X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=samtools.git;a=blobdiff_plain;f=misc%2Fsamtools.pl;h=d03c1c7f148dcca835946e7ccd9ee52f3b6dc00b;hp=86b285c5fa6c144a1d60ed61cb48a23bde7f134f;hb=HEAD;hpb=d363084f0412f3bcdeb0304aeb0974c9a10c7649 diff --git a/misc/samtools.pl b/misc/samtools.pl index 86b285c..d03c1c7 100755 --- a/misc/samtools.pl +++ b/misc/samtools.pl @@ -10,8 +10,8 @@ my $version = '0.3.3'; &usage if (@ARGV < 1); my $command = shift(@ARGV); -my %func = (showALEN=>\&showALEN, pileup2fq=>\&pileup2fq, varFilter=>\&varFilter, - unique=>\&unique, uniqcmp=>\&uniqcmp); +my %func = (showALEN=>\&showALEN, pileup2fq=>\&pileup2fq, varFilter=>\&varFilter, plp2vcf=>\&plp2vcf, + unique=>\&unique, uniqcmp=>\&uniqcmp, sra2hdr=>\&sra2hdr, sam2fq=>\&sam2fq); die("Unknown command \"$command\".\n") if (!defined($func{$command})); &{$func{$command}}; @@ -37,9 +37,21 @@ sub showALEN { # varFilter # +# +# 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); - getopts('pq:d:D:l:Q:w:W:N:G:', \%opts); + 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 varFilter [options] @@ -47,6 +59,8 @@ 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}] @@ -70,7 +84,8 @@ Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] next if (uc($t[2]) eq uc($t[3]) || $t[3] eq '*/*'); # skip non-var sites # clear the out-of-range elements while (@staging) { - last if ($staging[0][2] eq $t[0] && $staging[0][3] + $max_dist >= $t[1]); + # Still on the same chromosome and the first element's window still affects this position? + last if ($staging[0][3] eq $t[0] && $staging[0][4] + $staging[0][2] + $max_dist >= $t[1]); varFilter_aux(shift(@staging), $opts{p}); # calling a function is a bit slower, not much } my ($flt, $score) = (0, -1); @@ -80,14 +95,31 @@ Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] } elsif ($t[7] > $opts{D}) { $flt = 3; } + if ($t[2] eq '*') { # an indel + if ($opts{i} && $opts{i}>$t[5]) { $flt = 8; } + } + elsif ($opts{S} && $opts{S}>$t[5]) { $flt = 7; } # SNP + # 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 $x (@staging) { - next if ($x->[0] >= 0 || $x->[3] + $ow < $t[1]); + # Is it a SNP and is it outside the SNP filter window? + next if ($x->[0] >= 0 || $x->[4] + $x->[2] + $ow < $t[1]); $x->[1] = 5 if ($x->[1] == 0); } } @@ -97,7 +129,8 @@ Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] $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 < $t[1]); + # Is it a SNP and is it outside the gap filter window + next if ($x->[0] < 0 || $x->[4] + $x->[2] + $ol < $t[1]); if ($x->[0] < $score) { $x->[1] = 6; } else { @@ -109,17 +142,17 @@ Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] # 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 || $x->[1] == 5)); + ++$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->[3] + $oW >= $t[1] && $x->[1] == 0); + $x->[1] = 4 if ($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->[3] + $ow < $t[1]); + next if ($x->[0] < 0 || $x->[4] + $x->[2] + $ow < $t[1]); if ($x->[0] >= $opts{G}) { $flt = 5; last; } @@ -127,7 +160,7 @@ Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] } } } - push(@staging, [$score, $flt, @t]); + push(@staging, [$score, $flt, $len, @t]); } # output the last few elements in the staging list while (@staging) { @@ -138,9 +171,9 @@ Options: -Q INT minimum RMS mapping quality for SNPs [$opts{Q}] sub varFilter_aux { my ($first, $is_print) = @_; if ($first->[1] == 0) { - print join("\t", @$first[2 .. @$first-1]), "\n"; + print join("\t", @$first[3 .. @$first-1]), "\n"; } elsif ($is_print) { - print STDERR join("\t", substr("UQdDWGgX", $first->[1], 1), @$first[2 .. @$first-1]), "\n"; + print STDERR join("\t", substr("UQdDWGgsiX", $first->[1], 1), @$first[3 .. @$first-1]), "\n"; } } @@ -216,16 +249,113 @@ sub p2q_print_str { } } +# +# sam2fq +# + +sub sam2fq { + my %opts = (n=>20, p=>''); + getopts('n:p:', \%opts); + die("Usage: samtools.pl sam2fq [-n 20] [-p ] \n") if (@ARGV == 0 && -t STDIN); + if ($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:', \%opts); + 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; @@ -243,16 +373,16 @@ sub unique { } $score = 1 if ($score < 1); if ($t[0] ne $last) { - &unique_aux(\@a, $opts{f}, $recal_Q) if (@a); + &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) if (@a); + &unique_aux(\@a, $opts{f}, $recal_Q, $multi_only) if (@a); } sub unique_aux { - my ($a, $fac, $is_recal) = @_; + my ($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] > $max) { @@ -262,9 +392,11 @@ sub unique_aux { } } if ($is_recal) { - my $q = int($fac * ($max - $max2) / $max + .499); - $q = 250 if ($q > 250); - $a->[$max_i][1][4] = $q < 250? $q : 250; + if (!$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 = (); @@ -341,6 +473,44 @@ sub uniqcmp_aux { 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 #