macroape 3.2.2
Sign up to get free protection for your applications and to get access to all the features.
- data/.gitignore +18 -0
- data/Gemfile +4 -0
- data/LICENSE +22 -0
- data/README.md +61 -0
- data/Rakefile +7 -0
- data/bin/eval_alignment +3 -0
- data/bin/eval_similarity +3 -0
- data/bin/find_pvalue +3 -0
- data/bin/find_threshold +3 -0
- data/bin/preprocess_collection +3 -0
- data/bin/scan_collection +3 -0
- data/lib/macroape/aligned_pair_intersection.rb +136 -0
- data/lib/macroape/aligned_pair_metrics.rb +24 -0
- data/lib/macroape/aligned_pair_transformations.rb +23 -0
- data/lib/macroape/collection.rb +15 -0
- data/lib/macroape/count_by_threshold.rb +34 -0
- data/lib/macroape/exec/eval_alignment.rb +141 -0
- data/lib/macroape/exec/eval_similarity.rb +107 -0
- data/lib/macroape/exec/find_pvalue.rb +80 -0
- data/lib/macroape/exec/find_threshold.rb +76 -0
- data/lib/macroape/exec/preprocess_collection.rb +94 -0
- data/lib/macroape/exec/scan_collection.rb +124 -0
- data/lib/macroape/extract_pwm.rb +32 -0
- data/lib/macroape/gauss_estimation.rb +30 -0
- data/lib/macroape/matrix_information.rb +29 -0
- data/lib/macroape/matrix_on_background.rb +16 -0
- data/lib/macroape/matrix_transformations.rb +29 -0
- data/lib/macroape/pair_metrics.rb +9 -0
- data/lib/macroape/pair_transformations.rb +28 -0
- data/lib/macroape/pwm_compare.rb +10 -0
- data/lib/macroape/pwm_compare_aligned.rb +13 -0
- data/lib/macroape/single_matrix.rb +45 -0
- data/lib/macroape/support.rb +34 -0
- data/lib/macroape/threshold_by_pvalue.rb +68 -0
- data/lib/macroape/version.rb +3 -0
- data/lib/macroape.rb +26 -0
- data/macroape.gemspec +17 -0
- data/test/data/AHR_si.pat +10 -0
- data/test/data/KLF4_f2.pat +11 -0
- data/test/data/KLF4_f2_scan_results_all.txt +4 -0
- data/test/data/KLF4_f2_scan_results_default_cutoff.txt +3 -0
- data/test/data/KLF4_f2_scan_results_precise_mode.txt +4 -0
- data/test/data/SP1_f1.pat +12 -0
- data/test/data/SP1_f1_revcomp.pat +12 -0
- data/test/data/test_collection/GABPA_f1.pat +14 -0
- data/test/data/test_collection/KLF4_f2.pat +11 -0
- data/test/data/test_collection/SP1_f1.pat +12 -0
- data/test/data/test_collection.yaml +186 -0
- data/test/macroape_test.rb +125 -0
- metadata +116 -0
data/.gitignore
ADDED
data/Gemfile
ADDED
data/LICENSE
ADDED
@@ -0,0 +1,22 @@
|
|
1
|
+
Copyright (c) 2011-2012 Ilya Vorontsov, Ivan Kulakovskiy, Vsevolod Makeev
|
2
|
+
|
3
|
+
MIT License
|
4
|
+
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining
|
6
|
+
a copy of this software and associated documentation files (the
|
7
|
+
"Software"), to deal in the Software without restriction, including
|
8
|
+
without limitation the rights to use, copy, modify, merge, publish,
|
9
|
+
distribute, sublicense, and/or sell copies of the Software, and to
|
10
|
+
permit persons to whom the Software is furnished to do so, subject to
|
11
|
+
the following conditions:
|
12
|
+
|
13
|
+
The above copyright notice and this permission notice shall be
|
14
|
+
included in all copies or substantial portions of the Software.
|
15
|
+
|
16
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
17
|
+
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
18
|
+
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
19
|
+
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
|
20
|
+
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
|
21
|
+
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
|
22
|
+
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
data/README.md
ADDED
@@ -0,0 +1,61 @@
|
|
1
|
+
# Macroape
|
2
|
+
|
3
|
+
Macroape is abbreviation for MAtrix CompaRisOn by Approximate P-value Estimation. It's a bioinformatic tool for evaluating similarity measure between a pair of Position Weight Matrices. Used approach and application described in manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw
|
4
|
+
|
5
|
+
## Installation
|
6
|
+
|
7
|
+
Add this line to your application's Gemfile:
|
8
|
+
|
9
|
+
gem 'macroape'
|
10
|
+
|
11
|
+
And then execute:
|
12
|
+
|
13
|
+
$ bundle
|
14
|
+
|
15
|
+
Or install it yourself as:
|
16
|
+
|
17
|
+
$ gem install macroape
|
18
|
+
|
19
|
+
## Usage
|
20
|
+
For more information read manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw (not last version but comprehensive description of approach)
|
21
|
+
|
22
|
+
## Basic usage as a command-line tool
|
23
|
+
MacroAPE have 6 command line tools:
|
24
|
+
|
25
|
+
### Tools for calculating thresholds and pvalues:
|
26
|
+
* find_threshold \<PWM file\> [-p \<pvalue\> (by default: 0.0005)]
|
27
|
+
* find_pvalue \<PWM file\> \<threshold\>
|
28
|
+
|
29
|
+
### Tools for evaluating Jaccard similarity measure in the best alignment and in certain alignment:
|
30
|
+
* eval_similarity \<first PWM file\> \<second PWM file\>
|
31
|
+
* eval_alignment \<first PWM file\> \<second PWM file\> \<shift of second matrix\> \<orientation of second matrix(direct|revcomp)\>
|
32
|
+
|
33
|
+
### Tools for looking through collection for the motifs most similar to a query motif
|
34
|
+
* preprocess_collection \<folder with motif files\> [-o \<collection output file\>]
|
35
|
+
* scan_collection \<query PWM file\> \<collection file\>
|
36
|
+
|
37
|
+
Also you can use -h option to print help for a tool in console.
|
38
|
+
There are lots of different command line options. Most useful option is -d <discretization=1|10|100|1000>. You can vary precision/speed rate by specifing a discretization. For more information look through a manual.
|
39
|
+
|
40
|
+
## Basic usage in your code
|
41
|
+
require 'macroape'
|
42
|
+
background = [1,1,1,1]
|
43
|
+
discretization = 10
|
44
|
+
first_pwm_matrix = [[1,2,3,4], [1,2,3,4], [4,1,2,3,], [5,3,2,4], [4,1,2,3], [7,8,9,11]]
|
45
|
+
pwm_first = PWM::SingleMatrix.new(first_pwm_matrix).with_background(background).discrete(discretization)
|
46
|
+
pwm_second = PWM::SingleMatrix.load_pat('another_pwm.pat').with_background(background).discrete(discretization)
|
47
|
+
cmp = PWMCompare::PWMCompare.new(pwm_first, pwm_second)
|
48
|
+
first_threshold = pwm_first.threshold(pvalue)
|
49
|
+
second_threshold = pwm_second.threshold(pvalue)
|
50
|
+
similarity_info = cmp.jaccard(first_threshold, second_threshold)
|
51
|
+
puts "Jaccard similarity: #{similarity_info[:similarity]}"
|
52
|
+
|
53
|
+
## Contributing
|
54
|
+
|
55
|
+
1. Fork it
|
56
|
+
2. Create your feature branch (`git checkout -b my-new-feature`)
|
57
|
+
3. Commit your changes (`git commit -am 'Added some feature'`)
|
58
|
+
4. Push to the branch (`git push origin my-new-feature`)
|
59
|
+
5. Create new Pull Request
|
60
|
+
|
61
|
+
Copyright (c) 2011-2012 Ilya Vorontsov, Ivan Kulakovskiy, Vsevolod Makeev
|
data/Rakefile
ADDED
data/bin/eval_alignment
ADDED
data/bin/eval_similarity
ADDED
data/bin/find_pvalue
ADDED
data/bin/find_threshold
ADDED
data/bin/scan_collection
ADDED
@@ -0,0 +1,136 @@
|
|
1
|
+
module PWMCompare
|
2
|
+
module AlignedPairIntersection
|
3
|
+
|
4
|
+
def counts_for_two_matrices(threshold_first, threshold_second)
|
5
|
+
if first.probabilities == second.probabilities
|
6
|
+
if first.probabilities == [1,1,1,1]
|
7
|
+
common_words_for_two_matrices(threshold_first, threshold_second)
|
8
|
+
else
|
9
|
+
counts_for_two_matrices_with_same_probabilities(threshold_first, threshold_second)
|
10
|
+
end
|
11
|
+
else
|
12
|
+
counts_for_two_matrices_with_different_probabilities(threshold_first, threshold_second)
|
13
|
+
end
|
14
|
+
end
|
15
|
+
|
16
|
+
def counts_for_two_matrices_with_different_probabilities(threshold_first, threshold_second)
|
17
|
+
scores = { 0 => {0 => [1,1]} } # scores_on_first_pwm, scores_on_second_pwm --> count_on_first_probabilities, count_on_second_probabilities
|
18
|
+
result_first = 0.0
|
19
|
+
result_second = 0.0
|
20
|
+
length.times do |column|
|
21
|
+
ending_weight_first = first.sum_of_probabilities ** (length - column - 1)
|
22
|
+
ending_weight_second = second.sum_of_probabilities ** (length - column - 1)
|
23
|
+
already_enough_first = threshold_first - first.worst_suffix[column + 1]
|
24
|
+
already_enough_second = threshold_second - second.worst_suffix[column + 1]
|
25
|
+
least_sufficient_first = threshold_first - first.best_suffix[column + 1]
|
26
|
+
least_sufficient_second = threshold_second - second.best_suffix[column + 1]
|
27
|
+
|
28
|
+
new_scores = Hash.new{|h,k| h[k]=Hash.new{|h2,k2| h2[k2]=[0,0]}}
|
29
|
+
scores.each do |score_first, second_scores|
|
30
|
+
second_scores.each do |score_second, count|
|
31
|
+
4.times do |letter|
|
32
|
+
new_score_first = score_first + first.matrix[column][letter]
|
33
|
+
if new_score_first >= already_enough_first
|
34
|
+
new_score_second = score_second + second.matrix[column][letter]
|
35
|
+
if new_score_second >= already_enough_second
|
36
|
+
result_first += count[0] * first.probabilities[letter] * ending_weight_first
|
37
|
+
result_second += count[1] * second.probabilities[letter] * ending_weight_second
|
38
|
+
elsif new_score_second >= least_sufficient_second
|
39
|
+
new_scores[new_score_first][new_score_second][0] += count[0] * first.probabilities[letter]
|
40
|
+
new_scores[new_score_first][new_score_second][1] += count[1] * second.probabilities[letter]
|
41
|
+
end
|
42
|
+
elsif new_score_first >= least_sufficient_first
|
43
|
+
new_score_second = score_second + second.matrix[column][letter]
|
44
|
+
if new_score_second >= least_sufficient_second
|
45
|
+
new_scores[new_score_first][new_score_second][0] += count[0] * first.probabilities[letter]
|
46
|
+
new_scores[new_score_first][new_score_second][1] += count[1] * second.probabilities[letter]
|
47
|
+
end
|
48
|
+
end
|
49
|
+
end
|
50
|
+
end
|
51
|
+
end
|
52
|
+
raise 'Hash overflow in PWMCompare::AlignedPairIntersection#counts_for_two_matrices_with_different_probabilities' if new_scores.inject(0){|sum,hsh|sum+hsh.size} > MaxHashSize
|
53
|
+
scores = new_scores
|
54
|
+
end
|
55
|
+
[result_first, result_second]
|
56
|
+
end
|
57
|
+
|
58
|
+
def counts_for_two_matrices_with_same_probabilities(threshold_first, threshold_second)
|
59
|
+
scores = { 0 => {0 => 1} } # scores_on_first_pwm, scores_on_second_pwm --> count_on_first_probabilities, count_on_second_probabilities
|
60
|
+
result = 0.0
|
61
|
+
probabilities = first.probabilities
|
62
|
+
length.times do |column|
|
63
|
+
ending_weight = first.sum_of_probabilities ** (length - column - 1)
|
64
|
+
already_enough_first = threshold_first - first.worst_suffix[column + 1]
|
65
|
+
already_enough_second = threshold_second - second.worst_suffix[column + 1]
|
66
|
+
least_sufficient_first = threshold_first - first.best_suffix[column + 1]
|
67
|
+
least_sufficient_second = threshold_second - second.best_suffix[column + 1]
|
68
|
+
|
69
|
+
new_scores = Hash.new{|h,k| h[k]=Hash.new{|h2,k2| h2[k2]=0} }
|
70
|
+
scores.each do |score_first, second_scores|
|
71
|
+
second_scores.each do |score_second, count|
|
72
|
+
4.times do |letter|
|
73
|
+
new_score_first = score_first + first.matrix[column][letter]
|
74
|
+
if new_score_first >= already_enough_first
|
75
|
+
new_score_second = score_second + second.matrix[column][letter]
|
76
|
+
if new_score_second >= already_enough_second
|
77
|
+
result += count * probabilities[letter] * ending_weight
|
78
|
+
elsif new_score_second >= least_sufficient_second
|
79
|
+
new_scores[new_score_first][new_score_second] += count * probabilities[letter]
|
80
|
+
end
|
81
|
+
elsif new_score_first >= least_sufficient_first
|
82
|
+
new_score_second = score_second + second.matrix[column][letter]
|
83
|
+
if new_score_second >= least_sufficient_second
|
84
|
+
new_scores[new_score_first][new_score_second] += count * probabilities[letter]
|
85
|
+
end
|
86
|
+
end
|
87
|
+
end
|
88
|
+
end
|
89
|
+
end
|
90
|
+
raise 'Hash overflow in PWMCompare::AlignedPairIntersection#counts_for_two_matrices_with_same_probabilities' if new_scores.inject(0){|sum,hsh|sum+hsh.size} > MaxHashSize
|
91
|
+
scores = new_scores
|
92
|
+
end
|
93
|
+
[result, result]
|
94
|
+
end
|
95
|
+
|
96
|
+
|
97
|
+
def common_words_for_two_matrices(threshold_first, threshold_second)
|
98
|
+
scores = { 0 => {0 => 1} } # scores_on_first_pwm, scores_on_second_pwm --> count_on_first_probabilities, count_on_second_probabilities
|
99
|
+
result = 0
|
100
|
+
length.times do |column|
|
101
|
+
ending_weight = 4 ** (length - column - 1)
|
102
|
+
already_enough_first = threshold_first - first.worst_suffix[column + 1]
|
103
|
+
already_enough_second = threshold_second - second.worst_suffix[column + 1]
|
104
|
+
least_sufficient_first = threshold_first - first.best_suffix[column + 1]
|
105
|
+
least_sufficient_second = threshold_second - second.best_suffix[column + 1]
|
106
|
+
|
107
|
+
new_scores = Hash.new{|h,k| h[k]=Hash.new{|h2,k2| h2[k2]=0} }
|
108
|
+
scores.each do |score_first, second_scores|
|
109
|
+
second_scores.each do |score_second, count|
|
110
|
+
4.times do |letter|
|
111
|
+
new_score_first = score_first + first.matrix[column][letter]
|
112
|
+
if new_score_first >= already_enough_first
|
113
|
+
new_score_second = score_second + second.matrix[column][letter]
|
114
|
+
if new_score_second >= already_enough_second
|
115
|
+
result += count * ending_weight
|
116
|
+
elsif new_score_second >= least_sufficient_second
|
117
|
+
new_scores[new_score_first][new_score_second] += count
|
118
|
+
end
|
119
|
+
elsif new_score_first >= least_sufficient_first
|
120
|
+
new_score_second = score_second + second.matrix[column][letter]
|
121
|
+
if new_score_second >= least_sufficient_second
|
122
|
+
new_scores[new_score_first][new_score_second] += count
|
123
|
+
end
|
124
|
+
end
|
125
|
+
end
|
126
|
+
end
|
127
|
+
end
|
128
|
+
|
129
|
+
raise 'Hash overflow in PWMCompare::AlignedPairIntersection#common_words_for_two_matrices' if defined? MaxHashSize and new_scores.inject(0){|sum,hsh|sum+hsh.size} > MaxHashSize
|
130
|
+
scores = new_scores
|
131
|
+
end
|
132
|
+
[result, result]
|
133
|
+
end
|
134
|
+
|
135
|
+
end
|
136
|
+
end
|
@@ -0,0 +1,24 @@
|
|
1
|
+
module PWMCompare
|
2
|
+
module AlignedPairMetrics
|
3
|
+
def jaccard(first_threshold, second_threshold)
|
4
|
+
f = first.counts_by_thresholds(first_threshold).first
|
5
|
+
s = second.counts_by_thresholds(second_threshold).first
|
6
|
+
if f == 0 or s == 0
|
7
|
+
return {similarity: -1, tanimoto: -1, recognized_by_both: 0,
|
8
|
+
recognized_by_first: f,
|
9
|
+
recognized_by_second: s,
|
10
|
+
}
|
11
|
+
end
|
12
|
+
|
13
|
+
intersect = counts_for_two_matrices(first_threshold, second_threshold)
|
14
|
+
intersect = Math.sqrt(intersect[0] * intersect[1])
|
15
|
+
union = f + s - intersect
|
16
|
+
similarity = intersect.to_f / union
|
17
|
+
{ similarity: similarity,
|
18
|
+
tanimoto: 1.0 - similarity,
|
19
|
+
recognized_by_both: intersect,
|
20
|
+
recognized_by_first: f,
|
21
|
+
recognized_by_second: s }
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
@@ -0,0 +1,23 @@
|
|
1
|
+
module PWMCompare
|
2
|
+
module AlignedPairTransformations
|
3
|
+
|
4
|
+
#def discrete(rate)
|
5
|
+
# PWMCompareAligned.new(first.discrete(rate), second.discrete(rate))
|
6
|
+
#end
|
7
|
+
|
8
|
+
def sort_pair_of_matrices_by(&block)
|
9
|
+
mat = first.pwm.zip(second.pwm).sort_by(&block).transpose
|
10
|
+
PWMCompareAligned.new(SinglePWM(mat[0],first.probabilities), SinglePWM(mat[1], second.probabilities))
|
11
|
+
end
|
12
|
+
def sort_decreasing_max
|
13
|
+
PWMCompareAligned.new(*sort_pair_of_matrices_by{|col_pair| -col_pair[0].max} )
|
14
|
+
end
|
15
|
+
def sort_increasing_min
|
16
|
+
PWMCompareAligned.new(*sort_pair_of_matrices_by{|col_pair| col_pair[0].min} )
|
17
|
+
end
|
18
|
+
def permute_columns(permutation_index)
|
19
|
+
PWMCompareAligned.new(first.permute(permutation_index), second.permute(permutation_index))
|
20
|
+
end
|
21
|
+
|
22
|
+
end
|
23
|
+
end
|
@@ -0,0 +1,15 @@
|
|
1
|
+
module PWM
|
2
|
+
class Collection
|
3
|
+
attr_reader :rough_discretization, :precise_discretization, :background, :pvalues, :pwms, :infos
|
4
|
+
def initialize(rough_discretization, precise_discretization, background, pvalues)
|
5
|
+
@rough_discretization, @precise_discretization, @background, @pvalues = rough_discretization, precise_discretization, background, pvalues
|
6
|
+
@pwms={}
|
7
|
+
@infos={}
|
8
|
+
end
|
9
|
+
def add_pwm(pwm,info)
|
10
|
+
#@pwms[pwm] = info
|
11
|
+
@pwms[pwm.name] = pwm
|
12
|
+
@infos[pwm.name] = info
|
13
|
+
end
|
14
|
+
end
|
15
|
+
end
|
@@ -0,0 +1,34 @@
|
|
1
|
+
module PWM
|
2
|
+
module CountByThreshold
|
3
|
+
def counts_by_thresholds(*thresholds)
|
4
|
+
scores = calculate_count_distribution_after_threshold(thresholds.min)
|
5
|
+
=begin
|
6
|
+
thresholds.map{ |threshold|
|
7
|
+
#scores.select{|score,count| score >= threshold}.map{|score,count| count}.inject(0){|sum,val|sum+val}
|
8
|
+
scores.inject(0){|sum,(score,count)| (score >= threshold) ? sum + count : sum}
|
9
|
+
}
|
10
|
+
=end
|
11
|
+
s_thr= thresholds.map.with_index{|threshold,index|[threshold,index]}.sort_by{|threshold,index| threshold}
|
12
|
+
|
13
|
+
cnt = 0
|
14
|
+
thr_cnts=[]
|
15
|
+
|
16
|
+
scores.sort.reverse.each do |score,count|
|
17
|
+
while !s_thr.empty? and score < s_thr.last[0]
|
18
|
+
thr_cnts.push(cnt)
|
19
|
+
s_thr.pop
|
20
|
+
end
|
21
|
+
cnt += count
|
22
|
+
end
|
23
|
+
s_thr = thresholds.map.with_index{|threshold,index|[threshold,index]}.sort_by{|threshold,index| threshold}
|
24
|
+
while thr_cnts.size < s_thr.size
|
25
|
+
thr_cnts.push(cnt)
|
26
|
+
end
|
27
|
+
s_thr.reverse.zip(thr_cnts).sort_by{|(threshold,index), count| index}.map{|(threshold,index), count| count.to_f}
|
28
|
+
end
|
29
|
+
|
30
|
+
def pvalue_by_threshold(threshold)
|
31
|
+
counts_by_thresholds(threshold).first / number_of_words
|
32
|
+
end
|
33
|
+
end
|
34
|
+
end
|
@@ -0,0 +1,141 @@
|
|
1
|
+
help_string = %q{
|
2
|
+
Command-line format:
|
3
|
+
ruby eval_alignment.rb <1st matrix pat-file> <2nd matrix pat-file> <shift> <orientation(direct/revcomp)> [options]
|
4
|
+
type <1st matrix pat-file> <2nd matrix pat-file> | ruby eval_alignment.rb .stdin .stdin <shift> <orientation(direct/revcomp)> [options]
|
5
|
+
or in linux
|
6
|
+
cat <1st matrix pat-file> <2nd matrix pat-file> | ruby eval_alignment.rb .stdin .stdin <shift> <orientation(direct/revcomp)> [options]
|
7
|
+
|
8
|
+
Options:
|
9
|
+
[-p <P-value>]
|
10
|
+
[-d <discretization level>]
|
11
|
+
[-b <background probabilities, ACGT - 4 numbers, space-delimited, sum should be equal to 1>]
|
12
|
+
|
13
|
+
Output format:
|
14
|
+
<jaccard similarity coefficient>
|
15
|
+
<number of words recognized by both 1st and 2nd matrices | probability to draw a word recognized by both 1st and 2nd matrices> <length of the given alignment>
|
16
|
+
<aligned 1st matrix>
|
17
|
+
<aligned 2nd matrix>
|
18
|
+
<shift> <orientation>
|
19
|
+
|
20
|
+
Examples:
|
21
|
+
ruby eval_alignment.rb motifs/KLF4_f2.pat motifs/SP1_f1.pat -1 direct -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
|
22
|
+
or on windows
|
23
|
+
type motifs/SP1.pat | ruby eval_alignment.rb motifs/KLF4.pat .stdin 0 revcomp -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
|
24
|
+
or in linux
|
25
|
+
cat motifs/KLF4.pat motifs/SP1.pat | ruby eval_alignment.rb .stdin .stdin 3 direct -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
|
26
|
+
}
|
27
|
+
|
28
|
+
require 'macroape'
|
29
|
+
|
30
|
+
if ARGV.empty? or ARGV.include? '-h' or ARGV.include? '-help' or ARGV.include? '--help' or ARGV.include? '--h'
|
31
|
+
STDERR.puts help_string
|
32
|
+
exit
|
33
|
+
end
|
34
|
+
|
35
|
+
pvalue = 0.0005
|
36
|
+
discretization = 10
|
37
|
+
|
38
|
+
first_background = [1,1,1,1]
|
39
|
+
second_background = [1,1,1,1]
|
40
|
+
|
41
|
+
begin
|
42
|
+
first_file = ARGV.shift
|
43
|
+
second_file = ARGV.shift
|
44
|
+
|
45
|
+
shift = ARGV.shift
|
46
|
+
orientation = ARGV.shift
|
47
|
+
|
48
|
+
raise "You'd specify two input sources (each is filename or .stdin)" unless first_file and second_file
|
49
|
+
raise 'You\'d specify shift' unless shift
|
50
|
+
raise 'You\'d specify orientation' unless orientation
|
51
|
+
|
52
|
+
shift = shift.to_i
|
53
|
+
orientation = orientation.to_sym
|
54
|
+
|
55
|
+
case orientation
|
56
|
+
when :direct
|
57
|
+
reverse = false
|
58
|
+
when :revcomp
|
59
|
+
reverse = true
|
60
|
+
else
|
61
|
+
raise 'Unknown orientation(direct/revcomp)'
|
62
|
+
end
|
63
|
+
|
64
|
+
|
65
|
+
until ARGV.empty?
|
66
|
+
case ARGV.shift
|
67
|
+
when '-p'
|
68
|
+
pvalue = ARGV.shift.to_f
|
69
|
+
when '-d'
|
70
|
+
discretization = ARGV.shift.to_f
|
71
|
+
when '-m'
|
72
|
+
PWM::MaxHashSize = ARGV.shift.to_f
|
73
|
+
when '-md'
|
74
|
+
PWMCompare::MaxHashSize = ARGV.shift.to_f
|
75
|
+
when '-b'
|
76
|
+
second_background = first_background = ARGV.shift(4).map(&:to_f)
|
77
|
+
when '-b1'
|
78
|
+
first_background = ARGV.shift(4).map(&:to_f)
|
79
|
+
when '-b2'
|
80
|
+
second_background = ARGV.shift(4).map(&:to_f)
|
81
|
+
end
|
82
|
+
end
|
83
|
+
raise 'background should be symmetric' unless first_background == first_background.reverse
|
84
|
+
raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless second_background == second_background.reverse
|
85
|
+
|
86
|
+
|
87
|
+
PWM::MaxHashSize = 1000000 unless defined? PWM::MaxHashSize
|
88
|
+
PWMCompare::MaxHashSize = 1000 unless defined? PWMCompare::MaxHashSize
|
89
|
+
|
90
|
+
if first_file == '.stdin' || second_file == '.stdin'
|
91
|
+
r_stream, w_stream = IO.pipe
|
92
|
+
STDIN.readlines.each{|line| w_stream.write(line)}
|
93
|
+
w_stream.close
|
94
|
+
end
|
95
|
+
|
96
|
+
if first_file == '.stdin'
|
97
|
+
r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
|
98
|
+
pwm_first = PWM::SingleMatrix.load_from_line_array(extracted_pwm).with_background(first_background).discrete(discretization)
|
99
|
+
else
|
100
|
+
raise "Error! File #{first_file} don't exist" unless File.exist?(first_file)
|
101
|
+
pwm_first = PWM::SingleMatrix.load_pat(first_file).with_background(first_background).discrete(discretization)
|
102
|
+
end
|
103
|
+
|
104
|
+
if second_file == '.stdin'
|
105
|
+
r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
|
106
|
+
pwm_second = PWM::SingleMatrix.load_from_line_array(extracted_pwm).with_background(second_background).discrete(discretization)
|
107
|
+
else
|
108
|
+
raise "Error! File #{second_file} don't exist" unless File.exist?(second_file)
|
109
|
+
pwm_second = PWM::SingleMatrix.load_pat(second_file).with_background(second_background).discrete(discretization)
|
110
|
+
end
|
111
|
+
|
112
|
+
r_stream.close if first_file == '.stdin' || second_file == '.stdin'
|
113
|
+
|
114
|
+
|
115
|
+
pwm_second = pwm_second.reverse_complement if reverse
|
116
|
+
|
117
|
+
first_pwm_alignment = '.' * [-shift, 0].max + '>' * pwm_first.length
|
118
|
+
second_pwm_alignment = '.' * [shift, 0].max + (orientation == :direct ? '>' : '<') * pwm_second.length
|
119
|
+
overlap = [pwm_first.length + [-shift,0].max, pwm_second.length + [shift,0].max].min - shift.abs
|
120
|
+
alignment_length = [first_pwm_alignment.length, second_pwm_alignment.length].max
|
121
|
+
(first_pwm_alignment.length...alignment_length).each{|i| first_pwm_alignment[i] = '.'}
|
122
|
+
(second_pwm_alignment.length...alignment_length).each{|i| second_pwm_alignment[i] = '.'}
|
123
|
+
|
124
|
+
cmp = PWMCompare::PWMCompareAligned.new(pwm_first.left_augment([-shift,0].max),
|
125
|
+
pwm_second.left_augment([shift,0].max))
|
126
|
+
|
127
|
+
first_threshold = pwm_first.threshold(pvalue)
|
128
|
+
second_threshold = pwm_second.threshold(pvalue)
|
129
|
+
|
130
|
+
info = cmp.jaccard(first_threshold, second_threshold).merge(
|
131
|
+
text: "#{first_pwm_alignment}\n#{second_pwm_alignment}",
|
132
|
+
shift: shift,
|
133
|
+
orientation: orientation,
|
134
|
+
overlap: overlap,
|
135
|
+
alignment_length: alignment_length)
|
136
|
+
|
137
|
+
puts "#{info[:similarity]}\n#{info[:recognized_by_both]}\t#{info[:alignment_length]}\n#{info[:text]}\n#{info[:shift]}\t#{info[:orientation]}"
|
138
|
+
|
139
|
+
rescue => err
|
140
|
+
STDERR.puts "\n#{err}\n#{err.backtrace.first(5).join("\n")}\n\nUse -help option for help\n"
|
141
|
+
end
|
@@ -0,0 +1,107 @@
|
|
1
|
+
help_string = %q{
|
2
|
+
Command-line format:
|
3
|
+
ruby eval_similarity.rb <1st matrix pat-file> <2nd matrix pat-file> [options]
|
4
|
+
or on windows
|
5
|
+
type <1st matrix pat-file> <2nd matrix pat-file> | ruby eval_similarity.rb .stdin .stdin [options]
|
6
|
+
or in linux
|
7
|
+
cat <1st matrix pat-file> <2nd matrix pat-file> | ruby eval_similarity.rb .stdin .stdin [options]
|
8
|
+
|
9
|
+
Options:
|
10
|
+
[-p <P-value>]
|
11
|
+
[-d <discretization level>]
|
12
|
+
[-b <background probabilities, ACGT - 4 numbers, space-delimited, sum should be equal to 1>]
|
13
|
+
|
14
|
+
Output has format:
|
15
|
+
<jaccard similarity coefficient>
|
16
|
+
<number of words recognized by both 1st and 2nd matrices | probability to draw a word recognized by both 1st and 2nd matrices> <length of the optimal alignment>
|
17
|
+
<optimal alignment, the 1st matrix>
|
18
|
+
<optimal alignment, the 2nd matrix>
|
19
|
+
<shift> <orientation>
|
20
|
+
|
21
|
+
Examples:
|
22
|
+
ruby eval_similarity.rb motifs/KLF4.pat motifs/SP1.pat -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
|
23
|
+
or on windows
|
24
|
+
type motifs/SP1.pat | ruby eval_similarity.rb motifs/KLF4.pat .stdin -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
|
25
|
+
or in linux
|
26
|
+
cat motifs/KLF4.pat motifs/SP1.pat | ruby eval_similarity.rb .stdin .stdin -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
|
27
|
+
}
|
28
|
+
|
29
|
+
require 'macroape'
|
30
|
+
|
31
|
+
if ARGV.empty? or ARGV.include? '-h' or ARGV.include? '-help' or ARGV.include? '--help' or ARGV.include? '--h'
|
32
|
+
STDERR.puts help_string
|
33
|
+
exit
|
34
|
+
end
|
35
|
+
|
36
|
+
pvalue = 0.0005
|
37
|
+
discretization = 10
|
38
|
+
|
39
|
+
first_background = [1,1,1,1]
|
40
|
+
second_background = [1,1,1,1]
|
41
|
+
|
42
|
+
begin
|
43
|
+
first_file = ARGV.shift
|
44
|
+
second_file = ARGV.shift
|
45
|
+
raise "You'd specify two input sources (each is filename or .stdin)" unless first_file and second_file
|
46
|
+
|
47
|
+
until ARGV.empty?
|
48
|
+
case ARGV.shift
|
49
|
+
when '-p'
|
50
|
+
pvalue = ARGV.shift.to_f
|
51
|
+
when '-d'
|
52
|
+
discretization = ARGV.shift.to_f
|
53
|
+
when '-m'
|
54
|
+
PWM::MaxHashSize = ARGV.shift.to_f
|
55
|
+
when '-md'
|
56
|
+
PWMCompare::MaxHashSize = ARGV.shift.to_f
|
57
|
+
when '-b'
|
58
|
+
second_background = first_background = ARGV.shift(4).map(&:to_f)
|
59
|
+
when '-b1'
|
60
|
+
first_background = ARGV.shift(4).map(&:to_f)
|
61
|
+
when '-b2'
|
62
|
+
second_background = ARGV.shift(4).map(&:to_f)
|
63
|
+
end
|
64
|
+
end
|
65
|
+
raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless first_background == first_background.reverse
|
66
|
+
raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless second_background == second_background.reverse
|
67
|
+
|
68
|
+
PWM::MaxHashSize = 1000000 unless defined? PWM::MaxHashSize
|
69
|
+
PWMCompare::MaxHashSize = 1000 unless defined? PWMCompare::MaxHashSize
|
70
|
+
|
71
|
+
|
72
|
+
if first_file == '.stdin' || second_file == '.stdin'
|
73
|
+
r_stream, w_stream = IO.pipe
|
74
|
+
STDIN.readlines.each{|line| w_stream.write(line)}
|
75
|
+
w_stream.close
|
76
|
+
end
|
77
|
+
|
78
|
+
if first_file == '.stdin'
|
79
|
+
r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
|
80
|
+
pwm_first = PWM::SingleMatrix.load_from_line_array(extracted_pwm).with_background(first_background).discrete(discretization)
|
81
|
+
else
|
82
|
+
raise "Error! File #{first_file} don't exist" unless File.exist?(first_file)
|
83
|
+
pwm_first = PWM::SingleMatrix.load_pat(first_file).with_background(first_background).discrete(discretization)
|
84
|
+
end
|
85
|
+
|
86
|
+
if second_file == '.stdin'
|
87
|
+
r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
|
88
|
+
pwm_second = PWM::SingleMatrix.load_from_line_array(extracted_pwm).with_background(second_background).discrete(discretization)
|
89
|
+
else
|
90
|
+
raise "Error! File #{second_file} don't exist" unless File.exist?(second_file)
|
91
|
+
pwm_second = PWM::SingleMatrix.load_pat(second_file).with_background(second_background).discrete(discretization)
|
92
|
+
end
|
93
|
+
|
94
|
+
r_stream.close if first_file == '.stdin' || second_file == '.stdin'
|
95
|
+
|
96
|
+
cmp = PWMCompare::PWMCompare.new(pwm_first, pwm_second)
|
97
|
+
|
98
|
+
first_threshold = pwm_first.threshold(pvalue)
|
99
|
+
second_threshold = pwm_second.threshold(pvalue)
|
100
|
+
|
101
|
+
info = cmp.jaccard(first_threshold, second_threshold)
|
102
|
+
|
103
|
+
puts "#{info[:similarity]}\n#{info[:recognized_by_both]}\t#{info[:alignment_length]}\n#{info[:text]}\n#{info[:shift]}\t#{info[:orientation]}"
|
104
|
+
|
105
|
+
rescue => err
|
106
|
+
STDERR.puts "\n#{err}\n#{err.backtrace.first(5).join("\n")}\n\nUse -help option for help\n"
|
107
|
+
end
|