From cdd50dc98887876cabc7fd6dce802971149e6424 Mon Sep 17 00:00:00 2001 From: Heng Li Date: Mon, 8 Jun 2009 13:14:26 +0000 Subject: [PATCH] * samtools.pl-0.3.0 * combine snpFilter and indelFilter --- misc/samtools.pl | 203 ++++++++++++++++++++++++----------------------- 1 file changed, 104 insertions(+), 99 deletions(-) diff --git a/misc/samtools.pl b/misc/samtools.pl index 3dc6fe7..e27479c 100755 --- a/misc/samtools.pl +++ b/misc/samtools.pl @@ -6,11 +6,11 @@ use strict; use warnings; use Getopt::Std; -my $version = '0.2.3'; +my $version = '0.3.0'; &usage if (@ARGV < 1); my $command = shift(@ARGV); -my %func = (snpFilter=>\&snpFilter, indelFilter=>\&indelFilter, showALEN=>\&showALEN, pileup2fq=>\&pileup2fq); +my %func = (showALEN=>\&showALEN, pileup2fq=>\&pileup2fq, varFilter=>\&varFilter); die("Unknown command \"$command\".\n") if (!defined($func{$command})); &{$func{$command}}; @@ -32,114 +32,120 @@ sub showALEN { } # -# 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 - +sub varFilter { + my %opts = (d=>3, D=>100, l=>30, Q=>25, G=>20, s=>100, w=>10, W=>10, N=>2); + getopts('d:D:l:Q:w:W:N:G:', \%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}] -\n/) if (@ARGV == 0 && -t STDIN); - - my (@arr1, @arr2); - my ($curr, $last) = (\@arr1, \@arr2); - my $is_ref = defined($opts{r})? 1 : 0; +Usage: samtools.pl varFilter [options] + +Options: -Q INT minimum RMS mapping quality [$opts{Q}] + -d INT minimum read depth [$opts{d}] + -D INT maximum read depth [$opts{D}] + + -G INT min indel score for nearby SNP filtering [$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 [$opts{W}] + + -l INT window size for filtering adjacent gaps [$opts{l}]\n +/) if (@ARGV == 0 && -t STDIN); + + # 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 $last->[3]; + next if ($t[2] eq $t[3] || $t[3] eq '*/*'); # skip non-var sites + # clear the out-of-range elements + while (@staging) { + if ($staging[0][2] ne $t[0] || $staging[0][3] + $max_dist < $t[1]) { + varFilter_aux(shift @staging); + } else { + last; } } - if ($do_swap) { - my $tmp = $curr; $curr = $last; $last = $tmp; + my ($flt, $score) = (0, -1); + # first a simple filter + if ($t[6] < $opts{Q}) { + $flt = 1; + } elsif ($t[7] < $opts{d}) { + $flt = 2; + } elsif ($t[7] > $opts{D}) { + $flt = 3; } - } - print $last->[3] if (defined $last->[0]); -} - -# -# snpFilter -# - -sub snpFilter { - my %opts = (f=>'', Q=>40, d=>3, w=>10, D=>0, N=>2, W=>10, q=>20, s=>50); - getopts('f:s:w:q:Q:d:D:W:N:', \%opts); - die(qq{ -Usage: samtools.pl snpFilter [options] - -Options: -d INT minimum depth to call a SNP [$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],$x} = 1; + # site dependent filters + if ($flt == 0) { + if ($t[2] eq '*') { # an indel + # filtering SNPs + if ($t[5] >= $opts{G}) { + for my $x (@staging) { + next if ($x->[0] >= 0 || $x->[3] + $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 $x (@staging) { + next if ($x->[0] < 0 || $x->[3] + $ol < $t[1]); + if ($x->[0] < $score) { + $x->[1] = 6; + } else { + $flt = 6; last; + } + } + } else { # a SNP + # check adjacent SNPs + my $k = 1; + for my $x (@staging) { + ++$k if ($x->[0] < 0 && $x->[3] + $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->[3] + $oW >= $t[1] && $x->[1] == 0); + } + } else { # then check gap filter + for my $x (@staging) { + next if ($x->[0] < 0 || $x->[3] + $ow < $t[1]); + if ($x->[0] >= $opts{G}) { + $flt = 5; last; + } + } + } } } - close($fh); + push(@staging, [$score, $flt, @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); } - # print the last few lines if applicable - map { print $_->{L} if ($_->{F}) } @last; } +sub varFilter_aux { + my $first = shift; + if ($first->[1] == 0) { + print join("\t", @$first[2 .. @$first-1]), "\n"; + } else { + print STDERR join("\t", substr("UQdDWGgX", $first->[1], 1), @$first[2 .. @$first-1]), "\n"; + } +} + +# +# pileup2fq +# + sub pileup2fq { my %opts = (d=>3, D=>255, Q=>25, G=>50, l=>10); getopts('d:D:Q:G:l:', \%opts); @@ -150,7 +156,7 @@ Options: -d INT minimum depth [$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); @@ -218,8 +224,7 @@ 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/); -- 2.39.2