10 &usage if (@ARGV < 1);
12 my $command = shift(@ARGV);
13 my %func = (showALEN=>\&showALEN, pileup2fq=>\&pileup2fq, varFilter=>\&varFilter,
14 unique=>\&unique, uniqcmp=>\&uniqcmp);
16 die("Unknown command \"$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..$#t]), "\n";
41 my %opts = (d=>3, D=>100, l=>30, Q=>25, q=>10, G=>25, s=>100, w=>10, W=>10, N=>2, p=>undef);
42 getopts('pq:d:D:l:Q:w:W:N:G:', \%opts);
44 Usage: samtools.pl varFilter [options] <in.cns-pileup>
46 Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}]
47 -q INT minimum RMS mapping quality for gaps [$opts{q}]
48 -d INT minimum read depth [$opts{d}]
49 -D INT maximum read depth [$opts{D}]
51 -G INT min indel score for nearby SNP filtering [$opts{G}]
52 -w INT SNP within INT bp around a gap to be filtered [$opts{w}]
54 -W INT window size for filtering dense SNPs [$opts{W}]
55 -N INT max number of SNPs in a window [$opts{N}]
57 -l INT window size for filtering adjacent gaps [$opts{l}]
59 -p print filtered variants
60 \n/) if (@ARGV == 0 && -t STDIN);
62 # calculate the window size
63 my ($ol, $ow, $oW) = ($opts{l}, $opts{w}, $opts{W});
64 my $max_dist = $ol > $ow? $ol : $ow;
65 $max_dist = $oW if ($max_dist < $oW);
67 my @staging; # (indel_filtering_score, flt_tag)
70 next if (uc($t[2]) eq uc($t[3]) || $t[3] eq '*/*'); # skip non-var sites
71 # clear the out-of-range elements
73 last if ($staging[0][2] eq $t[0] && $staging[0][3] + $max_dist >= $t[1]);
74 varFilter_aux(shift(@staging), $opts{p}); # calling a function is a bit slower, not much
76 my ($flt, $score) = (0, -1);
77 # first a simple filter
78 if ($t[7] < $opts{d}) {
80 } elsif ($t[7] > $opts{D}) {
83 # site dependent filters
85 if ($t[2] eq '*') { # an indel
86 $flt = 1 if ($t[6] < $opts{q});
88 if ($t[5] >= $opts{G}) {
89 for my $x (@staging) {
90 next if ($x->[0] >= 0 || $x->[3] + $ow < $t[1]);
91 $x->[1] = 5 if ($x->[1] == 0);
94 # calculate the filtering score (different from indel quality)
96 $score += $opts{s} * $t[10] if ($t[8] ne '*');
97 $score += $opts{s} * $t[11] if ($t[9] ne '*');
98 # check the staging list for indel filtering
99 for my $x (@staging) {
100 next if ($x->[0] < 0 || $x->[3] + $ol < $t[1]);
101 if ($x->[0] < $score) {
108 $flt = 1 if ($t[6] < $opts{Q});
109 # check adjacent SNPs
111 for my $x (@staging) {
112 ++$k if ($x->[0] < 0 && $x->[3] + $oW >= $t[1] && ($x->[1] == 0 || $x->[1] == 4 || $x->[1] == 5));
114 # filtering is necessary
117 for my $x (@staging) {
118 $x->[1] = 4 if ($x->[0] < 0 && $x->[3] + $oW >= $t[1] && $x->[1] == 0);
120 } else { # then check gap filter
121 for my $x (@staging) {
122 next if ($x->[0] < 0 || $x->[3] + $ow < $t[1]);
123 if ($x->[0] >= $opts{G}) {
130 push(@staging, [$score, $flt, @t]);
132 # output the last few elements in the staging list
134 varFilter_aux(shift @staging, $opts{p});
139 my ($first, $is_print) = @_;
140 if ($first->[1] == 0) {
141 print join("\t", @$first[2 .. @$first-1]), "\n";
142 } elsif ($is_print) {
143 print STDERR join("\t", substr("UQdDWGgX", $first->[1], 1), @$first[2 .. @$first-1]), "\n";
152 my %opts = (d=>3, D=>255, Q=>25, G=>25, l=>10);
153 getopts('d:D:Q:G:l:', \%opts);
155 Usage: samtools.pl pileup2fq [options] <in.cns-pileup>
157 Options: -d INT minimum depth [$opts{d}]
158 -D INT maximum depth [$opts{D}]
159 -Q INT min RMS mapQ [$opts{Q}]
160 -G INT minimum indel score [$opts{G}]
161 -l INT indel filter winsize [$opts{l}]\n
162 /) if (@ARGV == 0 && -t STDIN);
164 my ($last_chr, $seq, $qual, @gaps, $last_pos);
172 if ($last_chr ne $t[0]) {
173 &p2q_post_process($last_chr, \$seq, \$qual, \@gaps, $opts{l}) if ($last_chr);
176 $seq = ''; $qual = '';
179 if ($t[1] - $last_pos != 1) {
180 $seq .= 'n' x ($t[1] - $last_pos - 1);
181 $qual .= '!' x ($t[1] - $last_pos - 1);
184 push(@gaps, $t[1]) if ($t[5] >= $opts{G});
186 $seq .= ($t[6] >= $_Q && $t[7] >= $_d && $t[7] <= $_D)? uc($t[3]) : lc($t[3]);
188 $q = 126 if ($q > 126);
193 &p2q_post_process($last_chr, \$seq, \$qual, \@gaps, $opts{l});
196 sub p2q_post_process {
197 my ($chr, $seq, $qual, $gaps, $l) = @_;
198 &p2q_filter_gaps($seq, $gaps, $l);
199 print "\@$chr\n"; &p2q_print_str($seq);
200 print "+\n"; &p2q_print_str($qual);
203 sub p2q_filter_gaps {
204 my ($seq, $gaps, $l) = @_;
206 my $x = $g > $l? $g - $l : 0;
207 substr($$seq, $x, $l + $l) = lc(substr($$seq, $x, $l + $l));
214 for (my $i = 0; $i < $l; $i += 60) {
215 print substr($$s, $i, 60), "\n";
224 my %opts = (f=>250.0, q=>5, r=>2, a=>1, b=>3);
225 getopts('Qf:q:r:a:b:', \%opts);
226 die("Usage: samtools.pl unique [-f $opts{f}] <in.sam>\n") if (@ARGV == 0 && -t STDIN);
228 my $recal_Q = !defined($opts{Q});
233 $score = $1 if (/AS:i:(\d+)/);
236 if ($score < 0) { # AS tag is unavailable
238 my ($mm, $go, $ge) = (0, 0, 0);
239 $cigar =~ s/(\d+)[ID]/++$go,$ge+=$1/eg;
241 $cigar =~ s/(\d+)M/$mm+=$1/eg;
242 $score = $mm * $opts{a} - $go * $opts{q} - $ge * $opts{r}; # no mismatches...
244 $score = 1 if ($score < 1);
245 if ($t[0] ne $last) {
246 &unique_aux(\@a, $opts{f}, $recal_Q) if (@a);
249 push(@a, [$score, \@t]);
251 &unique_aux(\@a, $opts{f}, $recal_Q) if (@a);
255 my ($a, $fac, $is_recal) = @_;
256 my ($max, $max2, $max_i) = (0, 0, -1);
257 for (my $i = 0; $i < @$a; ++$i) {
258 if ($a->[$i][0] > $max) {
259 $max2 = $max; $max = $a->[$i][0]; $max_i = $i;
260 } elsif ($a->[$i][0] > $max2) {
265 my $q = int($fac * ($max - $max2) / $max + .499);
266 $q = 250 if ($q > 250);
267 $a->[$max_i][1][4] = $q < 250? $q : 250;
269 print join("\t", @{$a->[$max_i][1]});
274 # uniqcmp: compare two SAM files
278 my %opts = (q=>10, s=>100);
279 getopts('pq:s:', \%opts);
280 die("Usage: samtools.pl uniqcmp <in1.sam> <in2.sam>\n") if (@ARGV < 2);
282 warn("[uniqcmp] read the first file...\n");
283 &uniqcmp_aux($ARGV[0], \%a, 0);
284 warn("[uniqcmp] read the second file...\n");
285 &uniqcmp_aux($ARGV[1], \%a, 1);
286 warn("[uniqcmp] stats...\n");
288 $cnt[$_] = 0 for (0..9);
289 for my $x (keys %a) {
292 if (defined($p->[0]) && defined($p->[1])) {
293 $z = ($p->[0][0] == $p->[1][0] && $p->[0][1] eq $p->[1][1] && abs($p->[0][2] - $p->[1][2]) < $opts{s})? 0 : 1;
294 if ($p->[0][3] >= $opts{q} && $p->[1][3] >= $opts{q}) {
296 } elsif ($p->[0][3] >= $opts{q}) {
298 } elsif ($p->[1][3] >= $opts{q}) {
301 print STDERR "$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",
302 $p->[0][5]-$p->[1][5], "\n" if ($z && defined($opts{p}) && ($p->[0][3] >= $opts{q} || $p->[1][3] >= $opts{q}));
303 } elsif (defined($p->[0])) {
304 ++$cnt[$p->[0][3]>=$opts{q}? 6 : 7];
306 ++$cnt[$p->[1][3]>=$opts{q}? 8 : 9];
309 print "Consistent (high, high): $cnt[0]\n";
310 print "Consistent (high, low ): $cnt[1]\n";
311 print "Consistent (low , high): $cnt[2]\n";
312 print "Inconsistent (high, high): $cnt[3]\n";
313 print "Inconsistent (high, low ): $cnt[4]\n";
314 print "Inconsistent (low , high): $cnt[5]\n";
315 print "Second missing (high): $cnt[6]\n";
316 print "Second missing (low ): $cnt[7]\n";
317 print "First missing (high): $cnt[8]\n";
318 print "First missing (low ): $cnt[9]\n";
322 my ($fn, $a, $which) = @_;
324 $fn = "samtools view $fn |" if ($fn =~ /\.bam/);
325 open($fh, $fn) || die;
329 my $l = ($t[5] =~ /^(\d+)S/)? $1 : 0;
330 my ($x, $nm) = (0, 0);
331 $nm = $1 if (/NM:i:(\d+)/);
333 s/(\d+)[MI]/$x+=$1/eg;
334 @{$a->{$t[0]}[$which]} = (($t[1]&0x10)? 1 : 0, $t[2], $t[3]-$l, $t[4], "$x:$nm", $x - 4 * $nm);
345 Program: samtools.pl (helper script for SAMtools)
347 Contact: Heng Li <lh3\@sanger.ac.uk>\n
348 Usage: samtools.pl <command> [<arguments>]\n
349 Command: varFilter filtering SNPs and short indels
350 pileup2fq generate fastq from `pileup -c'
351 showALEN print alignment length (ALEN) following CIGAR