rbbt-sources 3.0.32 → 3.0.33
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/rbbt/sources/go.rb +101 -2
- data/test/rbbt/sources/test_go.rb +58 -3
- metadata +2 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 1f7fdfe0006f7704b66f77aadeed785f2cd1ec4a
|
4
|
+
data.tar.gz: bda1ec6492f1dcb961d0bc5b1b2775a78a19c16c
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: aab537323d616bf284b5f54231179a808bb054d50f60fce8afd3a7b596aee4925063f737bdbc1d87ba749e786bbfec7d45356995b35c38414a614c049a37810d
|
7
|
+
data.tar.gz: dbb2e870406ea8cb7bd5bc7571b64e9c57841cb364d1553165a8510612607079917e054794e448fcdb8b70a81066c6c469463d774195a522d9bce431e02c296c
|
data/lib/rbbt/sources/go.rb
CHANGED
@@ -1,4 +1,4 @@
|
|
1
|
-
require 'rbbt'
|
1
|
+
require 'rbbt-util'
|
2
2
|
require 'rbbt/resource'
|
3
3
|
require 'rbbt/persist/tsv'
|
4
4
|
|
@@ -83,6 +83,80 @@ module GO
|
|
83
83
|
info[id]['namespace']
|
84
84
|
end
|
85
85
|
end
|
86
|
+
|
87
|
+
def self.ancestors_in(term, valid)
|
88
|
+
ancestors = id2ancestors(term)
|
89
|
+
return ancestors if FalseClass === valid
|
90
|
+
valid_ancestors = ancestors & valid
|
91
|
+
return valid_ancestors if valid_ancestors.any?
|
92
|
+
valid_ancestors.inject([]) do |acc,ancestor|
|
93
|
+
valid_a = ancestors_in ancestor, valid
|
94
|
+
acc = acc + valid_a
|
95
|
+
end
|
96
|
+
end
|
97
|
+
|
98
|
+
def self.group_genes(list, valid = nil)
|
99
|
+
valid = %w(GO:0005886 GO:0005634 GO:0005730 GO:0005829) if valid.nil?
|
100
|
+
|
101
|
+
compartment_leaves = {}
|
102
|
+
list.zip(list.go_cc_terms).each do |gene,terms|
|
103
|
+
valid_terms = terms.collect{|term|
|
104
|
+
(valid.include?(term) ? term : ancestors_in(term, valid))
|
105
|
+
}.flatten
|
106
|
+
valid_terms - GOTerm.setup(valid_terms).flat_ancestry.flatten
|
107
|
+
valid_terms.each do |term|
|
108
|
+
compartment_leaves[term] ||= []
|
109
|
+
compartment_leaves[term].push(gene)
|
110
|
+
end
|
111
|
+
end
|
112
|
+
|
113
|
+
groups = {}
|
114
|
+
while compartment_leaves.length > 1
|
115
|
+
|
116
|
+
# Group common
|
117
|
+
group = false
|
118
|
+
new_compartment_leaves = {}
|
119
|
+
compartment_leaves.each do |c,l|
|
120
|
+
if l.length > 1
|
121
|
+
groups[c] = l
|
122
|
+
group = true
|
123
|
+
new_compartment_leaves[c] = [c]
|
124
|
+
else
|
125
|
+
new_compartment_leaves[c] = l
|
126
|
+
end
|
127
|
+
end
|
128
|
+
compartment_leaves = new_compartment_leaves
|
129
|
+
|
130
|
+
# Pull-up leaves
|
131
|
+
if group == false
|
132
|
+
new_compartment_leaves = {}
|
133
|
+
final = compartment_leaves.keys
|
134
|
+
compartment_leaves.each do |c,l|
|
135
|
+
final = final - GOTerm.setup(c.dup).flat_ancestry
|
136
|
+
end
|
137
|
+
|
138
|
+
compartment_leaves.each do |c,l|
|
139
|
+
if final.include? c
|
140
|
+
valid_an = ancestors_in c, valid
|
141
|
+
valid_an.each do |ancestor|
|
142
|
+
ancestor = valid.first
|
143
|
+
next if ancestor.nil?
|
144
|
+
new_compartment_leaves[ancestor] ||= []
|
145
|
+
new_compartment_leaves[ancestor].concat(l)
|
146
|
+
end
|
147
|
+
else
|
148
|
+
new_compartment_leaves[c] = l
|
149
|
+
end
|
150
|
+
end
|
151
|
+
compartment_leaves = new_compartment_leaves
|
152
|
+
end
|
153
|
+
end
|
154
|
+
ng = {}
|
155
|
+
groups.keys.reverse.each do |k|
|
156
|
+
ng[k] = {items: groups[k], id: k, name: id2name(k)}
|
157
|
+
end
|
158
|
+
ng
|
159
|
+
end
|
86
160
|
end
|
87
161
|
|
88
162
|
if defined? Entity
|
@@ -112,6 +186,31 @@ if defined? Entity
|
|
112
186
|
description
|
113
187
|
end
|
114
188
|
|
189
|
+
property :ancestors => :single2array do
|
190
|
+
GOTerm.setup(GO.id2ancestors(self))
|
191
|
+
end
|
192
|
+
|
193
|
+
property :ancestor => :single2array do
|
194
|
+
ancestors.first
|
195
|
+
end
|
196
|
+
|
197
|
+
property :ancestry => :single2array do
|
198
|
+
ancestry = {}
|
199
|
+
self.ancestors.each do |ancestor|
|
200
|
+
prev = ancestor.ancestry
|
201
|
+
ancestry[ancestor] = prev
|
202
|
+
end
|
203
|
+
ancestry
|
204
|
+
end
|
205
|
+
|
206
|
+
property :flat_ancestry => :single2array do
|
207
|
+
ancestry = []
|
208
|
+
self.ancestors.each do |ancestor|
|
209
|
+
prev = ancestor.flat_ancestry
|
210
|
+
ancestry = ancestry + [ancestor] + prev
|
211
|
+
end
|
212
|
+
ancestry
|
213
|
+
end
|
115
214
|
end
|
116
215
|
|
117
216
|
if defined? Gene and Entity === Gene
|
@@ -131,7 +230,7 @@ if defined? Entity
|
|
131
230
|
property :go_mf_terms => :array2single do
|
132
231
|
@go_mf_terms ||= Organism.gene_go_mf(organism).tsv(:persist => true, :key_field => "Ensembl Gene ID", :fields => ["GO ID"], :type => :flat, :merge => true, :namespace => organism).chunked_values_at self.ensembl
|
133
232
|
end
|
134
|
-
|
233
|
+
|
135
234
|
end
|
136
235
|
end
|
137
236
|
end
|
@@ -1,21 +1,76 @@
|
|
1
1
|
require File.expand_path(File.dirname(__FILE__) + '/../../test_helper')
|
2
2
|
|
3
|
+
require 'rbbt/workflow'
|
4
|
+
Workflow.require_workflow "Genomics"
|
5
|
+
|
6
|
+
require 'rbbt/entity/gene'
|
3
7
|
require 'rbbt/sources/go'
|
4
8
|
require 'test/unit'
|
5
9
|
|
10
|
+
|
6
11
|
class TestGo < Test::Unit::TestCase
|
7
|
-
def
|
12
|
+
def _test_go
|
8
13
|
assert_match('vacuole inheritance',GO::id2name('GO:0000011'))
|
9
14
|
assert_equal(['vacuole inheritance','alpha-glucoside transport'], GO::id2name(['GO:0000011','GO:0000017']))
|
10
15
|
end
|
11
16
|
|
12
|
-
def
|
17
|
+
def _test_ancestors
|
13
18
|
assert GO.id2ancestors('GO:0000001').include? 'GO:0048308'
|
14
19
|
end
|
15
20
|
|
16
|
-
def
|
21
|
+
def _test_namespace
|
17
22
|
assert_equal 'biological_process', GO.id2namespace('GO:0000001')
|
18
23
|
end
|
24
|
+
|
25
|
+
def _test_ancestors
|
26
|
+
term = GOTerm.setup("GO:0005634")
|
27
|
+
end
|
28
|
+
|
29
|
+
def _test_ancestry
|
30
|
+
term = GOTerm.setup("GO:0005634")
|
31
|
+
term.ancestry.include? "GO:0005634"
|
32
|
+
end
|
33
|
+
|
34
|
+
def _test_ancestry_in
|
35
|
+
term = GOTerm.setup("GO:0005634")
|
36
|
+
valid = %w(GO:0005886 GO:0005634 GO:0005730 GO:0005829)
|
37
|
+
iii GO.ancestors_in(term, valid)
|
38
|
+
end
|
39
|
+
|
40
|
+
def test_groups
|
41
|
+
list = Gene.setup(%w(ENSG00000009413
|
42
|
+
ENSG00000038295
|
43
|
+
ENSG00000038427
|
44
|
+
ENSG00000047457
|
45
|
+
ENSG00000058668
|
46
|
+
ENSG00000065361
|
47
|
+
ENSG00000070778
|
48
|
+
ENSG00000072364
|
49
|
+
ENSG00000073711
|
50
|
+
ENSG00000075420
|
51
|
+
ENSG00000088387
|
52
|
+
ENSG00000096384
|
53
|
+
ENSG00000100345
|
54
|
+
ENSG00000102804
|
55
|
+
ENSG00000102910
|
56
|
+
ENSG00000103657
|
57
|
+
ENSG00000104043
|
58
|
+
ENSG00000106772
|
59
|
+
ENSG00000107186
|
60
|
+
ENSG00000108262
|
61
|
+
ENSG00000261163
|
62
|
+
ENSG00000263077
|
63
|
+
ENSG00000101654
|
64
|
+
ENSG00000111012), :organism => Organism.default_code("Hsa"))
|
65
|
+
|
66
|
+
valid = %w(GO:0005886 GO:0005634 GO:0005730 GO:0005829 )
|
67
|
+
iii GO.group_genes(list, valid)
|
68
|
+
end
|
69
|
+
|
70
|
+
def _test_nucleolus
|
71
|
+
nuo = "GO:0005730"
|
72
|
+
nu = "GO:0005634"
|
73
|
+
end
|
19
74
|
end
|
20
75
|
|
21
76
|
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: rbbt-sources
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 3.0.
|
4
|
+
version: 3.0.33
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Miguel Vazquez
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2015-
|
11
|
+
date: 2015-09-30 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: rbbt-util
|