Rather than using manifest variables for accumulating results you can use a hash keyed by chromosome:

use strict;
use warnings;

my %chroms;

# Count frequency of start sited in the column-2.
while (<DATA>) {
    next if /^#/;    #ignore comments in the input file
    chomp(my $line = $_);
    my ($chr, $start, $end, $strand) = (split /\t/, $line)[0, 1, 2, 5]
+;
    my $startPoint;

    if ($strand eq '+') {
        $startPoint = $start;
        $chroms{$chr}{startPoints_plusStrand}{$startPoint}++;
    } elsif ($strand eq '-') {
        $startPoint = $end;
        $chroms{$chr}{startPoints_minusStrand}{$startPoint}++;
    } else {
        die "Goofy BED Data on line $.:\n$_";
    }
}

print "#Plus Strand BedGraph:\n";

for my $chromKey (keys %chroms) {
    my $chrom = $chroms{$chromKey};

    print "--- $chromKey\n";
    for my $key (sort {$a <=> $b} keys %{$chrom->{startPoints_plusStra
+nd}}) {
        print $key + 1, "\t", "+", "$chrom->{startPoints_plusStrand}{$
+key}\n";
    }
}

__DATA__
#chromosome    start    end    position    col-5    strand
chr7    127471196    127472363    Pos1    0    +
chr7    127471196    127472363    Pos1    0    +
chr7    127471196    127472363    Pos1    0    +
chr7    127471196    127472363    Pos1    0    +
chr7    127471196    127472363    Pos1    0    +
chr7    127472363    127473530    Pos2    0    +
chr7    127473530    127474697    Pos3    0    -
chr7    127474697    127475864    Pos4    0    -
chr8    40000    40500    Pos5    0    +
chr8    40000    40600    Pos5    0    +
chr8    40000    40650    Pos5    0    +
chr8    41000    41200    Pos6    0    -
[download]