line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
#!/usr/bin/env perl |
2
|
|
|
|
|
|
|
# ABSTRACT: Compute distance matrix for any distance metric |
3
|
|
|
|
|
|
|
|
4
|
|
|
|
|
|
|
package Algorithm::DistanceMatrix; |
5
|
|
|
|
|
|
|
BEGIN { |
6
|
1
|
|
|
1
|
|
47086
|
$Algorithm::DistanceMatrix::VERSION = '0.04'; |
7
|
|
|
|
|
|
|
} |
8
|
1
|
|
|
1
|
|
1811
|
use Moose; |
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
|
10
|
|
|
|
|
|
|
has 'mode' =>( |
11
|
|
|
|
|
|
|
is => 'rw', |
12
|
|
|
|
|
|
|
isa => 'Str', |
13
|
|
|
|
|
|
|
default => 'lower', |
14
|
|
|
|
|
|
|
); |
15
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
has 'metric' => ( |
18
|
|
|
|
|
|
|
is=>'rw', |
19
|
|
|
|
|
|
|
isa=>'CodeRef', |
20
|
|
|
|
|
|
|
default=>sub{abs($_[0]-$_[1])}, |
21
|
|
|
|
|
|
|
); |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
has 'objects' => ( |
25
|
|
|
|
|
|
|
is => 'rw', |
26
|
|
|
|
|
|
|
isa => 'ArrayRef', |
27
|
|
|
|
|
|
|
); |
28
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
|
30
|
|
|
|
|
|
|
sub distancematrix { |
31
|
|
|
|
|
|
|
my ($self, ) = @_; |
32
|
|
|
|
|
|
|
# Callback function |
33
|
|
|
|
|
|
|
my $metric = $self->metric; |
34
|
|
|
|
|
|
|
my $objects = $self->objects; |
35
|
|
|
|
|
|
|
my $n = @$objects; |
36
|
|
|
|
|
|
|
my $distances = []; |
37
|
|
|
|
|
|
|
for (my $i = 0; $i < $n; $i++) { |
38
|
|
|
|
|
|
|
# This initialization is required to prevent 'undef' at [0,0], |
39
|
|
|
|
|
|
|
$distances->[$i] ||= []; |
40
|
|
|
|
|
|
|
# Diagonal or full matrix? |
41
|
|
|
|
|
|
|
my $start = $self->mode =~ /full/i ? 0 : $i+1; |
42
|
|
|
|
|
|
|
for (my $j = $start; $j < $n; $j++) { |
43
|
|
|
|
|
|
|
# Use a pointer, then determine if it's row-major or col-major order |
44
|
|
|
|
|
|
|
# Swap i and j if lower diagonal (default) |
45
|
|
|
|
|
|
|
my $ref = $self->mode =~ /lower/i ? |
46
|
|
|
|
|
|
|
\$distances->[$j][$i] : \$distances->[$i][$j]; |
47
|
|
|
|
|
|
|
# Callback function provides the distance |
48
|
|
|
|
|
|
|
$$ref = $metric->($objects->[$i], $objects->[$j]); |
49
|
|
|
|
|
|
|
} |
50
|
|
|
|
|
|
|
} |
51
|
|
|
|
|
|
|
# Last diagonal element is undef, unless explicitly computed |
52
|
|
|
|
|
|
|
$distances->[$n-1] = [(undef)x$n] if $self->mode =~ /upper/i; |
53
|
|
|
|
|
|
|
return $distances; |
54
|
|
|
|
|
|
|
} |
55
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
|
57
|
|
|
|
|
|
|
__PACKAGE__->meta->make_immutable; |
58
|
|
|
|
|
|
|
no Moose; |
59
|
|
|
|
|
|
|
1; |
60
|
|
|
|
|
|
|
__END__ |
61
|
|
|
|
|
|
|
=pod |
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
=head1 NAME |
64
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
Algorithm::DistanceMatrix - Compute distance matrix for any distance metric |
66
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
=head1 VERSION |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
version 0.04 |
70
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
=head1 SYNOPSIS |
72
|
|
|
|
|
|
|
|
73
|
|
|
|
|
|
|
use Algorithm::DistanceMatrix; |
74
|
|
|
|
|
|
|
my $m = Algorithm::DistanceMatrix->new( |
75
|
|
|
|
|
|
|
metric=>\&mydistance,objects=\@myarray); |
76
|
|
|
|
|
|
|
my $distmatrix = $m->distancematrix; |
77
|
|
|
|
|
|
|
|
78
|
|
|
|
|
|
|
use Algorithm::Cluster qw/treecluster/; |
79
|
|
|
|
|
|
|
# method=> |
80
|
|
|
|
|
|
|
# s: single-linkage clustering |
81
|
|
|
|
|
|
|
# http://en.wikipedia.org/wiki/Single-linkage_clustering |
82
|
|
|
|
|
|
|
# m: maximum- (or complete-) linkage clustering |
83
|
|
|
|
|
|
|
# http://en.wikipedia.org/wiki/Complete_linkage_clustering |
84
|
|
|
|
|
|
|
# a: average-linkage clustering (UPGMA) |
85
|
|
|
|
|
|
|
# http://en.wikipedia.org/wiki/UPGMA |
86
|
|
|
|
|
|
|
|
87
|
|
|
|
|
|
|
my $tree = treecluster(data=>$distmat, method=>'a'); |
88
|
|
|
|
|
|
|
|
89
|
|
|
|
|
|
|
# Get your objects and the cluster IDs they belong to, assuming 5 clusters |
90
|
|
|
|
|
|
|
my $cluster_ids = $tree->cut(5); |
91
|
|
|
|
|
|
|
# Index corresponds to that of the original objects |
92
|
|
|
|
|
|
|
print $objects->[2], ' belongs to cluster ', $cluster_ids->[2], "\n"; |
93
|
|
|
|
|
|
|
|
94
|
|
|
|
|
|
|
=head1 DESCRIPTION |
95
|
|
|
|
|
|
|
|
96
|
|
|
|
|
|
|
This is a small helper package for L<Algorithm::Cluster>. That module provides |
97
|
|
|
|
|
|
|
many facilities for clustering data. It also provides a C<distancematrix> function, |
98
|
|
|
|
|
|
|
but assumes tabular data, which is the standard for gene expression data. |
99
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
If your data is tabular, you should first have a look at C<distancematrix> in |
101
|
|
|
|
|
|
|
L<Algorithm::Cluster> |
102
|
|
|
|
|
|
|
|
103
|
|
|
|
|
|
|
http://cpansearch.perl.org/src/MDEHOON/Algorithm-Cluster-1.48/doc/cluster.pdf |
104
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
Otherwise, this package provides a simple distance matrix, given an arbitrary |
106
|
|
|
|
|
|
|
distance function. It does not assume anything about your data. You simply |
107
|
|
|
|
|
|
|
provide a callback function for measuring the distance between any two objects. |
108
|
|
|
|
|
|
|
It produces a lower diagonal (by default) distance matrix that is fit to be used |
109
|
|
|
|
|
|
|
by the clustering algorithms of L<Algorithm::Cluster>. |
110
|
|
|
|
|
|
|
|
111
|
|
|
|
|
|
|
=head1 NAME |
112
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
Algorithm::DistanceMatrix - Compute distance matrix for any distance metric |
114
|
|
|
|
|
|
|
|
115
|
|
|
|
|
|
|
=head1 VERSION |
116
|
|
|
|
|
|
|
|
117
|
|
|
|
|
|
|
version 0.04 |
118
|
|
|
|
|
|
|
|
119
|
|
|
|
|
|
|
=head1 METHODS |
120
|
|
|
|
|
|
|
|
121
|
|
|
|
|
|
|
=head2 mode |
122
|
|
|
|
|
|
|
|
123
|
|
|
|
|
|
|
One of C<qw/lower upper full/> for a lower diagonal, upper diagonal, or full |
124
|
|
|
|
|
|
|
distance matrix. |
125
|
|
|
|
|
|
|
|
126
|
|
|
|
|
|
|
=head2 metric |
127
|
|
|
|
|
|
|
|
128
|
|
|
|
|
|
|
Callback for computing the distance, similarity, or whatever measure you like. |
129
|
|
|
|
|
|
|
|
130
|
|
|
|
|
|
|
$matrix->metric(\@mydistance); |
131
|
|
|
|
|
|
|
|
132
|
|
|
|
|
|
|
Where C<mydistance> receives two objects as it's first two arguments. |
133
|
|
|
|
|
|
|
|
134
|
|
|
|
|
|
|
If you need to pass special parameters to your method: |
135
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
$matrix->metric(sub{my($x,$y)=@_;mydistance(first=>$x,second=>$y,mode=>'fast')}; |
137
|
|
|
|
|
|
|
|
138
|
|
|
|
|
|
|
You may use any metric, and may return any number or object. Note that if you |
139
|
|
|
|
|
|
|
plan to use this with L<Algorithm::Cluster> this needs to be a distance metric. |
140
|
|
|
|
|
|
|
So, if you're measure how similar two things are, on a scale of 1-10, then you |
141
|
|
|
|
|
|
|
should return C<10-$similarity> to get a distance. |
142
|
|
|
|
|
|
|
|
143
|
|
|
|
|
|
|
Default is the absolute values of the scalar difference (i.e. C<abs(X-Y)>) |
144
|
|
|
|
|
|
|
|
145
|
|
|
|
|
|
|
=head2 objects |
146
|
|
|
|
|
|
|
|
147
|
|
|
|
|
|
|
Array reference. Doesn't matter what kind of objects are in the array, as long |
148
|
|
|
|
|
|
|
as your C<metric> can process them. |
149
|
|
|
|
|
|
|
|
150
|
|
|
|
|
|
|
=head2 distancematrix |
151
|
|
|
|
|
|
|
|
152
|
|
|
|
|
|
|
2D array of distances (or similarities, or whatever) between your objects. |
153
|
|
|
|
|
|
|
|
154
|
|
|
|
|
|
|
(An ArrayRef of ArrayRefs.) |
155
|
|
|
|
|
|
|
|
156
|
|
|
|
|
|
|
=head1 AUTHOR |
157
|
|
|
|
|
|
|
|
158
|
|
|
|
|
|
|
Chad A. Davis <chad.a.davis@gmail.com> |
159
|
|
|
|
|
|
|
|
160
|
|
|
|
|
|
|
=head1 COPYRIGHT AND LICENSE |
161
|
|
|
|
|
|
|
|
162
|
|
|
|
|
|
|
This software is copyright (c) 2011 by Chad A. Davis. |
163
|
|
|
|
|
|
|
|
164
|
|
|
|
|
|
|
This is free software; you can redistribute it and/or modify it under |
165
|
|
|
|
|
|
|
the same terms as the Perl 5 programming language system itself. |
166
|
|
|
|
|
|
|
|
167
|
|
|
|
|
|
|
=cut |
168
|
|
|
|
|
|
|
|