pipeline/script/filter_fusion.pl

#!/usr/bin/perl
use strict;
use warnings;

die "usage:perl $0 depth_file input out project" unless @ARGV == 4;
my ($depth_file, $input, $out, $project) = @ARGV;

my $public_path = defined $ENV{'PUBLIC'} ? $ENV{'PUBLIC'} : "/dataseq/jmdna/codes/public/";

print "Fusion过滤使用public路径：$public_path\n";

open IN, "$input";
open OUT1, "> $out";

open SD, "$public_path/gene_strand.txt";
my %strand;
while (<SD>) {
    chomp;
    my @line = split;
    $strand{$line[1]} = $line[0];
}

my $fusion = info();
my @fusions = @$fusion;

while (<IN>) {
    if (/^##/) {
        next;
    }
    if (/^#CHROM/) {
        chomp;
        print OUT1 "$_\tCHROM2\tPOS2\tINFO2\tGENE1\tGENE2\tFUSION\tFREQ1\tFREQ2\tref_gene\n";
        next;
    }
    chomp;
    my @line = split(/\t/);
    $line[7] =~ /SVTYPE=(.*?);/;
    my $SVTYPE = $1;

    if ($SVTYPE eq 'INV') {
        $line[7] =~ /SVLEN=(.*?);/;
        if ($1 > 1000) {
            my $freq_inv = sprintf("%.2f", (split(/:/, $line[9]))[1] / (split(/:/, $line[9]))[7] * 100);
            next if $freq_inv < 0.3;
            my $pt_inv = $_;
            my $gene1 = &site2gene($line[0], $line[1]);
            $line[7] =~ /END=(.*?);/;
            my $end = $1;
            my $gene2 = &site2gene($line[0], $end);

            # if (!((grep {$_ eq $gene1} @fusions) or (grep {$_ eq $gene2} @fusions))) {
            #     next;
            # }
            my $ref_gene;
            if (grep {$_ eq $gene1} @fusions) {
                $ref_gene = $gene1;
            }
            if (grep {$_ eq $gene2} @fusions) {
                $ref_gene = $gene2;
            }
            if (!(defined $ref_gene)) {
                next;
            }

            my $depth = &depth($line[0], $line[1], $line[0], $end);
            my $freq_inv2;
            if (defined $depth) {$freq_inv2 = sprintf("%.2f", (split(/:/, $line[9]))[1] / $depth * 100)}
            else {$freq_inv2 = "-"};
            print OUT1 "$pt_inv\t$line[0]\t$end\t-\t$gene1\t$gene2\t$gene1-$gene2\t$freq_inv\t$freq_inv2\t$ref_gene\n";
            print OUT1 "$pt_inv\t$line[0]\t$end\t-\t$gene2\t$gene1\t$gene2-$gene1\t$freq_inv\t$freq_inv2\t$ref_gene\n";
            next;
        }
    }
    elsif ($SVTYPE eq 'BND') {
        if ($line[2] =~ /(\d+)_1/) {
            $line[7] =~ /EVENT=(.*?);\S+Func.refGene=(.*?);Gene.refGene=(.*?);/;
            my ($event, $region1, $gene1) = ($1, $2, $3);
            my $pt = "$_";
            my $mate = <IN>;
            chomp $mate;
            my @line2 = split(/\t/, $mate);
            $line2[7] =~ /EVENT=(.*?);\S+Func.refGene=(.*?);Gene.refGene=(.*?);/;
            my ($event2, $region2, $gene2) = ($1, $2, $3);
            $gene1 = (split /\\x3b/, $gene1)[0];
            $gene2 = (split /\\x3b/, $gene2)[0];

            # if (!(grep {$_ eq $gene1} @fusions or grep {$_ eq $gene2} @fusions)) {
            #     next;
            # }
            my $ref_gene;
            if (grep {$_ eq $gene1} @fusions) {
                $ref_gene = $gene1;
            }
            if (grep {$_ eq $gene2} @fusions) {
                $ref_gene = $gene2;
            }
            if (!(defined $ref_gene)) {
                next;
            }
            if ($event eq $event2 and $line2[2] =~ /_2/ and $gene1 ne $gene2 and $region1 !~ /upstream|downstream|intergenic|UTR/ and $region2 !~ /upstream|downstream|intergenic|UTR/) {
                my $freq = sprintf("%.2f", (split(/:/, $line[9]))[1] / (split(/:/, $line[9]))[7] * 100);
                next if $freq < 0.3;
                my $strand1 = &gene2strand($gene1);
                my $strand2 = &gene2strand($gene2);
                my $depth = &depth($line[0], $line[1], $line2[0], $line2[1]);
                my $freq_bnd;
                if (defined $depth) {$freq_bnd = sprintf("%.2f", (split(/:/, $line[9]))[1] / $depth * 100)}
                else {$freq_bnd = "-"};
                if ($line[4] =~ /^N\[/) {
                    if ($strand1 eq '+' and $strand2 eq '+') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene1\t$gene2\t$gene1-$gene2\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                    elsif ($strand1 eq '-' and $strand2 eq '-') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene2\t$gene1\t$gene2-$gene1\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                    else {
                        print($gene1, '.....\n')
                    }
                }
                elsif ($line[4] =~ /^N\]/) {
                    if ($strand1 eq '+' and $strand2 eq '-') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene1\t$gene2\t$gene1-$gene2\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                    elsif ($strand1 eq '-' and $strand2 eq '+') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene2\t$gene1\t$gene2-$gene1\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                }
                elsif ($line[4] =~ /\]N$/) {
                    if ($strand1 eq '+' and $strand2 eq '+') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene2\t$gene1\t$gene2-$gene1\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                    elsif ($strand1 eq '-' and $strand2 eq '-') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene1\t$gene2\t$gene1-$gene2\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                }
                elsif ($line[4] =~ /\[N$/) {
                    if ($strand1 eq '+' and $strand2 eq '-') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene2\t$gene1\t$gene2-$gene1\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                    elsif ($strand1 eq '-' and $strand2 eq '+') {
                        print OUT1 "$pt\t$line2[0]\t$line2[1]\t$line2[7]\t$gene1\t$gene2\t$gene1-$gene2\t$freq\t$freq_bnd\t$ref_gene\n";
                    }
                }
                else {
                    print($gene1 'NNN.......\n')
                }
            }
        }
    }
}

sub site2gene {
    my ($chr, $pos) = @_;
    my %gene;
    open REFGENE, "$public_path/hg19_refGene.txt";
    my @gene;
    while (<REFGENE>) {
        my @line = split(/\t/, $_);
        if ($chr eq $line[2] and $pos >= $line[4] and $pos <= $line[5]) {
            $gene{$line[-4]}++;
            push @gene, $line[-4] if $gene{$line[-4]} < 2;
        }
    }
    return join("|", @gene);
}

sub depth {
    my ($chr1, $pos1, $chr2, $pos2) = @_;
    my $depth1 = `less $depth_file|grep -w $chr1|grep -w $pos1`;
    my $depth2 = `less $depth_file|grep -w $chr2|grep -w $pos2`;
    $depth1 = (split(/\t/, $depth1))[3] if defined $depth1;
    $depth2 = (split(/\t/, $depth2))[3] if defined $depth2;
    my $depth;
    if ($depth1 and $depth2) {
        $depth = ($depth1 + $depth2) / 2
    }
    elsif ($depth1 and !$depth2) {
        $depth = $depth1;
    }
    elsif (!$depth1 and $depth2) {
        $depth = $depth2;
    }
    return $depth;
}

sub gene2strand {
    my $gene = shift @_;
    return $strand{$gene};
}

sub info {
    open INFO, "$public_path/info.csv";
    # 读取并解析表头
    my $header = <INFO>;
    chomp($header);
    my @column_names = split(',', $header);
    my (@fusion);
    while (<INFO>) {
        chomp;
        my @line = split(/,/, $_);
        # 将数据与表头对应
        my %record;
        @record{@column_names} = @line;
        if ($record{'project'} eq $project) {
            if ($record{'fusion'} ne "NA") {
                @fusion = split(/\//, $record{'fusion'});
            }
        }
    }
    return \@fusion
}