Arrays 将包含在数字范围内的哈希键分组

我有一个数据集，包含了几种不同方法的组合尝试 （

接近1到3

）以确定基因组中的位置：

source  chromosome1 bp1 chromosome2 bp2
attempt1    2L  5890205 2L  5890720
attempt2    2L  5890205 2L  5890721
attempt1    2L  22220720    2L  22255744
attempt1    3L  15568694    3L  15568866
attempt3    3R  14006279    3R  14008254
attempt1    3R  14006281    3R  14008253
attempt2    3R  14006282    3R  14008254
attempt3    3R  14006286    3R  14008254
attempt1    3R  32060908    3R  32061196
attempt1    3R  32066206    3R  32068392
attempt3    3R  32066206    3R  32068392
attempt2    3R  32066207    3R  32068393
attempt2    X   4574312 X   4576608
attempt1    X   4574313 X   4576607
attempt3    X   4574313 X   4576608

我希望找到并分组每次尝试都已确定的位置，为错误留出一点空间。例如，我想把前两行

source  chromosome1 bp1 chromosome2 bp2
attempt1    2L  5890205 2L  5890720
attempt2    2L  5890205 2L  5890721

…作为单个事件（

event 1

），已通过两次不同的尝试（

attempt1

和

attempt2

）识别。我只想在以下情况下将此类实例归类为单个事件：

• 就

  bp1

  +/-

  5的位置达成一致（即在窗口

  5890200..5890210

  内）
• 识别相同的

  chromosome1

  和

  chromosome2

  （

  2L

  ）
• 就bp2的位置达成一致（即在窗口内

  5890715..5890725

  ）

我尝试使用每个染色体和bp作为散列中的一个单独的键来实现这一点

my %SVs;
my $header;

# Make hash 
while(<$in>){
  chomp;
  if ($. == 1){
      $header = $_;
      next;
  }
  my ($source, $chromosome1, $bp1, $chromosome2, $bp2) = split;

  push @{$SVs{$chromosome1}{$bp1}{$chromosome2}{$bp2}}, $_;

  }
}

---

下面通过添加到等价类的最后一个条目定义了每个等价类（基于我对您上述评论的理解）：

---

如果您有三次尝试，并且#1和#2同意，以及#2和#3同意（但是#1和#3不同意），该怎么办。那么，第二次尝试属于哪一个事件呢？好问题。这对我来说意味着它们应该被分组，然后我想扩展我的窗口（例如，到+/-10），以允许这项工作非常出色。你能解释一下

，除非（@{$events[-1]}）

正在做什么吗？@fugu对代码进行了一些简化，以摆脱这种检查。原则上，您还应该检查读取标题和第一行数据是否成功，但我猜输入在某种程度上是由您控制的。

my %events;
for my $chr1 ( sort keys %SVs ){
  for my $bp1 ( sort { $a <=> $b } keys $SVs{$chr1} ){
    my $w1_start = ( $bp1 - 5 );
    my $w1_end = ( $bp1 + 5 );
    my $window1 = "$w1_start-$w1_end";

    for my $chr2 ( sort keys $SVs{$chr1}{$bp1} ){
      for my $bp2 ( sort { $a <=> $b } keys $SVs{$chr1}{$bp1}{$chr2} ){

        my $w2_start = ( $bp2 - 5 );
        my $w2_end = ( $bp2 + 5 );
        my $window2 = "$w2_start-$w2_end";

        for ( $w1_start .. $w1_end ){
          if ($bp1 == $_){
            push @{$events{$chr1}{$window1}}, @{$SVs{$chr1}{$bp1}{$chr2}{$bp2}};
          }
        }

        for ( $w2_start .. $w2_end ){
          if ($bp2 == $_){
            push @{$events{$chr2}{$window2}}, @{$SVs{$chr1}{$bp1}{$chr2}{$bp2}};
          }
        }

      }
    }
  }
}

print Dumper \%events;

event   source  chromosome1 bp1 chromosome2 bp2
1   attempt1    2L  5890205 2L  5890720
1   attempt2    2L  5890205 2L  5890721
2   attempt1    2L  22220720    2L  22255744
3   attempt1    3L  15568694    3L  15568866
4   attempt3    3R  14006279    3R  14008254
4   attempt1    3R  14006281    3R  14008253
4   attempt2    3R  14006282    3R  14008254
4   attempt3    3R  14006286    3R  14008254
5   attempt1    3R  32060908    3R  32061196
6   attempt1    3R  32066206    3R  32068392
6   attempt3    3R  32066206    3R  32068392
6   attempt2    3R  32066207    3R  32068393
7   attempt2    X   4574312 X   4576608
7   attempt1    X   4574313 X   4576607
7   attempt3    X   4574313 X   4576608 

#!/usr/bin/env perl

use strict;
use warnings;

run(\*DATA);

sub run {
    my $fh = shift;
    my @header = split ' ', scalar <$fh>;

    my @events = ([ get_next_event($fh, \@header)]);

    while (my $event = get_next_event($fh, \@header)) {
        # change the -1 in the second subscript to 0
        # if you want to always compare to the first
        # event added to the equivalence class
        if (same_event($events[-1][-1], $event, 5)) {
            push @{ $events[-1] }, $event;
            next;
        }

        push @events, [ $event ];
    }

    print join("\t", event => @header), "\n";
    for my $i (1 .. @events) {
        for my $ev (@{ $events[$i - 1] }) {
            print join("\t", $i, @{$ev}{@header}), "\n";
        }
    }
}

sub get_next_event {
    my $fh = shift;
    my $header = shift;
    return unless defined(my $line = <$fh>);
    return unless $line =~ /\S/;

    my %event;
    @event{ @$header } = split ' ', $line;
    return \%event;
}

sub same_event {
    my ($x, $y, $threshold) = @_;

    return if $x->{chromosome1} ne $y->{chromosome1};
    return if abs($x->{bp1} - $y->{bp1}) > $threshold;
    return if abs($x->{bp2} - $y->{bp2}) > $threshold;
    return 1;
}

__DATA__
source  chromosome1 bp1 chromosome2 bp2
attempt1    2L  5890205 2L  5890720
attempt2    2L  5890205 2L  5890721
attempt1    2L  22220720    2L  22255744
attempt1    3L  15568694    3L  15568866
attempt3    3R  14006279    3R  14008254
attempt1    3R  14006281    3R  14008253
attempt2    3R  14006282    3R  14008254
attempt3    3R  14006286    3R  14008254
attempt1    3R  32060908    3R  32061196
attempt1    3R  32066206    3R  32068392
attempt3    3R  32066206    3R  32068392
attempt2    3R  32066207    3R  32068393
attempt2    X   4574312 X   4576608
attempt1    X   4574313 X   4576607
attempt3    X   4574313 X   4576608