Module: GO

Extended by:
Resource
Defined in:
lib/rbbt/sources/go.rb

Overview

This module holds helper methods to deal with the Gene Ontology files. Right now all it does is provide a translation form id to the actual names.

Constant Summary collapse

MULTIPLE_VALUE_FIELDS = 
%w(is_a)

Class Method Summary collapse

Class Method Details

.ancestors_in(term, valid = false) ⇒ Object 



120
121
122
123
124
125
126
127
128
129
 
# File 'lib/rbbt/sources/go.rb', line 120

def self.ancestors_in(term, valid = false)
  ancestors = id2ancestors(term)
  return ancestors if FalseClass === valid
  valid_ancestors = ancestors & valid
  return valid_ancestors if valid_ancestors.any?
  ancestors.inject([]) do |acc,ancestor|
    valid_a = ancestors_in ancestor, valid
    acc = acc + valid_a
  end.uniq
end

.descendants(id) ⇒ Object 



69
70
71
72
73
74
75
76
77
78
79
80
81
82
 
# File 'lib/rbbt/sources/go.rb', line 69

def self.descendants(id)
  list = Set.new
  new = Set.new
  new << id
  while new.any?
    list += new
    new = Set.new
    info.each do |new_id,values|
      next unless values['is_a']
      new << new_id if values['is_a'].select{|e| list.include? e.split("!").first[/GO:\d+/] }.any? && ! list.include?(new_id)
    end
  end
  list
end

.goterms ⇒ Object 



56
57
58
 
# File 'lib/rbbt/sources/go.rb', line 56

def self.goterms
  info.keys
end

.group_genes(list, valid = nil) ⇒ Object 



131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
 
# File 'lib/rbbt/sources/go.rb', line 131

def self.group_genes(list, valid = nil)
  valid = %w(GO:0005886 GO:0005634 GO:0005730 GO:0005829) if valid.nil?

  compartment_leaves = {}
  list.zip(list.go_cc_terms).each do |gene,terms|
    terms = [] if terms.nil?
    valid_terms = terms & valid
    valid_terms = terms.collect{|term| 
      (valid.include?(term) ? term : ancestors_in(term, valid))
    }.flatten
    #valid_terms - GOTerm.setup(valid_terms).flat_ancestry.flatten
    valid_terms.each do |term|
      compartment_leaves[term] ||= []
      compartment_leaves[term].push(gene)
    end
  end

  groups = {}
  while compartment_leaves.length > 1
    # Group common
    group = false
    new_compartment_leaves = {}
    compartment_leaves.each do |c,l|
      if l.length > 1
        groups[c] = l
        group = true
        new_compartment_leaves[c] = [c]
      else
        new_compartment_leaves[c] = l
      end
    end
    compartment_leaves = new_compartment_leaves

    # Pull-up leaves
    if group == false
      new_compartment_leaves = {}
      final = compartment_leaves.keys
      compartment_leaves
      compartment_leaves.each do |c,l|
        final = final - GOTerm.setup(c.dup).flat_ancestry
      end

      compartment_leaves.each do |c,l|
        if final.include? c
          valid_an = ancestors_in c, valid
          valid_an.each do |ancestor|
            new_compartment_leaves[ancestor] ||= []
            new_compartment_leaves[ancestor].push(c)
            groups[ancestor].push(c) if groups[ancestor]
          end
        else
          new_compartment_leaves[c] = l
        end
      end
      compartment_leaves = new_compartment_leaves
    end
  end
  ng = {}
  groups.keys.reverse.each do |k|
    ng[k] = {items: groups[k].uniq, id: k, name: id2name(k)}
  end
  ng
end

.id2ancestors(id) ⇒ Object 



106
107
108
 
# File 'lib/rbbt/sources/go.rb', line 106

def self.id2ancestors(id)
  id2ancestors_by_type(id, 'is_a') + id2ancestors_by_type(id, 'relationship')
end

.id2ancestors_by_type(id, type = 'is_a') ⇒ Object 



84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
 
# File 'lib/rbbt/sources/go.rb', line 84

def self.id2ancestors_by_type(id, type='is_a')
  if id.kind_of? Array
    info.values_at(*id).
      select{|i| ! i[type].nil?}.
      collect{|i| 
        res = i[type]
        res = [res] unless Array === res
        res.collect{|id| 
          id.match(/(GO:\d+)/)[1] if id.match(/(GO:\d+)/)
        }.compact
      }
  else
    return [] if id.nil? or info[id].nil? or info[id][type].nil?
    res = info[id][type]
    res = [res] unless Array === res
    res.
      collect{|id| 
      id.match(/(GO:\d+)/)[1] if id.match(/(GO:\d+)/)
    }.compact
  end
end

.id2name(id) ⇒ Object 



60
61
62
63
64
65
66
67
 
# File 'lib/rbbt/sources/go.rb', line 60

def self.id2name(id)
  if id.kind_of? Array
    info.values_at(*id).collect{|i| i['name'] if i}
  else
    return nil if info[id].nil?
    info[id]['name']
  end
end

.id2namespace(id) ⇒ Object 



110
111
112
113
114
115
116
117
118
 
# File 'lib/rbbt/sources/go.rb', line 110

def self.id2namespace(id)
  self.init unless info
  if id.kind_of? Array
    info.values_at(*id).collect{|i| i['namespace'] if i}
  else
    return nil if info[id].nil?
    info[id]['namespace']
  end
end

.info ⇒ Object 



52
53
54
 
# File 'lib/rbbt/sources/go.rb', line 52

def self.info
  @@info ||= self.init
end

.init ⇒ Object

This method needs to be called before any translations can be made, it is called automatically the first time the id2name method is called. It loads the gene_ontology.obo file and extracts all the fields, although right now, only the name field is used.



27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
 
# File 'lib/rbbt/sources/go.rb', line 27

def self.init
  Persist.persist_tsv(nil, 'gene_ontology', {}, :persist => true, serializer: :marshal) do |info|
    #info.serializer = :marshal if info.respond_to? :serializer
    Rbbt.share.databases.GO.gene_ontology.produce.read.split(/\[Term\]/).each{|term| 
      term_info = {}

      term.split(/\n/). select{|l| l =~ /:/}.each{|l| 
        key, value = l.chomp.match(/(.*?):(.*)/).values_at(1,2)
        if MULTIPLE_VALUE_FIELDS.include? key.strip
          term_info[key.strip] ||= []
          term_info[key.strip] << value.strip
        else
          term_info[key.strip] = value.strip
        end
      }

      next if term_info["id"].nil?
      info[term_info["id"]] = term_info
      info[term_info["alt_id"]] = term_info if term_info["alt_id"]
    }

    info
  end.tap{|o| o.unnamed = true}
end