Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Dear all, I have hundreds of pathways that I want to compare. Some of the pathways have overlapping nodes. I am interested in obtaining node pairs in the shortest pathway for each.

pwy nodes
A a b c d e f
B a b c

[download]

The shortest pathway that abc appear in is pwy B, so I want to output abc as pairwise. Similarily, the shortest pathway that def appear in is A.

desired output

de A
ef A
ab B
ac B
bc B

[download]

I have tried all sorts of crazy moves with hashes, but now I think that maybe I should read all of the nodes into a hash of pathways and then loop through comparing all of the arrays. I know that right now I am just printing out the next door neighbor nodes,but I also think there must be a better way to do this. thanks

example code 

my $in=$ARGV[0] || "pathways.col";
open (IN,$in) or die "cannot open $in\n";
my %HoCplx2ID; my %HoPwyPair;
while (my $lines=<IN>){
        next if ($lines =~/^#/);
        next if ($lines =~/^UNIQUE-ID/);
        chomp $lines;
        my @cols=split(/\t/,$lines);
        my $cmplxID=$cols[0];
        #print $cmplxID."\n";
        my $cmplxNm=$cols[1];
        my @restCols=@cols[2..$#cols];
        my @cycIDs=grep(/^GCXG-/, @restCols);
        @cycIDs=grep($_ ne '',@cycIDs);
        print "cycIDs array\n";
        print Dumper(@cycIDs);
        my $pwySize=scalar(@cycIDs);
        push (@{$HoCplx2ID{$cmplxID}},@cycIDs);
        for (my $i=0; $i < ($pwySize-1); $i++){
                my $pair =join("\t",$cycIDs[$i],$cycIDs[$i+1]);
                $HoPwyPair{$pair}{$cmplxID}=$pwySize;
        }

}
close(IN);
########## print out pairwise with PA01 locusIDs ######
my $org=$ARGV[1]|| "PA01";
my $outfile="$in.$org.pairwise.nxtNeighb.tab";
#open (OUT,">",$outfile);
### step 1 for each pair find smallest pathway
my %HoSmPwy;
foreach my $pair (keys %HoPwyPair){
        $HoSmPwy{$pair}=100;
        foreach my $pwy (keys %{$HoPwyPair{$pair}}){
                if ($HoPwyPair{$pair}{$pwy} < $HoSmPwy{$pair}) {
                        $HoSmPwy{$pair}=$HoPwyPair{$pair}{$pwy};
                }
        }
}
print "hash of smallest pathways\n";
#print Dumper(%HoSmPwy);
### step 2 for each pathway, look at each pair if that pwy size = smal
+lest pathway , then print ##
print "output\n";
foreach my $pwy (keys(%HoCplx2ID)){
        my @units=@{$HoCplx2ID{$pwy}};
        my $pwySize=scalar(@units);
        for (my $i=0; $i < ($pwySize-1); $i++){
        my $pair =join("\t",$units[$i],$units[$i+1]);
                if ($pwySize = $HoSmPwy{$pair}) {
                #       print $pair."\n";
                }
        }
}

####

[download]

Replies are listed 'Best First'.
Re: array comparisons
by McA (Priest) on Oct 27, 2014 at 16:06 UTC

    Hi,

    hopefully this helps to get your mind around your problem:

    #!/usr/bin/perl
use strict;
use warnings;
use 5.010;

use Data::Dumper;

my %neigbours;

while (my $line = <DATA>) {
    chomp $line;
    my ($id_pathway, @pathnodes) = split ' ', $line;
    my $path_length = @pathnodes;
    if($path_length < 2) {
        warn "Don't know whether paths with less than two nodes are al
+lowed. Skipping.";
        next;
    }
    for (my $i = 0; $i < $path_length - 1; $i++) {
        my $key = "$pathnodes[$i]$pathnodes[$i+1]";
        if(defined $neigbours{$key}) {
            if($neigbours{$key}->{'minpathlen'} >= $path_length) {
                $neigbours{$key}->{'minpathlen'} = $path_length;
                $neigbours{$key}->{'id_pathway'} = $id_pathway;
            }
        }
        else {
            $neigbours{$key}->{'minpathlen'} = $path_length;
            $neigbours{$key}->{'id_pathway'} = $id_pathway;
        }
    }
}

say Dumper(\%neigbours);

__DATA__
A a b c d e f
B a b c

    [download]

    Allow me an annotation: Please use meaningful variable names. Your code is hard to read.

    Regards
    McA

[reply]
[d/l]

Back to Seekers of Perl Wisdom