10 &usage if (@ARGV < 1);
12 my $command = shift(@ARGV);
13 my %func = (showALEN=>\&showALEN, pileup2fq=>\&pileup2fq, varFilter=>\&varFilter, plp2vcf=>\&plp2vcf,
14 unique=>\&unique, uniqcmp=>\&uniqcmp, sra2hdr=>\&sra2hdr, sam2fq=>\&sam2fq);
16 die("Unknown command \"$command\".\n") if (!defined($func{$command}));
25 die(qq/Usage: samtools.pl showALEN <in.sam>\n/) if (@ARGV == 0 && -t STDIN);
28 next if (/^\@/ || @t < 11);
31 s/(\d+)[MI]/$l+=$1/eg;
32 print join("\t", @t[0..5]), "\t$l\t", join("\t", @t[6..$#t]), "\n";
45 # W too many SNPs in a window (SNP only)
46 # G close to a high-quality indel (SNP only)
47 # Q low RMS mapping quality (SNP only)
48 # g close to another indel with higher quality (indel only)
49 # s low SNP quality (SNP only)
50 # i low indel quality (indel only)
53 my %opts = (d=>3, D=>100, l=>30, Q=>25, q=>10, G=>25, s=>100, w=>10, W=>10, N=>2, p=>undef, S=>'', i=>'');
54 getopts('pq:d:D:l:Q:w:W:N:G:S:i:', \%opts);
56 Usage: samtools.pl varFilter [options] <in.cns-pileup>
58 Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}]
59 -q INT minimum RMS mapping quality for gaps [$opts{q}]
60 -d INT minimum read depth [$opts{d}]
61 -D INT maximum read depth [$opts{D}]
62 -S INT minimum SNP quality [$opts{S}]
63 -i INT minimum indel quality [$opts{i}]
65 -G INT min indel score for nearby SNP filtering [$opts{G}]
66 -w INT SNP within INT bp around a gap to be filtered [$opts{w}]
68 -W INT window size for filtering dense SNPs [$opts{W}]
69 -N INT max number of SNPs in a window [$opts{N}]
71 -l INT window size for filtering adjacent gaps [$opts{l}]
73 -p print filtered variants
74 \n/) if (@ARGV == 0 && -t STDIN);
76 # calculate the window size
77 my ($ol, $ow, $oW) = ($opts{l}, $opts{w}, $opts{W});
78 my $max_dist = $ol > $ow? $ol : $ow;
79 $max_dist = $oW if ($max_dist < $oW);
81 my @staging; # (indel_filtering_score, flt_tag)
84 next if (uc($t[2]) eq uc($t[3]) || $t[3] eq '*/*'); # skip non-var sites
85 # clear the out-of-range elements
87 # Still on the same chromosome and the first element's window still affects this position?
88 last if ($staging[0][3] eq $t[0] && $staging[0][4] + $staging[0][2] + $max_dist >= $t[1]);
89 varFilter_aux(shift(@staging), $opts{p}); # calling a function is a bit slower, not much
91 my ($flt, $score) = (0, -1);
92 # first a simple filter
93 if ($t[7] < $opts{d}) {
95 } elsif ($t[7] > $opts{D}) {
98 if ($t[2] eq '*') { # an indel
99 if ($opts{i} && $opts{i}>$t[5]) { $flt = 8; }
101 elsif ($opts{S} && $opts{S}>$t[5]) { $flt = 7; } # SNP
103 # site dependent filters
106 if ($t[2] eq '*') { # an indel
107 # If deletion, remember the length of the deletion
108 my ($a,$b) = split(m{/},$t[3]);
109 my $alen = length($a) - 1;
110 my $blen = length($b) - 1;
113 if ( substr($a,0,1) eq '-' ) { $len=$alen; }
115 elsif ( substr($b,0,1) eq '-' ) { $len=$blen; }
117 $flt = 1 if ($t[6] < $opts{q});
119 if ($t[5] >= $opts{G}) {
120 for my $x (@staging) {
121 # Is it a SNP and is it outside the SNP filter window?
122 next if ($x->[0] >= 0 || $x->[4] + $x->[2] + $ow < $t[1]);
123 $x->[1] = 5 if ($x->[1] == 0);
126 # calculate the filtering score (different from indel quality)
128 $score += $opts{s} * $t[10] if ($t[8] ne '*');
129 $score += $opts{s} * $t[11] if ($t[9] ne '*');
130 # check the staging list for indel filtering
131 for my $x (@staging) {
132 # Is it a SNP and is it outside the gap filter window
133 next if ($x->[0] < 0 || $x->[4] + $x->[2] + $ol < $t[1]);
134 if ($x->[0] < $score) {
141 $flt = 1 if ($t[6] < $opts{Q});
142 # check adjacent SNPs
144 for my $x (@staging) {
145 ++$k if ($x->[0] < 0 && $x->[4] + $x->[2] + $oW >= $t[1] && ($x->[1] == 0 || $x->[1] == 4 || $x->[1] == 5));
147 # filtering is necessary
150 for my $x (@staging) {
151 $x->[1] = 4 if ($x->[0] < 0 && $x->[4] + $x->[2] + $oW >= $t[1] && $x->[1] == 0);
153 } else { # then check gap filter
154 for my $x (@staging) {
155 next if ($x->[0] < 0 || $x->[4] + $x->[2] + $ow < $t[1]);
156 if ($x->[0] >= $opts{G}) {
163 push(@staging, [$score, $flt, $len, @t]);
165 # output the last few elements in the staging list
167 varFilter_aux(shift @staging, $opts{p});
172 my ($first, $is_print) = @_;
173 if ($first->[1] == 0) {
174 print join("\t", @$first[3 .. @$first-1]), "\n";
175 } elsif ($is_print) {
176 print STDERR join("\t", substr("UQdDWGgsiX", $first->[1], 1), @$first[3 .. @$first-1]), "\n";
185 my %opts = (d=>3, D=>255, Q=>25, G=>25, l=>10);
186 getopts('d:D:Q:G:l:', \%opts);
188 Usage: samtools.pl pileup2fq [options] <in.cns-pileup>
190 Options: -d INT minimum depth [$opts{d}]
191 -D INT maximum depth [$opts{D}]
192 -Q INT min RMS mapQ [$opts{Q}]
193 -G INT minimum indel score [$opts{G}]
194 -l INT indel filter winsize [$opts{l}]\n
195 /) if (@ARGV == 0 && -t STDIN);
197 my ($last_chr, $seq, $qual, @gaps, $last_pos);
205 if ($last_chr ne $t[0]) {
206 &p2q_post_process($last_chr, \$seq, \$qual, \@gaps, $opts{l}) if ($last_chr);
209 $seq = ''; $qual = '';
212 if ($t[1] - $last_pos != 1) {
213 $seq .= 'n' x ($t[1] - $last_pos - 1);
214 $qual .= '!' x ($t[1] - $last_pos - 1);
217 push(@gaps, $t[1]) if ($t[5] >= $opts{G});
219 $seq .= ($t[6] >= $_Q && $t[7] >= $_d && $t[7] <= $_D)? uc($t[3]) : lc($t[3]);
221 $q = 126 if ($q > 126);
226 &p2q_post_process($last_chr, \$seq, \$qual, \@gaps, $opts{l});
229 sub p2q_post_process {
230 my ($chr, $seq, $qual, $gaps, $l) = @_;
231 &p2q_filter_gaps($seq, $gaps, $l);
232 print "\@$chr\n"; &p2q_print_str($seq);
233 print "+\n"; &p2q_print_str($qual);
236 sub p2q_filter_gaps {
237 my ($seq, $gaps, $l) = @_;
239 my $x = $g > $l? $g - $l : 0;
240 substr($$seq, $x, $l + $l) = lc(substr($$seq, $x, $l + $l));
247 for (my $i = 0; $i < $l; $i += 60) {
248 print substr($$s, $i, 60), "\n";
257 my %opts = (n=>20, p=>'');
258 getopts('n:p:', \%opts);
259 die("Usage: samtools.pl sam2fq [-n 20] [-p <prefix>] <inp.sam>\n") if (@ARGV == 0 && -t STDIN);
260 if ($opts{p} && $opts{n} > 1) {
263 for (0 .. $opts{n}-1) {
264 open($fh[$_], sprintf("| gzip > $pre.%.3d.fq.gz", $_)) || die;
271 next if ($t[9] eq '*');
272 my ($name, $seq, $qual);
273 if ($t[1] & 16) { # reverse strand
274 $seq = reverse($t[9]);
275 $qual = reverse($t[10]);
276 $seq =~ tr/ACGTacgt/TGCAtgca/;
278 ($seq, $qual) = @t[9,10];
281 $name .= "/1" if ($t[1] & 0x40);
282 $name .= "/2" if ($t[1] & 0x80);
283 print {$fh[$i]} "\@$name\n$seq\n";
285 print {$fh[$i]} "+\n$qual\n";
287 $i = 0 if (++$i == $opts{n});
289 close($fh[$_]) for (0 .. $opts{n}-1);
291 die("To be implemented.\n");
299 # This subroutine does not use an XML parser. It requires that the SRA
300 # XML files are properly formated.
304 die("Usage: samtools.pl sra2hdr <SRA.prefix>\n") if (@ARGV == 0);
308 my $sample = 'UNKNOWN';
309 open($fh, "$pre.sample.xml") || die;
311 $sample = $1 if (/<SAMPLE.*alias="([^"]+)"/i);
316 open($fh, "$pre.experiment.xml") || die;
318 if (/<EXPERIMENT.*accession="([^\s"]+)"/i) {
320 } elsif (/<LIBRARY_NAME>\s*(\S+)\s*<\/LIBRARY_NAME>/i) {
327 open($fh, "$pre.run.xml") || die;
329 if (/<RUN.*accession="([^\s"]+)"/i) {
331 } elsif (/<EXPERIMENT_REF.*accession="([^\s"]+)"/i) {
332 print "\@RG\tID:$run\tSM:$sample\tLB:$exp2lib{$1}\n";
333 } elsif (/<FILE.*filename="([^\s"]+)"/i) {
335 } elsif (/<\/RUN>/i) {
337 print STDERR "$fn[0]\t$run\n";
340 print STDERR "$fn[$_]\t$run", "_", $_+1, "\n";
353 my %opts = (f=>250.0, q=>5, r=>2, a=>1, b=>3);
354 getopts('Qf:q:r:a:b:m', \%opts);
355 die("Usage: samtools.pl unique [-f $opts{f}] <in.sam>\n") if (@ARGV == 0 && -t STDIN);
357 my $recal_Q = !defined($opts{Q});
358 my $multi_only = defined($opts{m});
363 $score = $1 if (/AS:i:(\d+)/);
366 if ($score < 0) { # AS tag is unavailable
368 my ($mm, $go, $ge) = (0, 0, 0);
369 $cigar =~ s/(\d+)[ID]/++$go,$ge+=$1/eg;
371 $cigar =~ s/(\d+)M/$mm+=$1/eg;
372 $score = $mm * $opts{a} - $go * $opts{q} - $ge * $opts{r}; # no mismatches...
374 $score = 1 if ($score < 1);
375 if ($t[0] ne $last) {
376 &unique_aux(\@a, $opts{f}, $recal_Q, $multi_only) if (@a);
379 push(@a, [$score, \@t]);
381 &unique_aux(\@a, $opts{f}, $recal_Q, $multi_only) if (@a);
385 my ($a, $fac, $is_recal, $multi_only) = @_;
386 my ($max, $max2, $max_i) = (0, 0, -1);
387 for (my $i = 0; $i < @$a; ++$i) {
388 if ($a->[$i][0] > $max) {
389 $max2 = $max; $max = $a->[$i][0]; $max_i = $i;
390 } elsif ($a->[$i][0] > $max2) {
395 if (!$multi_only || @$a > 1) {
396 my $q = int($fac * ($max - $max2) / $max + .499);
397 $q = 250 if ($q > 250);
398 $a->[$max_i][1][4] = $q < 250? $q : 250;
401 print join("\t", @{$a->[$max_i][1]});
406 # uniqcmp: compare two SAM files
410 my %opts = (q=>10, s=>100);
411 getopts('pq:s:', \%opts);
412 die("Usage: samtools.pl uniqcmp <in1.sam> <in2.sam>\n") if (@ARGV < 2);
414 warn("[uniqcmp] read the first file...\n");
415 &uniqcmp_aux($ARGV[0], \%a, 0);
416 warn("[uniqcmp] read the second file...\n");
417 &uniqcmp_aux($ARGV[1], \%a, 1);
418 warn("[uniqcmp] stats...\n");
420 $cnt[$_] = 0 for (0..9);
421 for my $x (keys %a) {
424 if (defined($p->[0]) && defined($p->[1])) {
425 $z = ($p->[0][0] == $p->[1][0] && $p->[0][1] eq $p->[1][1] && abs($p->[0][2] - $p->[1][2]) < $opts{s})? 0 : 1;
426 if ($p->[0][3] >= $opts{q} && $p->[1][3] >= $opts{q}) {
428 } elsif ($p->[0][3] >= $opts{q}) {
430 } elsif ($p->[1][3] >= $opts{q}) {
433 print STDERR "$x\t$p->[0][1]:$p->[0][2]\t$p->[0][3]\t$p->[0][4]\t$p->[1][1]:$p->[1][2]\t$p->[1][3]\t$p->[1][4]\t",
434 $p->[0][5]-$p->[1][5], "\n" if ($z && defined($opts{p}) && ($p->[0][3] >= $opts{q} || $p->[1][3] >= $opts{q}));
435 } elsif (defined($p->[0])) {
436 ++$cnt[$p->[0][3]>=$opts{q}? 6 : 7];
437 print STDERR "$x\t$p->[0][1]:$p->[0][2]\t$p->[0][3]\t$p->[0][4]\t*\t0\t*\t",
438 $p->[0][5], "\n" if (defined($opts{p}) && $p->[0][3] >= $opts{q});
440 print STDERR "$x\t*\t0\t*\t$p->[1][1]:$p->[1][2]\t$p->[1][3]\t$p->[1][4]\t",
441 -$p->[1][5], "\n" if (defined($opts{p}) && $p->[1][3] >= $opts{q});
442 ++$cnt[$p->[1][3]>=$opts{q}? 8 : 9];
445 print "Consistent (high, high): $cnt[0]\n";
446 print "Consistent (high, low ): $cnt[1]\n";
447 print "Consistent (low , high): $cnt[2]\n";
448 print "Inconsistent (high, high): $cnt[3]\n";
449 print "Inconsistent (high, low ): $cnt[4]\n";
450 print "Inconsistent (low , high): $cnt[5]\n";
451 print "Second missing (high): $cnt[6]\n";
452 print "Second missing (low ): $cnt[7]\n";
453 print "First missing (high): $cnt[8]\n";
454 print "First missing (low ): $cnt[9]\n";
458 my ($fn, $a, $which) = @_;
460 $fn = "samtools view $fn |" if ($fn =~ /\.bam/);
461 open($fh, $fn) || die;
465 # my $l = ($t[5] =~ /^(\d+)S/)? $1 : 0;
467 my ($x, $nm) = (0, 0);
468 $nm = $1 if (/NM:i:(\d+)/);
470 s/(\d+)[MI]/$x+=$1/eg;
471 @{$a->{$t[0]}[$which]} = (($t[1]&0x10)? 1 : 0, $t[2], $t[3]-$l, $t[4], "$x:$nm", $x - 4 * $nm);
479 next if ($t[3] eq '*/*');
480 if ($t[2] eq '*') { # indel
481 my @s = split("/", $t[3]);
487 push(@a, 'N'.substr($_, 1));
491 push(@b, 'N'.substr($_, 1));
494 if ($a[0] && $a[1]) {
495 if (length($a[0]) < length($a[1])) {
497 $alt = ($b[0] . ('N' x (length($a[1]) - length($a[0])))) . ",$b[1]";
498 } elsif (length($a[0]) > length($a[1])) {
500 $alt = ($b[1] . ('N' x (length($a[0]) - length($a[1])))) . ",$b[0]";
503 $alt = ($b[0] eq $b[1])? $b[0] : "$b[0],$b[1]";
506 $ref = $a[0]; $alt = $b[0];
508 print join("\t", @t[0,1], '.', $ref, $alt, $t[5], '.', '.'), "\n";
520 Program: samtools.pl (helper script for SAMtools)
522 Contact: Heng Li <lh3\@sanger.ac.uk>\n
523 Usage: samtools.pl <command> [<arguments>]\n
524 Command: varFilter filtering SNPs and short indels
525 pileup2fq generate fastq from `pileup -c'
526 showALEN print alignment length (ALEN) following CIGAR