X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=misc%2Fsamtools.pl;h=d03c1c7f148dcca835946e7ccd9ee52f3b6dc00b;hb=0ccd9d36ebf35ce620a8248ecf4336c84065e6c0;hp=3dc6fe7d82df9f6255581345790d1477c34f92b7;hpb=ce9e1dd6d5aeffea22acbf8a02abf4642a8be84e;p=samtools.git diff --git a/misc/samtools.pl b/misc/samtools.pl index 3dc6fe7..d03c1c7 100755 --- a/misc/samtools.pl +++ b/misc/samtools.pl @@ -6,11 +6,12 @@ use strict; use warnings; use Getopt::Std; -my $version = '0.2.3'; +my $version = '0.3.3'; &usage if (@ARGV < 1); my $command = shift(@ARGV); -my %func = (snpFilter=>\&snpFilter, indelFilter=>\&indelFilter, showALEN=>\&showALEN, pileup2fq=>\&pileup2fq); +my %func = (showALEN=>\&showALEN, pileup2fq=>\&pileup2fq, varFilter=>\&varFilter, plp2vcf=>\&plp2vcf, + unique=>\&unique, uniqcmp=>\&uniqcmp, sra2hdr=>\&sra2hdr, sam2fq=>\&sam2fq); die("Unknown command \"$command\".\n") if (!defined($func{$command})); &{$func{$command}}; @@ -24,124 +25,164 @@ sub showALEN { die(qq/Usage: samtools.pl showALEN \n/) if (@ARGV == 0 && -t STDIN); while (<>) { my @t = split; + next if (/^\@/ || @t < 11); my $l = 0; $_ = $t[5]; - s/(\d+)[SMI]/$l+=$1/eg; + s/(\d+)[MI]/$l+=$1/eg; print join("\t", @t[0..5]), "\t$l\t", join("\t", @t[6..$#t]), "\n"; } } # -# indelFilter +# varFilter # -sub indelFilter { - my %opts = (D=>100, m=>10, r=>undef, s=>100); - getopts('D:m:rs:', \%opts); # -s for scaling factor in score calculation +# +# Filtration code: +# +# d low depth +# D high depth +# W too many SNPs in a window (SNP only) +# G close to a high-quality indel (SNP only) +# Q low RMS mapping quality (SNP only) +# g close to another indel with higher quality (indel only) +# s low SNP quality (SNP only) +# i low indel quality (indel only) +sub varFilter { + my %opts = (d=>3, D=>100, l=>30, Q=>25, q=>10, G=>25, s=>100, w=>10, W=>10, N=>2, p=>undef, S=>'', i=>''); + getopts('pq:d:D:l:Q:w:W:N:G:S:i:', \%opts); die(qq/ -Usage: samtools.pl indelFilter [options] \n -Options: -D INT maximum read depth [$opts{D}] - -m INT minimum distance between two adjacent indels [$opts{m}] +Usage: samtools.pl varFilter [options] + +Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] + -q INT minimum RMS mapping quality for gaps [$opts{q}] + -d INT minimum read depth [$opts{d}] + -D INT maximum read depth [$opts{D}] + -S INT minimum SNP quality [$opts{S}] + -i INT minimum indel quality [$opts{i}] + + -G INT min indel score for nearby SNP filtering [$opts{G}] + -w INT SNP within INT bp around a gap to be filtered [$opts{w}] + + -W INT window size for filtering dense SNPs [$opts{W}] + -N INT max number of SNPs in a window [$opts{N}] + + -l INT window size for filtering adjacent gaps [$opts{l}] + + -p print filtered variants \n/) if (@ARGV == 0 && -t STDIN); - my (@arr1, @arr2); - my ($curr, $last) = (\@arr1, \@arr2); - my $is_ref = defined($opts{r})? 1 : 0; + # calculate the window size + my ($ol, $ow, $oW) = ($opts{l}, $opts{w}, $opts{W}); + my $max_dist = $ol > $ow? $ol : $ow; + $max_dist = $oW if ($max_dist < $oW); + # the core loop + my @staging; # (indel_filtering_score, flt_tag) while (<>) { my @t = split; - next if ($t[2] ne '*'); - if (!$is_ref) { - next if ($t[3] eq '*/*'); - next if ($t[5] == 0); - } - next if ($t[7] > $opts{D}); - # calculate indel score - my $score = $t[5]; - $score += $opts{s} * $t[10] if ($t[8] ne '*'); - $score += $opts{s} * $t[11] if ($t[9] ne '*'); - @$curr = ($t[0], $t[1], $score, $_); - my $do_swap = 1; - if (defined $last->[0]) { - if ($curr->[0] eq $last->[0] && $last->[1] + $opts{m} > $curr->[1]) { - $do_swap = 0 if ($last->[2] > $curr->[2]); - } else { # then print - print $last->[3]; - } + next if (uc($t[2]) eq uc($t[3]) || $t[3] eq '*/*'); # skip non-var sites + # clear the out-of-range elements + while (@staging) { + # Still on the same chromosome and the first element's window still affects this position? + last if ($staging[0][3] eq $t[0] && $staging[0][4] + $staging[0][2] + $max_dist >= $t[1]); + varFilter_aux(shift(@staging), $opts{p}); # calling a function is a bit slower, not much } - if ($do_swap) { - my $tmp = $curr; $curr = $last; $last = $tmp; + my ($flt, $score) = (0, -1); + # first a simple filter + if ($t[7] < $opts{d}) { + $flt = 2; + } elsif ($t[7] > $opts{D}) { + $flt = 3; } - } - print $last->[3] if (defined $last->[0]); -} + if ($t[2] eq '*') { # an indel + if ($opts{i} && $opts{i}>$t[5]) { $flt = 8; } + } + elsif ($opts{S} && $opts{S}>$t[5]) { $flt = 7; } # SNP -# -# snpFilter -# - -sub snpFilter { - my %opts = (f=>'', Q=>40, d=>3, w=>10, D=>0, N=>2, W=>10, q=>20, s=>50); - getopts('f:s:w:q:Q:d:D:W:N:', \%opts); - die(qq{ -Usage: samtools.pl snpFilter [options] - -Options: -d INT minimum depth to call a SNP [$opts{d}] - -D INT maximum depth, 0 to ignore [$opts{D}] - -Q INT required max mapping quality of the reads covering the SNP [$opts{Q}] - -q INT minimum SNP quality [$opts{q}] - - -f FILE filtered samtools indels [null] - -s INT minimum samtols indel score [$opts{s}] - -w INT SNP within INT bp around an indel to be filtered [$opts{w}] - - -W INT window size for filtering dense SNPs [$opts{W}] - -N INT maximum number of SNPs in a window [$opts{N}] -\n}) unless (@ARGV); - my (%hash, $fh); - my $skip = $opts{w}; - $opts{D} = 100000000 if ($opts{D} == 0); - if ($opts{f}) { # filtered samtools indel - my $n = 0; - open($fh, $opts{f}) || die; - while (<$fh>) { - my @t = split; - next if ($t[2] ne '*' || $t[3] eq '*/*' || $t[5] < $opts{s}); - for (my $x = $t[1] - $skip + 1; $x < $t[1] + $skip; ++$x) { - $hash{$t[0],$x} = 1; + # site dependent filters + my $len=0; + if ($flt == 0) { + if ($t[2] eq '*') { # an indel + # If deletion, remember the length of the deletion + my ($a,$b) = split(m{/},$t[3]); + my $alen = length($a) - 1; + my $blen = length($b) - 1; + if ( $alen>$blen ) + { + if ( substr($a,0,1) eq '-' ) { $len=$alen; } + } + elsif ( substr($b,0,1) eq '-' ) { $len=$blen; } + + $flt = 1 if ($t[6] < $opts{q}); + # filtering SNPs + if ($t[5] >= $opts{G}) { + for my $x (@staging) { + # Is it a SNP and is it outside the SNP filter window? + next if ($x->[0] >= 0 || $x->[4] + $x->[2] + $ow < $t[1]); + $x->[1] = 5 if ($x->[1] == 0); + } + } + # calculate the filtering score (different from indel quality) + $score = $t[5]; + $score += $opts{s} * $t[10] if ($t[8] ne '*'); + $score += $opts{s} * $t[11] if ($t[9] ne '*'); + # check the staging list for indel filtering + for my $x (@staging) { + # Is it a SNP and is it outside the gap filter window + next if ($x->[0] < 0 || $x->[4] + $x->[2] + $ol < $t[1]); + if ($x->[0] < $score) { + $x->[1] = 6; + } else { + $flt = 6; last; + } + } + } else { # a SNP + $flt = 1 if ($t[6] < $opts{Q}); + # check adjacent SNPs + my $k = 1; + for my $x (@staging) { + ++$k if ($x->[0] < 0 && $x->[4] + $x->[2] + $oW >= $t[1] && ($x->[1] == 0 || $x->[1] == 4 || $x->[1] == 5)); + } + # filtering is necessary + if ($k > $opts{N}) { + $flt = 4; + for my $x (@staging) { + $x->[1] = 4 if ($x->[0] < 0 && $x->[4] + $x->[2] + $oW >= $t[1] && $x->[1] == 0); + } + } else { # then check gap filter + for my $x (@staging) { + next if ($x->[0] < 0 || $x->[4] + $x->[2] + $ow < $t[1]); + if ($x->[0] >= $opts{G}) { + $flt = 5; last; + } + } + } } } - close($fh); + push(@staging, [$score, $flt, $len, @t]); } - my (@last, $last_chr); - $last_chr = ''; - while (<>) { - my @t = split; - next if ($t[2] eq '*' || $hash{$t[0],$t[1]}); - next if ($t[2] eq $t[3]); - my $is_good = ($t[7] >= $opts{d} && $t[7] <= $opts{D} && $t[6] >= $opts{Q} && $t[5] >= $opts{q})? 1 : 0; - next unless ($is_good); # drop - if ($t[0] ne $last_chr) { # a different chr, print - map { print $_->{L} if ($_->{F}) } @last; - @last = (); - $last_chr = $t[0]; - } - # The following block implemented by Nathans Weeks. - push(@last, {L => $_, X => $t[1], F => 1}); # Enqueue current SNP - if ($#last == $opts{N}) { # number of SNPs in queue is N+1 - if ($last[$#last]{X} - $last[0]{X} < $opts{W}) { # if all within window W - map {$_->{F} = 0} @last; # all SNPs in the window of size W are "bad" - } - print STDOUT $last[0]{L} if ($last[0]{F}); # print first SNP if good - shift @last # dequeue first SNP - } + # output the last few elements in the staging list + while (@staging) { + varFilter_aux(shift @staging, $opts{p}); + } +} + +sub varFilter_aux { + my ($first, $is_print) = @_; + if ($first->[1] == 0) { + print join("\t", @$first[3 .. @$first-1]), "\n"; + } elsif ($is_print) { + print STDERR join("\t", substr("UQdDWGgsiX", $first->[1], 1), @$first[3 .. @$first-1]), "\n"; } - # print the last few lines if applicable - map { print $_->{L} if ($_->{F}) } @last; } +# +# pileup2fq +# + sub pileup2fq { - my %opts = (d=>3, D=>255, Q=>25, G=>50, l=>10); + my %opts = (d=>3, D=>255, Q=>25, G=>25, l=>10); getopts('d:D:Q:G:l:', \%opts); die(qq/ Usage: samtools.pl pileup2fq [options] @@ -150,7 +191,7 @@ Options: -d INT minimum depth [$opts{d}] -D INT maximum depth [$opts{D}] -Q INT min RMS mapQ [$opts{Q}] -G INT minimum indel score [$opts{G}] - -l INT indel filter winsize [$opts{l}] + -l INT indel filter winsize [$opts{l}]\n /) if (@ARGV == 0 && -t STDIN); my ($last_chr, $seq, $qual, @gaps, $last_pos); @@ -208,6 +249,268 @@ sub p2q_print_str { } } +# +# sam2fq +# + +sub sam2fq { + my %opts = (n=>20, p=>''); + getopts('n:p:', \%opts); + die("Usage: samtools.pl sam2fq [-n 20] [-p ] \n") if (@ARGV == 0 && -t STDIN); + if ($opts{p} && $opts{n} > 1) { + my $pre = $opts{p}; + my @fh; + for (0 .. $opts{n}-1) { + open($fh[$_], sprintf("| gzip > $pre.%.3d.fq.gz", $_)) || die; + } + my $i = 0; + while (<>) { + next if (/^@/); + chomp; + my @t = split("\t"); + next if ($t[9] eq '*'); + my ($name, $seq, $qual); + if ($t[1] & 16) { # reverse strand + $seq = reverse($t[9]); + $qual = reverse($t[10]); + $seq =~ tr/ACGTacgt/TGCAtgca/; + } else { + ($seq, $qual) = @t[9,10]; + } + $name = $t[0]; + $name .= "/1" if ($t[1] & 0x40); + $name .= "/2" if ($t[1] & 0x80); + print {$fh[$i]} "\@$name\n$seq\n"; + if ($qual ne '*') { + print {$fh[$i]} "+\n$qual\n"; + } + $i = 0 if (++$i == $opts{n}); + } + close($fh[$_]) for (0 .. $opts{n}-1); + } else { + die("To be implemented.\n"); + } +} + +# +# sra2hdr +# + +# This subroutine does not use an XML parser. It requires that the SRA +# XML files are properly formated. +sub sra2hdr { + my %opts = (); + getopts('', \%opts); + die("Usage: samtools.pl sra2hdr \n") if (@ARGV == 0); + my $pre = $ARGV[0]; + my $fh; + # read sample + my $sample = 'UNKNOWN'; + open($fh, "$pre.sample.xml") || die; + while (<$fh>) { + $sample = $1 if (/) { + if (/\s*(\S+)\s*<\/LIBRARY_NAME>/i) { + $exp2lib{$exp} = $1; + } + } + close($fh); + # read run + my ($run, @fn); + open($fh, "$pre.run.xml") || die; + while (<$fh>) { + if (//i) { + if (@fn == 1) { + print STDERR "$fn[0]\t$run\n"; + } else { + for (0 .. $#fn) { + print STDERR "$fn[$_]\t$run", "_", $_+1, "\n"; + } + } + } + } + close($fh); +} + +# +# unique +# + +sub unique { + my %opts = (f=>250.0, q=>5, r=>2, a=>1, b=>3); + getopts('Qf:q:r:a:b:m', \%opts); + die("Usage: samtools.pl unique [-f $opts{f}] \n") if (@ARGV == 0 && -t STDIN); + my $last = ''; + my $recal_Q = !defined($opts{Q}); + my $multi_only = defined($opts{m}); + my @a; + while (<>) { + my $score = -1; + print $_ if (/^\@/); + $score = $1 if (/AS:i:(\d+)/); + my @t = split("\t"); + next if (@t < 11); + if ($score < 0) { # AS tag is unavailable + my $cigar = $t[5]; + my ($mm, $go, $ge) = (0, 0, 0); + $cigar =~ s/(\d+)[ID]/++$go,$ge+=$1/eg; + $cigar = $t[5]; + $cigar =~ s/(\d+)M/$mm+=$1/eg; + $score = $mm * $opts{a} - $go * $opts{q} - $ge * $opts{r}; # no mismatches... + } + $score = 1 if ($score < 1); + if ($t[0] ne $last) { + &unique_aux(\@a, $opts{f}, $recal_Q, $multi_only) if (@a); + $last = $t[0]; + } + push(@a, [$score, \@t]); + } + &unique_aux(\@a, $opts{f}, $recal_Q, $multi_only) if (@a); +} + +sub unique_aux { + my ($a, $fac, $is_recal, $multi_only) = @_; + my ($max, $max2, $max_i) = (0, 0, -1); + for (my $i = 0; $i < @$a; ++$i) { + if ($a->[$i][0] > $max) { + $max2 = $max; $max = $a->[$i][0]; $max_i = $i; + } elsif ($a->[$i][0] > $max2) { + $max2 = $a->[$i][0]; + } + } + if ($is_recal) { + if (!$multi_only || @$a > 1) { + my $q = int($fac * ($max - $max2) / $max + .499); + $q = 250 if ($q > 250); + $a->[$max_i][1][4] = $q < 250? $q : 250; + } + } + print join("\t", @{$a->[$max_i][1]}); + @$a = (); +} + +# +# uniqcmp: compare two SAM files +# + +sub uniqcmp { + my %opts = (q=>10, s=>100); + getopts('pq:s:', \%opts); + die("Usage: samtools.pl uniqcmp \n") if (@ARGV < 2); + my ($fh, %a); + warn("[uniqcmp] read the first file...\n"); + &uniqcmp_aux($ARGV[0], \%a, 0); + warn("[uniqcmp] read the second file...\n"); + &uniqcmp_aux($ARGV[1], \%a, 1); + warn("[uniqcmp] stats...\n"); + my @cnt; + $cnt[$_] = 0 for (0..9); + for my $x (keys %a) { + my $p = $a{$x}; + my $z; + if (defined($p->[0]) && defined($p->[1])) { + $z = ($p->[0][0] == $p->[1][0] && $p->[0][1] eq $p->[1][1] && abs($p->[0][2] - $p->[1][2]) < $opts{s})? 0 : 1; + if ($p->[0][3] >= $opts{q} && $p->[1][3] >= $opts{q}) { + ++$cnt[$z*3+0]; + } elsif ($p->[0][3] >= $opts{q}) { + ++$cnt[$z*3+1]; + } elsif ($p->[1][3] >= $opts{q}) { + ++$cnt[$z*3+2]; + } + print STDERR "$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", + $p->[0][5]-$p->[1][5], "\n" if ($z && defined($opts{p}) && ($p->[0][3] >= $opts{q} || $p->[1][3] >= $opts{q})); + } elsif (defined($p->[0])) { + ++$cnt[$p->[0][3]>=$opts{q}? 6 : 7]; + print STDERR "$x\t$p->[0][1]:$p->[0][2]\t$p->[0][3]\t$p->[0][4]\t*\t0\t*\t", + $p->[0][5], "\n" if (defined($opts{p}) && $p->[0][3] >= $opts{q}); + } else { + print STDERR "$x\t*\t0\t*\t$p->[1][1]:$p->[1][2]\t$p->[1][3]\t$p->[1][4]\t", + -$p->[1][5], "\n" if (defined($opts{p}) && $p->[1][3] >= $opts{q}); + ++$cnt[$p->[1][3]>=$opts{q}? 8 : 9]; + } + } + print "Consistent (high, high): $cnt[0]\n"; + print "Consistent (high, low ): $cnt[1]\n"; + print "Consistent (low , high): $cnt[2]\n"; + print "Inconsistent (high, high): $cnt[3]\n"; + print "Inconsistent (high, low ): $cnt[4]\n"; + print "Inconsistent (low , high): $cnt[5]\n"; + print "Second missing (high): $cnt[6]\n"; + print "Second missing (low ): $cnt[7]\n"; + print "First missing (high): $cnt[8]\n"; + print "First missing (low ): $cnt[9]\n"; +} + +sub uniqcmp_aux { + my ($fn, $a, $which) = @_; + my $fh; + $fn = "samtools view $fn |" if ($fn =~ /\.bam/); + open($fh, $fn) || die; + while (<$fh>) { + my @t = split; + next if (@t < 11); +# my $l = ($t[5] =~ /^(\d+)S/)? $1 : 0; + my $l = 0; + my ($x, $nm) = (0, 0); + $nm = $1 if (/NM:i:(\d+)/); + $_ = $t[5]; + s/(\d+)[MI]/$x+=$1/eg; + @{$a->{$t[0]}[$which]} = (($t[1]&0x10)? 1 : 0, $t[2], $t[3]-$l, $t[4], "$x:$nm", $x - 4 * $nm); + } + close($fh); +} + +sub plp2vcf { + while (<>) { + my @t = split; + next if ($t[3] eq '*/*'); + if ($t[2] eq '*') { # indel + my @s = split("/", $t[3]); + my (@a, @b); + my ($ref, $alt); + for (@s) { + next if ($_ eq '*'); + if (/^-/) { + push(@a, 'N'.substr($_, 1)); + push(@b, 'N'); + } elsif (/^\+/) { + push(@a, 'N'); + push(@b, 'N'.substr($_, 1)); + } + } + if ($a[0] && $a[1]) { + if (length($a[0]) < length($a[1])) { + $ref = $a[1]; + $alt = ($b[0] . ('N' x (length($a[1]) - length($a[0])))) . ",$b[1]"; + } elsif (length($a[0]) > length($a[1])) { + $ref = $a[0]; + $alt = ($b[1] . ('N' x (length($a[0]) - length($a[1])))) . ",$b[0]"; + } else { + $ref = $a[0]; + $alt = ($b[0] eq $b[1])? $b[0] : "$b[0],$b[1]"; + } + } else { + $ref = $a[0]; $alt = $b[0]; + } + print join("\t", @t[0,1], '.', $ref, $alt, $t[5], '.', '.'), "\n"; + } else { # SNP + } + } +} + # # Usage # @@ -218,8 +521,7 @@ Program: samtools.pl (helper script for SAMtools) Version: $version Contact: Heng Li \n Usage: samtools.pl []\n -Command: indelFilter filter indels generated by `pileup -c' - snpFilter filter SNPs generated by `pileup -c' +Command: varFilter filtering SNPs and short indels pileup2fq generate fastq from `pileup -c' showALEN print alignment length (ALEN) following CIGAR \n/);