcovs, $vars[$amp]->{ $p }->{ REF }->{ tcov });
} else {
push(@vcovs, 0);
}
push(@ref, $vars[$amp]->{ $p }->{ REF }) if ( $vars[$amp]->{ $p }->{ REF } );
}
foreach my $t (@vcovs) {
$nocov++ if ( $t < $maxcov/50 ); # The amplicon that has depth less than 1/50 of the max depth will be considered not working and thus not used.
}
@gvs = sort { $b->[0]->{ freq } <=> $a->[0]->{ freq } } @gvs if ( @gvs > 1 );
@ref = sort { $b->{ tcov } <=> $a->{ tcov } } @ref if ( @ref > 1 );
if ( @gvs < 1 ) { # Only referenece
if ( $opt_p ) {
if ( @ref ) {
$vref = $ref[0];
} else {
print join("\t", $sample, $gene, $chr, $p, $p, "", "", 0, 0, 0, 0, 0, 0, "", 0, "0;0", 0, 0, 0, 0, 0, "", 0, 0, 0, 0, 0, 0, "", "", 0, 0, "$chr:$p-$p", "", 0, 0, 0, 0), "\n";
next;
}
} else {
next;
}
} else {
$vref = $gvs[0]->[0];
}
if ( $flag ) { # different good variants detected in different amplicons
#next if ($gvs[0]->[0]->{ freq }/$gvs[1]->[0]->{ freq } < 10); # need 10 times difference to overwrite it.
my $gdnt = $gvs[0]->[0]->{ n };
my $gcnt = 0;
foreach my $amps (@$v) {
my ($amp, $chr, $S, $E) = @$amps;
$gcnt++ if ( $vars[$amp]->{ $p }->{ VARN }->{ $gdnt } && isGoodVar( $vars[$amp]->{ $p }->{ VARN }->{ $gdnt }, $vars[$amp]->{ $p }->{ REF }) );
}
$flag = 0 if ( $gcnt == @gvs+0 );
}
my @badv = ();
my $gvscnt = @gvs+0;
foreach my $amps (@$v) {
my ($amp, $chr, $S, $E, $iS, $iE) = @$amps;
next if ( $goodamp{"$amp-$vref->{ refallele }-$vref->{ varallele }"} );
my $tref = $vars[$amp]->{ $p }->{ VAR }->[0];
if ( $vref->{ sp } >= $iS && $vref->{ ep } <= $iE ) {
if ( $vars[$amp]->{ $p }->{ VAR } ) {
push( @badv, [$vars[$amp]->{ $p }->{ VAR }->[0], "$chr:$S-$E"] );
} elsif ( $vars[$amp]->{ $p }->{ REF } ) {
push( @badv, [$vars[$amp]->{ $p }->{ REF }, "$chr:$S-$E"] );
} else {
push( @badv, [undef, "$chr:$S-$E"] );
}
} elsif ( ($vref->{ sp } < $iE && $iE < $vref->{ ep }) || ($vref->{ sp } < $iS && $iS < $vref->{ ep }) ) { # the variant overlap with amplicon's primer
#print STDERR "$iS $iE $vref->{sp} $vref->{ep} $vref->{n} $gvscnt\n";
$gvscnt-- if ( $gvscnt > 1 );
}
}
$flag = 0 if ( $flag && $gvscnt < @gvs+0 );
my @hds = qw(sp ep refallele varallele tcov cov rfc rrc fwd rev genotype freq bias pmean pstd qual qstd mapq qratio hifreq extrafreq shift3 msi msint nm hicnt hicov leftseq rightseq);
my @hds2 = qw(tcov cov rfc rrc fwd rev genotype freq bias pmean pstd qual qstd mapq qratio hifreq extrafreq);
adjComplex($vref) if ( $vartype eq "Complex" );
print join("\t", $sample, $gene, $chr, (map { $vref->{ $_ } ? $vref->{ $_ } : 0; } @hds), $gvs[0]->[1] ? $gvs[0]->[1] : "", $vartype, $gvscnt, $gvscnt+@badv+0, $nocov, $flag);
print "\t", $vref->{ DEBUG } if ( $opt_D );
for(my $gvi = 1; $gvi < @gvs; $gvi++) {
print "\tGood$gvi ", join(" ", (map {$gvs[$gvi]->[0]->{ $_ }; } @hds2), $gvs[$gvi]->[1]) if ( $opt_y );
}
for(my $bvi = 0; $bvi < @badv; $bvi++) {
print "\tBad$bvi ", join(" ", (map {$badv[$bvi]->[0]->{ $_ } ? $badv[$bvi]->[0]->{ $_ } : 0; } @hds2), $badv[$bvi]->[1]) if ( $opt_y );
}
print "\n";
}
}
}
sub varType {
my ($ref, $var) = @_;
if ( length($ref) == 1 && length($var) == 1 ) {
return "SNV";
} elsif ( substr($ref, 0, 1) ne substr($var, 0, 1) ) {
return "Complex";
} elsif ( length($ref) == 1 && length($var) > 1 && $var =~ /^$ref/ ) {
return "Insertion";
} elsif (length($ref) > 1 && length($var) == 1 && $ref =~ /^$var/ ) {
return "Deletion";
}
return "Complex";
}
sub vardict {
my ($seg, $vars) = @_;
my ($chr, $S, $E, $G) = @{ $seg };
my @hds = qw(sp ep refallele varallele tcov cov rfc rrc fwd rev genotype freq bias pmean pstd qual qstd mapq qratio hifreq extrafreq shift3 msi msint nm hicnt hicov leftseq rightseq);
for(my $p = $S; $p <= $E; $p++) {
my @vts = ();
my @vrefs = ();
unless( $vars->{ $p }->{ VAR } ) {
next unless( $opt_p );
my $vref = $vars->{ $p }->{ REF };
unless($vref) {
print join("\t", $sample, $G, $chr, $p, $p, "", "", 0, 0, 0, 0, 0, 0, "", 0, "0;0", 0, 0, 0, 0, 0, "", 0, 0, 0, 0, 0, 0, "", "", 0, 0, "$chr:$S-$E", ""), "\n";
next;
}
push(@vts, "");
push(@vrefs, $vref);
} else {
for(my $i = 0; $i < @{ $vars->{$p}->{VAR} }; $i++) {
my $vref = $vars->{ $p }->{ VAR }->[$i];
last if ( $vref->{refallele} =~ /N/ );
my $vartype = varType($vref->{refallele}, $vref->{varallele});
unless( isGoodVar( $vars->{ $p }->{ VAR }->[$i], $vars->{ $p }->{ REF }, $vartype ) ) {
next unless( $opt_p );
}
push(@vts, $vartype);
push(@vrefs, $vref);
}
}
for(my $vi = 0; $vi < @vts; $vi++) {
my ($vartype, $vref) = ($vts[$vi], $vrefs[$vi]);
adjComplex($vref) if ( $vartype eq "Complex" );
print join("\t", $sample, $G, $chr, (map { $vref->{ $_ } ? $vref->{ $_ } : 0; } @hds), "$chr:$S-$E", $vartype);
print "\t", $vref->{ DEBUG } if ( $opt_D );
print "\n";
}
}
}
# Adjust the complex var
sub adjComplex {
my $vref = shift;
my ($refnt, $varnt) = ($vref->{ refallele }, $vref->{ varallele });
my $n = 0;
$n++ while( length($refnt) - $n > 1 && length($varnt) - $n > 1 && substr($refnt, $n, 1) eq substr($varnt, $n, 1) );
if ( $n ) {
$vref->{ sp } += $n;
$vref->{ refallele } = substr($refnt, $n);
$vref->{ varallele } = substr($varnt, $n);
$vref->{ leftseq } .= substr($refnt, 0, $n);
$vref->{ leftseq } = substr($vref->{ leftseq }, $n);
}
($refnt, $varnt) = ($vref->{ refallele }, $vref->{ varallele });
$n = 1;
$n++ while( length($refnt) - $n > 0 && length($varnt) - $n > 0 && substr($refnt, -$n, 1) eq substr($varnt, -$n, 1) );
if ( $n > 1 ) {
$vref->{ ep } -= $n - 1;
$vref->{ refallele } = substr($refnt, 0, -$n+1);
$vref->{ varallele } = substr($varnt, 0, -$n+1);
$vref->{ rightseq } = substr($refnt, -($n-1), $n-1) . substr($vref->{ rightseq }, 0, -$n+1);
}
}
sub somdict {
my ($seg, $vars1, $vars2) = @_;
my ($chr, $S, $E, $G) = @{ $seg };
my @hdrs = qw(tcov cov rfc rrc fwd rev genotype freq bias pmean pstd qual qstd mapq qratio hifreq extrafreq nm);
my @hd1 = ("sp", "ep", "refallele", "varallele");
my @hd2 = ("shift3", "msi", "msint", "leftseq", "rightseq");
my $FISHERP = 0.01;
for(my $p = $S; $p <= $E; $p++) {
next unless($vars1->{$p} || $vars2->{$p}); # both samples have no coverage
my $vartype = "";
if( ! $vars1->{ $p } ) { # no coverage for sample 1
$vartype = varType($vars2->{$p}->{ VAR }->[0]->{ refallele }, $vars2->{$p}->{ VAR }->[0]->{ varallele }) if ($vars2->{$p}->{ VAR });
next unless( $vars2->{$p}->{ VAR } && isGoodVar($vars2->{$p}->{ VAR }->[0], $vars2->{$p}->{REF}, $vartype) );
adjComplex($vars2->{$p}->{ VAR }->[0]) if ( $vartype eq "Complex" );
print join("\t", $sample, $G, $chr, (map{ $vars2->{$p}->{ VAR }->[0]->{ $_ }; } @hd1), (map{ 0; } @hdrs), (map { $vars2->{$p}->{ VAR }->[0]->{ $_ }; } (@hdrs, @hd2)), "$chr:$S-$E", "Deletion", $vartype), "\n";
} elsif( ! $vars2->{ $p } ) { # no coverage for sample 2
$vartype = varType($vars1->{$p}->{ VAR }->[0]->{ refallele }, $vars1->{$p}->{ VAR }->[0]->{ varallele }) if ($vars1->{$p}->{ VAR });
next unless( $vars1->{$p}->{ VAR } && isGoodVar($vars1->{$p}->{ VAR }->[0], $vars1->{$p}->{REF}, $vartype) );
adjComplex($vars1->{$p}->{ VAR }->[0]) if ( $vartype eq "Complex" );
print join("\t", $sample, $G, $chr, (map { $vars1->{$p}->{ VAR }->[0]->{ $_ }; } (@hd1, @hdrs)), (map { 0; } @hdrs), (map { $vars1->{$p}->{ VAR }->[0]->{$_};} @hd2), "$chr:$S-$E", "SampleSpecific", $vartype), "\n";
} else { # both samples have coverage
my ($v1, $v2) = ($vars1->{$p}, $vars2->{$p});
next unless ( $v1->{ VAR } || $v2->{ VAR } );
if ( $v1->{ VAR } ) {
my $N = 0;
while( $v1->{ VAR }->[$N] && isGoodVar($v1->{ VAR }->[$N], $v1->{ REF }, varType($v1->{ VAR }->[$N]->{ refallele }, $v1->{ VAR }->[$N]->{ varallele }))) {
my $VREF = $v1->{ VAR }->[$N];
my $nt = $VREF->{ n };
if ( length($nt) > 1 && length($VREF->{ refallele }) == length($VREF->{ varallele }) && (! isGoodVar($v2->{ VARN }->{ $nt })) && $VREF->{ genotype } !~ /-/ && $VREF->{ genotype } !~ /m/ && $VREF->{ genotype } !~ /i/ ) {
my $fnt = substr($nt, 0, -1); $fnt =~ s/&$//;
my $lnt = substr($nt, 1); $lnt =~ s/^&//; substr($lnt, 1, 0) = "&" if ( length($lnt) > 1 );
if ( $v2->{ VARN }->{ $fnt } && isGoodVar($v2->{ VARN }->{ $fnt }, $v2->{ REF })) {
$VREF->{ sp } += length($VREF->{ refallele }) - 1;
$VREF->{ refallele } = substr($VREF->{ refallele }, -1 );
$VREF->{ varallele } = substr($VREF->{ varallele }, -1 );
} elsif ( $vars2->{ $p + length($nt) -2 }->{ VARN }->{ $lnt } && isGoodVar($vars2->{ $p + length($nt) -2 }->{ VARN }->{ $lnt }, $vars2->{$p+length($nt)-2}->{REF})) {
$VREF->{ ep } += length($VREF->{ refallele }) - 1;
$VREF->{ refallele } = substr($VREF->{ refallele }, 0, -1 );
$VREF->{ varallele } = substr($VREF->{ varallele }, 0, -1 );
}
}
$vartype = varType($VREF->{ refallele }, $VREF->{ varallele });
adjComplex($VREF) if ( $vartype eq "Complex" );
if ( $v2->{ VARN }->{ $nt } ) {
#my $type = isGoodVar( $v2->{ VARN }->{ $nt }, $v2->{ REF } ) ? "Germline" : ($v2->{ VARN }->{ $nt }->{ freq } < $opt_V || $v2->{ VARN }->{ $nt }->{ cov } <= 1 ? "LikelySomatic" : "AFDiff");
my $type = isGoodVar( $v2->{ VARN }->{ $nt }, $v2->{ REF }, $vartype ) ? ($VREF->{ freq } > (1-$opt_V) && $v2->{ VARN }->{ $nt }->{ freq } < 0.8 && $v2->{ VARN }->{ $nt }->{ freq } > 0.2 ? "LikelyLOH" : "Germline") : ($v2->{ VARN }->{ $nt }->{ freq } < $opt_V || $v2->{ VARN }->{ $nt }->{ cov } <= 1 ? "LikelySomatic" : "AFDiff");
$type = "StrongSomatic" if ( isNoise($v2->{ VARN }->{ $nt }) && $vartype eq "SNV" );
#if ($type =~ /Somatic/) {
# my $newtype = combineAnalysis($VREF, $v2->{ VARN }->{ $nt }, $chr, $p, $nt);
# if ( $newtype eq "FALSE" ) {$N++; next;}
# $type = $newtype if ( $newtype );
#}
#$type = "StrongSomatic" if ( isNoise($v2->{ VARN }->{ $nt }) );
print join("\t", $sample, $G, $chr, (map { $VREF->{ $_ }; } (@hd1, @hdrs)), (map { $v2->{ VARN }->{ $nt }->{ $_ }; } (@hdrs, @hd2)), "$chr:$S-$E", $type, $vartype), "\n";
} else { # sample 1 only, should be strong somatic
my @tvf = $v2->{ REF } ? (map { $v2->{ REF }->{ $_ } ? $v2->{ REF }->{ $_ } : 0; } @hdrs) : ($v2->{ VAR }->[0]->{ tcov } ? $v2->{ VAR }->[0]->{ tcov } : 0, map { 0; } (1..17));
my $type = "StrongSomatic";
#if ($vartype ne "SNV") {
$v2->{ VARN }->{ $nt }->{ cov } = 0; # Ensure it's initialized before passing to combineAnalysis
my $newtype = combineAnalysis($VREF, $v2->{ VARN }->{ $nt }, $chr, $p, $nt);
if ( $newtype eq "FALSE" ) {$N++; next;}
$type = $newtype if ( $newtype );
#}
if ( $type ne "StrongSomatic" ) {
print join("\t", $sample, $G, $chr, (map { $VREF->{ $_ }; } (@hd1, @hdrs)), (map { $v2->{ VARN }->{ $nt }->{ $_ }; } @hdrs),(map { $VREF->{$_}; } @hd2), "$chr:$S-$E", $type, $vartype), "\n";
} else {
print join("\t", $sample, $G, $chr, (map { $VREF->{ $_ }; } (@hd1, @hdrs)), @tvf, (map { $VREF->{$_};} @hd2), "$chr:$S-$E", "StrongSomatic", $vartype), "\n";
}
}
$N++;
}
unless($N) {
next unless( $v2->{ VAR } );
$vartype = varType($v2->{ VAR }->[0]->{ refallele }, $v2->{ VAR }->[0]->{ varallele });
next unless( isGoodVar( $v2->{ VAR }->[0], $v2->{ REF }, $vartype ) );
# potentail LOH
my $nt = $v2->{ VAR }->[0]->{ n };
if ( $v1->{ VARN }->{ $nt } ) {
my $type = $v1->{ VARN }->{ $nt }->{ freq } < $opt_V ? "LikelyLOH" : "Germline";
adjComplex($v1->{ VARN }->{ $nt }) if ( $vartype eq "Complex" );
print join("\t", $sample, $G, $chr, (map { $v1->{ VARN }->{ $nt }->{ $_ }; } (@hd1, @hdrs)), (map { $v2->{ VAR }->[0]->{ $_ }; } (@hdrs, @hd2)), "$chr:$S-$E", $type, varType($v1->{ VARN }->{ $nt }->{ refallele }, $v1->{ VARN }->{ $nt }->{ varallele })), "\n";
} else {
my @th1 = $v1->{ REF } ? (map { $v1->{ REF }->{ $_ } } @hdrs) : ($v1->{ VAR }->[0]->{ tcov }, (map { 0; } @hdrs[1..$#hdrs]));
adjComplex($v2->{ VAR }->[0]) if ( $vartype eq "Complex" );
print join("\t", $sample, $G, $chr, (map { $v2->{ VAR }->[0]->{ $_ }; } @hd1), @th1, (map { $v2->{ VAR }->[0]->{ $_ }; } (@hdrs, @hd2)), "$chr:$S-$E", "StrongLOH", $vartype), "\n";
}
}
} elsif ( $v2->{ VAR } ) { # sample 1 has only reference
$vartype = varType($v2->{ VAR }->[0]->{ refallele }, $v2->{ VAR }->[0]->{ varallele });
next unless( isGoodVar( $v2->{ VAR }->[0], $v2->{ REF }, $vartype ) );
# potential LOH
my $nt = $v2->{ VAR }->[0]->{ n };
my $type = "StrongLOH";
$v1->{ VARN }->{ $nt }->{ cov } = 0;
my $newtype = combineAnalysis($v2->{ VARN }->{ $nt }, $v1->{ VARN }->{ $nt }, $chr, $p, $nt);
next if ( $newtype eq "FALSE" );
$type = $newtype if ( $newtype );
my @th1 = $newtype ? (map { $v1->{ VARN }->{ $nt }->{ $_ } ? $v1->{ VARN }->{ $nt }->{ $_ } : 0; } @hdrs) : (map { $v1->{ REF }->{ $_ } ? $v1->{ REF }->{ $_ } : 0 } @hdrs);
adjComplex($v2->{ VAR }->[0]) if ( $vartype eq "Complex" );
print join("\t", $sample, $G, $chr, (map { $v2->{ VAR }->[0]->{ $_ } ? $v2->{ VAR }->[0]->{ $_ } : 0; } @hd1), @th1, (map { $v2->{ VAR }->[0]->{ $_ } ? $v2->{ VAR }->[0]->{ $_ } : 0; } (@hdrs, @hd2)), "$chr:$S-$E", $type, $vartype), "\n";
}
}
}
}
# Taken a likely somatic indels and see whether combine two bam files still support somatic status. This is mainly
# for Indels that softclipping overhang is too short to positively being called in one bam file, but can be called
# in the other bam file, thus creating false positives
sub combineAnalysis {
my ($var1, $var2, $chr, $p, $nt) = @_;
my ($bam1, $bam2) = split(/\|/, $BAM);
print STDERR "Start Combine $p $nt\n" if ( $opt_y );
my $REF = getREF($chr, $var1->{ sp } - $RLEN, $var1->{ ep } + $RLEN);
my $vars = toVars($chr, $var1->{ sp } - $RLEN, $var1->{ ep } + $RLEN, "$bam1:$bam2", $REF);
if ( $vars->{ $p }->{ VARN }->{ $nt } ) {
my $vref = $vars->{ $p }->{ VARN }->{ $nt };
print STDERR "Combine: 1: $var1->{ cov } comb: $vref->{ cov }\n" if ( $opt_y );
if ( $vref->{ cov } - $var1->{ cov } >= $MINR ) {
map { $var2->{ $_ } = $vref->{ $_ } - $var1->{ $_ } > 0 ? $vref->{ $_ } - $var1->{ $_ } : 0; } qw(tcov cov rfc rrc fwd rev);
map { $var2->{ $_ } = sprintf("%.3f", ($vref->{ $_ }*$vref->{ cov } - $var1->{ $_ }*$var1->{ cov })/$var2->{ cov }); } qw(pmean qual mapq hifreq extrafreq nm);
map { $var2->{ $_ } = 1; } qw(pstd qstd);
return "FALSE" if ( $var2->{ tcov } <= 0 );
my $freq2 = sprintf("%.3f", $var2->{ cov }/$var2->{ tcov });
$var2->{ freq } = $freq2;
$var2->{ qratio } = $var1->{ qratio }; # Can't back calculate and should be inaccurate
$var2->{ genotype } = $vref->{ genotype };
$var2->{ bias } = strandBias($var2->{rfc}, $var2->{rrc}) . ";" . strandBias($var2->{fwd}, $var2->{rev});
return "Germline";
} elsif ($vref->{ cov } < $var1->{ cov } - 2) {
print STDERR "Combine produce less: $chr $p $nt $vref->{ cov } $var1->{ cov }\n" if ( $opt_y );
return "FALSE";
} else {
return "";
}
} else {
return "FALSE";
}
return "";
}
# A variance is considered noise if the quality is below $GOODQ and there're no more than 3 reads
sub isNoise {
my $vref = shift;
if ( (($vref->{ qual } < 4.5 || ($vref->{ qual } < 12 && $vref->{ qstd } == 0)) && $vref->{ cov } <= 3) || ($vref->{ qual } < $GOODQ && $vref->{ freq } < 2*$opt_V && $vref->{ cov } <= 1) ) {
$vref->{ tcov } -= $vref->{ cov };
$vref->{ cov } = 0;
$vref->{ fwd } = 0;
$vref->{ rev } = 0;
$vref->{ freq } = 0;
$vref->{ hifreq } = 0;
return 1;
}
return 0;
}
# Determine whether a variant meet specified criteria
sub isGoodVar {
my ($vref, $rref, $type) = @_;
return 0 unless($vref && $vref->{ refallele });
$type = varType($vref->{ refallele }, $vref->{ varallele }) unless( $type );
return 0 if ( $vref->{ freq } < $FREQ );
return 0 if ( $vref->{ hicnt } < $MINR );
return 0 if ( $vref->{ pmean } < $opt_P );
return 0 if ( $vref->{ qual } < $GOODQ );
#print STDERR "$vref->{sp} $vref->{ cov } $vref->{varallele} $type\n";
#return 0 if ( $vref->{ pstd } == 0 ); # Leave it to var2vcf, especially for targeted PCR, where pstd = 0 might be expected.
if ( $rref && $rref->{ hicnt } > $MINR && $vref->{ freq } < 0.25 ) {
#The reference allele has much better mean mapq than var allele, thus likely false positives
return 0 if ( ($vref->{ mapq } + (length($vref->{ refallele }) + length($vref->{ varallele })-2) < 5 && $rref->{ mapq } > 20) || ( 1 + ($vref->{ mapq } + (length($vref->{ refallele }) + length($vref->{ varallele }))))/($rref->{ mapq } + 1) < 0.25 );
#return 0 if ( $type && $type eq "SNV" && (($vref->{ mapq } < 5 && $rref->{ mapq } > 20) || $rref->{ mapq } - $vref->{ mapq } > 30 ));
}
return 0 if ( $type eq "Deletion" && $SPLICE{ $vref->{sp} . "-" . $vref->{ep} } );
return 0 if ( $vref->{ qratio } < $QRATIO );
return 1 if ( $vref->{ freq } > 0.35 );
return 0 if ( $vref->{ mapq } < $opt_O );
return 0 if ( $vref->{ msi } >= 13 && $vref->{ freq } <= 0.275 && $vref->{ msint } == 1);
return 0 if ( $vref->{ msi } >= 8 && $vref->{ freq } <= 0.2 && $vref->{ msint } > 1 );
if ( $vref->{ bias } eq "2;1" && $vref->{ freq } < 0.25 ) {
return 0 unless($type && $type ne "SNV" && (length($vref->{refallele}) > 3 || length($vref->{varallele}) > 3));
}
return 1;
}
sub addCnt {
my ($vref, $dir, $rp, $q, $Q, $nm) = @_; # ref dir read_position quality
$vref->{ cnt }++;
$vref->{ $dir }++;
$vref->{ pmean } += $rp;
$vref->{ qmean } += $q;
$vref->{ Qmean } += $Q;
$vref->{ nm } += $nm;
$q >= $GOODQ ? $vref->{ hicnt }++ : $vref->{ locnt }++;
}
# Given a region, return the REF
sub getREF {
my ($chr, $START, $END) = @_;
my %REF;
my $s_start = $START - $EXT - 700 < 1 ? 1 : $START - $EXT - 700;
my $s_end = $END + $EXT + 700 > $CHRS{ $chr } ? $CHRS{ $chr } : $END + $EXT + 700;
my ($header, $exon) = split(/\n/, `samtools faidx $fasta $chr:$s_start-$s_end`, 2);
$exon =~ s/\s+//g;
for(my $i = $s_start; $i <= $s_start + length($exon); $i++) {
$REF{ $i } = uc(substr( $exon, $i - ($s_start), 1 ));
}
return \%REF;
}
# Add deletion structural variant
sub addSVDel {
my ($sdref, $s, $e, $ms, $me, $dir, $rlen, $mlen) = @_;
$sdref->{ cnt }++;
$sdref->{ dir } = $dir;
$sdref->{ start } = $s unless( $sdref->{ start } && $sdref->{ start } < $s );
$sdref->{ end } = $e unless( $sdref->{ end } && $sdref->{ end } > $e );
push(@{ $sdref->{ mates } }, [$ms, $me, $mlen]);
$sdref->{ mstart } = $ms unless( $sdref->{ mstart } && $sdref->{ mstart } < $ms );
$sdref->{ mend } = $ms+$rlen unless( $sdref->{ mend } && $sdref->{ mend } > $me );
}
# Construct a variant structure given a segment and BAM files.
sub parseSAM {
my ($chr, $START, $END, $bam, $REF) = @_;
my %cov;
my %var;
my %hash;
my %dels5;
my %ins;
my %mnp; # Keep track of MNPs
my %sclip3; # soft clipped at 3'
my %sclip5; # soft clipped at 5'
my @bams = split(/:/, $bam);
$chr =~ s/^chr// if ( $opt_C );
$REF = getREF($chr, $START, $END);
foreach my $bami (@bams) {
#my $tsamcnt = `samtools view $bami $chr:$START-$END | head -1`;
#next unless( $tsamcnt || $opt_p ); # to avoid too much IO for exome and targeted while the BED is whole genome
# Get the reference sequence
my $SAMFILTER = $opt_F ? "-F $opt_F" : "";
open(SAM, "samtools view $SAMFILTER $bami $chr:$START-$END |");
my %DUP = ();
my $DUPP = 0;
while( <SAM> ) {
if ( $opt_Z ) {
next if ( rand() <= $opt_Z );
}
my @a = split(/\t/);
next if ( defined($opt_Q) && $a[4] < $opt_Q ); # ignore low mapping quality reads
#next if ( $opt_F && ($a[1] & 0x200 || $a[1] & 0x100) ); # ignore "non-primary alignment", or quality failed reads.
#next if ( $opt_F && ($a[1] & 0x100) ); # ignore "non-primary alignment"
if ( $a[1] & 0x100 ) {
next if ($opt_F); # -F 0 to turn it off
}
next if ( $a[9] eq "*" );
# filter duplicated reads if option -t is set
if ( $opt_t ) {
%DUP = () if ( $a[3] ne $DUPP );
if ( $a[7] =~ /\d/ ) {
next if ( $DUP{ "$a[3]-$a[6]-$a[7]" } );
$DUP{ "$a[3]-$a[6]-$a[7]" } = 1;
$DUPP = $a[3];
} elsif ( $a[1] & 0x8 ) { # mate not mapped
next if ( $DUP{ "$a[3]-$a[5]" } );
$DUP{ "$a[3]-$a[5]" } = 1;
$DUPP = $a[3];
}
}
my $nm = 0;
my @segid = $a[5] =~ /(\d+)[ID]/g; # For total indels
my $idlen = 0; $idlen += $_ foreach(@segid);
if ( /NM:i:(\d+)/i ) { # number of mismatches. Don't use NM since it includes gaps, which can be from indels
$nm = $1 - $idlen;
next if ( $opt_m && $nm > $opt_m ); # edit distance - indels is the # of mismatches
} else {
print STDERR "No XM tag for mismatches. $_\n" if ( $opt_y && $a[5] ne "*" );
next;
}
my $n = 0; # keep track the read position, including softclipped
my $p = 0; # keep track the position in the alignment, excluding softclipped
my $dir = $a[1] & 0x10 ? "-" : "+";
# Amplicon based calling
if ( $opt_a ) {
my ($dis, $ovlp) = split( /:/, $opt_a );
($dis, $ovlp) = (10, 0.95) unless($dis && $ovlp);
my $rlen3= 0; $rlen3 += $1 while( $a[5] =~ /(\d+)[MD]/g ); # The total aligned length, excluding soft-clipped bases and insertions
my ($segstart, $segend) = ($a[3], $a[3]+$rlen3-1);
if ( $a[5] =~ /^(\d+)S/ ) {
my $ts1 = $segstart > $START ? $segstart : $START;
my $te1 = $segend < $END ? $segend : $END;
next unless( abs($ts1-$te1)/($segend-$segstart) > $ovlp);
} elsif ($a[5] =~ /(\d+)S$/ ) {
my $ts1 = $segstart > $START ? $segstart : $START;
my $te1 = $segend < $END ? $segend : $END;
next unless( abs($te1-$ts1)/($segend-$segstart) > $ovlp);
} else {
if ($a[6] eq "=" && $a[8]) {
($segstart, $segend) = $a[8] > 0 ? ($segstart, $segstart+$a[8]-1) : ($a[7], $a[7]-$a[8]-1);
}
# No segment overlapping test since samtools should take care of it
my $ts1 = $segstart > $START ? $segstart : $START;
my $te1 = $segend < $END ? $segend : $END;
next unless( (abs($segstart - $START) <= $dis && abs($segend - $END) <= $dis ) && abs(($ts1-$te1)/($segend-$segstart)) > $ovlp);
}
}
if ( $a[1] & 0x8 ) { # Mate unmapped, potential insertion
# to be implemented
} else {
if ( $a[6] eq "=" ) {
if ( /\tSA:Z:(\S+)/ ) {
if ( $a[1] & 0x800 ) { # the supplementary alignment
next; # Ignore the supplmentary for now so that it won't skew the coverage
}
}
}
}
# Modify the CIGAR for potential mis-alignment for indels at the end of reads to softclipping and let VarDict's algorithm to figure out indels
my $offset = 0;
while( $idlen > 0 ) {
my $flag = 0;
if ($a[5] =~ /^(\d+)S(\d+)([ID])/) {
my $tslen = $1 + ($3 eq "I" ? $2 : 0);
$tslen .= "S";
$a[3] += $3 eq "D" ? $2 : 0;
$a[5] =~ s/^\d+S\d+[ID]/$tslen/;
$flag = 1;
}
if ($a[5] =~ /(\d+)([ID])(\d+)S$/) {
my $tslen = $3 + ($2 eq "I" ? $1 : 0);
$tslen .= "S";
$a[5] =~ s/\d+[ID]\d+S$/$tslen/;
$flag = 1;
}
if ($a[5] =~ /^(\d+)S(\d+)M(\d+)([ID])/) {
my $tslen = $1 + $2 + ($4 eq "I" ? $3 : 0);
my $tmid = $2;
$tslen .= "S";
if ( $tmid <= 10 ) {
$a[3] += $2 + ($4 eq "D" ? $3 : 0);
$a[5] =~ s/^\d+S\d+M\d+[ID]/$tslen/;
$flag = 1;
}
}
if ($a[5] =~ /(\d+)([ID])(\d+)M(\d+)S$/) {
my $tslen = $4 + $3 + ($2 eq "I" ? $1 : 0);
my $tmid = $3;
$tslen .= "S";
if ( $tmid <= 10 ) {
$a[5] =~ s/\d+[ID]\d+M\d+S$/$tslen/;
$flag = 1;
}
}
# The following two clauses to make indels at the end of reads as softly
# clipped reads and let VarDict's algorithm identify indels
if ($a[5] =~ /^(\d)M(\d+)([ID])(\d+)M/) {
my $tslen = $1 + ($3 eq "I" ? $2 : 0);
my $mlen = $4;
if ( $1 <= 8 ) {
$a[3] += $1 + ($3 eq "D" ? $2 : 0);
my $n = 0;
$n++ while( $n < $mlen && $REF->{ $a[3]+$n } && $REF->{ $a[3]+$n } ne substr($a[9], $tslen+$n, 1));
$tslen += $n;
$mlen -= $n;
$a[3] += $n;
$a[5] =~ s/^\dM\d+[ID]\d+M/${tslen}S${mlen}M/;
$flag = 1;
}
}
if ($a[5] =~ /(\d+)([ID])(\d)M$/) {
my $tslen = $3 + ($2 eq "I" ? $1 : 0);
$tslen .= "S";
if ( $3 <= 8 ) {
$a[5] =~ s/\d+[ID]\d+M$/$tslen/;
$flag = 1;
}
}
# Combine two deletions and insertion into one complex if they are close
if ($a[5] =~ /^(.*?)(\d+)M(\d+)D(\d)M(\d+)I(\d)M(\d+)D/) {
my $tslen = $4 + $5 + $6;
my $dlen = $3 + $4 + $6 + $7;
my $mid = $4 + $6;
my $ov5 = $1;
my $refoff = $a[3] + $2;
my $rdoff = $2;
my $RDOFF = $2;
if ( $ov5 ) {
my @rdp = $ov5 =~ /(\d+)[MIS]/g; # read position
my @rfp = $ov5 =~ /(\d+)[MD]/g; # reference position
$refoff += $_ foreach(@rfp);
$rdoff += $_ foreach(@rdp);
}
my $rn = 0;
$rn++ while( $REF->{ $refoff + $rn } && $REF->{ $refoff + $rn } eq substr($a[9], $rdoff + $rn, 1) );
$RDOFF += $rn;
$dlen -= $rn;
$tslen -= $rn;
if ( $mid <= 10 ) {
#print STDERR "B: $rn $RDOFF $dlen $tslen $a[5]\n";
$a[5] =~ s/\d+M\d+D\dM\d+I\dM\d+D/${RDOFF}M${dlen}D${tslen}I/;
$flag = 1;
}
}
# Combine two close deletions (<10bp) into one
if ($a[5] =~ /(\d+)D(\d)M(\d+)([DI])(\d+I)?/ ) {
my $dlen = $1 + $2 + ($4 eq "D" ? $3 : 0);
my $ilen = $2 + ($4 eq "I" ? $3 : 0);
my $istr = $5;
if ( $istr && $4 eq "D") {
$ilen += $1 if ( $istr =~ /(\d+)I/ );
}
$a[5] =~ s/\d+D\dM\d+[DI](\d+I)?/${dlen}D${ilen}I/;
$flag = 1;
}
# Combine two close indels (<10bp) into one
if ($a[5] =~ /(\d+)I(\d)M(\d+)([DI])(\d+I)?/ ) {
my $dlen = $2 + ($4 eq "D" ? $3 : 0);
my $ilen = $1 + $2 + ($4 eq "I" ? $3 : 0);
my $istr = $5;
if ( $istr && $4 eq "D" ) {
$ilen += $1 if ( $istr =~ /(\d+)I/ );
}
$a[5] =~ s/\d+I\dM\d+[DI](\d+I)?/${dlen}D${ilen}I/;
$flag = 1;
}
last unless ( $flag );
}
# The following two clauses to capture sometimes mis-softly clipped reads by aligner
if ($a[5] =~ /^(.*?)(\d+)M(\d+)S$/) {
my $ov5 = $1;
my $mch = $2;
my $soft = $3;
my $refoff = $a[3] + $2;
my $rdoff = $2;
if ( $ov5 ) {
my @rdp = $ov5 =~ /(\d+)[MIS]/g; # read position
my @rfp = $ov5 =~ /(\d+)[MD]/g; # reference position
$refoff += $_ foreach(@rfp);
$rdoff += $_ foreach(@rdp);
}
my $rn = 0;
$rn++ while( $rn + 1 < $soft && $REF->{ $refoff + $rn + 1} && $REF->{ $refoff + $rn + 1} eq substr($a[9], $rdoff + $rn + 1, 1) && ord(substr($a[10], $rdoff + $rn + 1, 1))-33 > $LOWQUAL );
if ( $rn > 3 || ($REF->{ $refoff } && $REF->{ $refoff } eq substr($a[9], $rdoff, 1) )) {
$mch += $rn + 1;
$soft -= $rn + 1;
if ( $soft > 0 ) {
$a[5] =~ s/\d+M\d+S$/${mch}M${soft}S/;
} else {
$a[5] =~ s/\d+M\d+S$/${mch}M/;
}
$rn++;
}
if ( $rn == 0 ) {
$rn++ while( $REF->{ $refoff - $rn - 1 } && $REF->{ $refoff - $rn - 1 } ne substr($a[9], $rdoff - $rn - 1, 1) );
if ( $rn > 0 ) {
$soft += $rn;
$mch -= $rn;
$a[5] =~ s/\d+M\d+S$/${mch}M${soft}S/;
}
}
}
if ($a[5] =~ /^(\d+)S(\d+)M/ ) {
my $mch = $2;
my $soft = $1;
my $rn = 0;
$rn++ while( $rn + 1 < $soft && $REF->{ $a[3] - $rn - 2} && $REF->{ $a[3] - $rn - 2} eq substr($a[9], $soft - $rn - 2, 1) && ord(substr($a[10], $soft - $rn - 2, 1))-33 > $LOWQUAL );
if ( $rn > 3 || ($REF->{ $a[3] - 1 } && $REF->{ $a[3] - 1 } eq substr($a[9], $soft - 1, 1))) {
$mch += $rn + 1;
$soft -= $rn + 1;
if ( $soft > 0 ) {
$a[5] =~ s/^\d+S\d+M/${soft}S${mch}M/;
} else {
$a[5] =~ s/^\d+S\d+M/${mch}M/;
}
$a[3] -= $rn + 1;
$rn++;
}
if ( $rn == 0 ) {
$rn++ while( $REF->{ $a[3] + $rn } && $REF->{ $a[3] + $rn } ne substr($a[9], $soft + $rn, 1));
if( $rn > 0 ) {
$soft += $rn;
$mch -= $rn;
$a[5] =~ s/^\d+S\d+M/${soft}S${mch}M/;
$a[3] += $rn;
}
}
}
my $start = $a[3];
my @cigar = $a[5] =~ /(\d+)([A-Z])/g;
my @segs = $a[5] =~ /(\d+)[MI]/g; # Only match and insertion counts toward read length
my @segs2 = $a[5] =~ /(\d+)[MIS]/g; # For total length, including soft-clipped bases
my $rlen = 0; $rlen += $_ foreach(@segs); #$stat->sum(\@segs); # The read length for matched bases
my $rlen2= 0; $rlen2 += $_ foreach(@segs2); #$stat->sum(\@segs2); # The total length, including soft-clipped bases
$RLEN = $rlen2 if ($rlen2 > $RLEN); # Determine the read length
for(my $ci = 0; $ci < @cigar; $ci += 2) {
my $m = $cigar[$ci];
my $C = $cigar[$ci+1];
$C = "S" if ( ($ci == 0 || $ci == @cigar - 2) && $C eq "I" ); # Treat insertions at the edge as softclipping
if ( $C eq "N" ) {
$SPLICE{ ($start-1) . "-" . ($start+$m-1) }++;
$start += $m;
$offset = 0;
next;
} elsif ( $C eq "S" ) {
if ( $ci == 0 ) { # 5' soft clipped
# align softclipped but matched sequences due to mis-softclipping
while( $m-1 >= 0 && $start - 1 > 0 && $start - 1 <= $CHRS{ $chr } && $REF->{ $start-1 } && $REF->{ $start-1 } eq substr($a[9], $m-1, 1) && ord(substr($a[10],$m-1, 1))-33 > 10) {
$hash{ $start - 1 }->{ $REF->{ $start - 1 } }->{ cnt } = 0 unless( $hash{ $start - 1 }->{ $REF->{ $start - 1 } }->{ cnt } );
addCnt($hash{ $start - 1 }->{ $REF->{ $start - 1 } }, $dir, $m, ord(substr($a[10],$m-1, 1))-33, $a[4], $nm);
$cov{ $start - 1 }++;
$start--; $m--;
}
if ( $m > 0 ) {
my $q = 0;
my $qn = 0;
my $lowqcnt = 0;
for(my $si = $m-1; $si >= 0; $si--) {
last if ( substr($a[9], $si, 1) eq "N" );
my $tq = ord(substr($a[10], $si, 1))-33;
$lowqcnt++ if ($tq <= 12);
last if ( $lowqcnt > 1 );
$q += $tq;
$qn++;
}
if ( $qn >= 1 && $qn > $lowqcnt && $start >= $START - $BUFFER && $start <= $END + $BUFFER ) {
for(my $si = $m-1; $m - $si <= $qn; $si--) {
$sclip5{ $start }->{ nt }->[$m-1-$si]->{ substr($a[9], $si, 1) }++;
$sclip5{ $start }->{ seq }->[$m-1-$si]->{ substr($a[9], $si, 1) }->{ cnt } = 0 unless( $sclip5{ $start }->{ seq }->[$m-1-$si]->{ substr($a[9], $si, 1) }->{ cnt });
addCnt($sclip5{ $start }->{ seq }->[$m-1-$si]->{ substr($a[9], $si, 1) }, $dir, $si - ($m-$qn), ord(substr($a[10], $si, 1))-33, $a[4], $nm);
}
$sclip5{ $start }->{ cnt } = 0 unless( $sclip5{ $start }->{ cnt } );
addCnt( $sclip5{ $start }, $dir, $m, $q/$qn, $a[4], $nm);
}
}
$m = $cigar[$ci];
} elsif ( $ci == @cigar - 2 ) { # 3' soft clipped
while( $n < length($a[9]) && $REF->{ $start } && $REF->{ $start } eq substr($a[9], $n, 1) && ord(substr($a[10], $n, 1))-33 > 10) {
$hash{ $start }->{ $REF->{ $start } }->{ cnt } = 0 unless( $hash{ $start }->{ $REF->{ $start } }->{ cnt } );
addCnt($hash{ $start }->{ $REF->{ $start } }, $dir, $rlen2-$p, ord(substr($a[10], $n, 1))-33, $a[4], $nm);
$cov{$start}++;
$n++; $start++; $m--; $p++;
}
if ( length($a[9]) - $n > 0 ) {
my $q = 0;
my $qn = 0;
my $lowqcnt = 0;
for(my $si = 0; $si < $m; $si++) {
last if ( substr($a[9], $n+$si, 1) eq "N" );
my $tq = ord(substr($a[10], $n+$si, 1))-33;
$lowqcnt++ if ($tq <= 12);
last if ( $lowqcnt > 1 );
$q += $tq;
$qn++;
}
if ( $qn >= 1 && $qn > $lowqcnt && $start >= $START - $BUFFER && $start <= $END + $BUFFER ) {
for(my $si = 0; $si < $qn; $si++) {
$sclip3{ $start }->{ nt }->[$si]->{ substr($a[9], $n+$si, 1) }++;
$sclip3{ $start }->{ seq }->[$si]->{ substr($a[9], $n+$si, 1) }->{ cnt } = 0 unless( $sclip3{ $start }->{ seq }->[$si]->{ substr($a[9], $n+$si, 1) }->{ cnt } );
addCnt($sclip3{ $start }->{ seq }->[$si]->{ substr($a[9], $n+$si, 1) }, $dir, $qn - $si, ord(substr($a[10], $n+$si, 1))-33, $a[4], $nm);
}
$sclip3{ $start }->{ cnt } = 0 unless( $sclip3{ $start }->{ cnt } );
addCnt( $sclip3{ $start }, $dir, $m, $q/$qn, $a[4], $nm);
}
}
}
$n += $m;
$offset = 0;
$start = $a[3]; # had to reset the start due to softclipping adjustment
next;
} elsif ( $C eq "H" ) {
$offset = 0;
next;
} elsif ( $C eq "I" ) {
$offset = 0;
my $s = substr($a[9], $n, $m);
my $q = substr($a[10], $n, $m);
my $ss = "";
my ($multoffs, $multoffp, $nmoff) = (0, 0, 0); # For multiple indels within 10bp
if ( $cigar[$ci+2] && $cigar[$ci+2] <= $VEXT && $cigar[$ci+3] =~ /M/ && $cigar[$ci+5] && $cigar[$ci+5] =~ /[ID]/ ) {
$s .= "#" . substr($a[9], $n+$m, $cigar[$ci+2]);
$q .= substr($a[10], $n+$m, $cigar[$ci+2]);
$s .= $cigar[$ci+5] eq "I" ? ("^" . substr($a[9], $n+$m+$cigar[$ci+2], $cigar[$ci+4])) : ("^" . $cigar[$ci+4]);
$q .= $cigar[$ci+5] eq "I" ? substr($a[10], $n+$m+$cigar[$ci+2], $cigar[$ci+4]) : substr($a[10], $n+$m+$cigar[$ci+2], 1);
$multoffs += $cigar[$ci+2] + ($cigar[$ci+5] eq "D" ? $cigar[$ci+4] : 0);
$multoffp += $cigar[$ci+2] + ($cigar[$ci+5] eq "I" ? $cigar[$ci+4] : 0);
$ci += 4;
if ( $cigar[$ci+6] && $cigar[$ci+7] eq "M" ) {
my ($toffset, $tss, $tq) = findOffset($start+$multoffs, $n+$m+$multoffp, $cigar[$ci+6], \$a[9], \$a[10], $REF, \%cov);
($offset, $ss) = ($toffset, $tss);
$q .= $tq;
}
} else {
if ( $cigar[$ci+3] && $cigar[$ci+3] =~ /M/ ) {
my $vsn = 0;
for(my $vi = 0; $vsn <= $VEXT && $vi < $cigar[$ci+2]; $vi++) {
last if ( substr($a[9], $n+$m+$vi, 1) eq "N" );
last if ( ord(substr($a[10], $n+$m+$vi, 1))-33 < $GOODQ );
if ($REF->{ $start+$vi } && substr($a[9], $n+$m+$vi, 1) ne $REF->{ $start+$vi }) {
$offset = $vi+1;
$nmoff++;
$vsn = 0;
} elsif ($REF->{ $start+$vi } && substr($a[9], $n+$m+$vi, 1) eq $REF->{ $start+$vi }) {
$vsn++;
}
}
if ($offset) {
$ss .= substr($a[9], $n+$m, $offset);
$q .= substr($a[10], $n+$m, $offset);
for( my $osi = 0; $osi < $offset; $osi++ ) {
$cov{ $start + $osi }++;
}
}
}
}
$s = "$s&$ss" if ( $offset > 0);
if ( $start - 1 >= $START && $start -1 <= $END && $s !~ /N/ ) {
$ins{ $start - 1 }->{ "+$s" }++;
$hash{ $start - 1 }->{ I }->{ "+$s" }->{ $dir }++;
my $hv = $hash{ $start - 1 }->{ I }->{ "+$s" };
$hv->{ cnt }++;
my $tp = $p < $rlen-$p ? $p + 1: $rlen-$p;
my $tmpq = 0;
for(my $i = 0; $i < length($q); $i++) {
$tmpq += ord(substr($q, $i, 1))-33;
}
$tmpq /= length($q);
unless( $hv->{ pstd } ) {
$hv->{ pstd } = 0;
$hv->{ pstd } = 1 if ($hv->{ pp } && $tp != $hv->{ pp });
}
unless( $hv->{ qstd } ) {
$hv->{ qstd } = 0;