Module: Organism

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

Defined Under Namespace

Classes: OrganismNotProcessedError

Class Method Summary collapse

Class Method Details

.allowed_biomart_archives ⇒ Object 



27
28
29
30
31
 
# File 'lib/rbbt/sources/organism.rb', line 27

def self.allowed_biomart_archives
  Rbbt.etc.allowed_biomart_archives.exists? ? 
    Rbbt.etc.allowed_biomart_archives.list :
    nil
end

.attach_translations(org, tsv, target = nil, fields = nil, options = {}) ⇒ Object 



110
111
112
113
114
115
116
117
118
119
120
 
# File 'lib/rbbt/sources/organism.rb', line 110

def self.attach_translations(org, tsv, target = nil, fields = nil, options = {})
  Log.high "Attaching Translations for #{ org.inspect }, target #{target.inspect}, fields #{fields.inspect}"
  options = Misc.add_defaults options, :persist => true, :case_insensitive => false

  options.merge! :key_field    => target unless target.nil?
  options.merge! :fields => fields unless fields.nil?

  index = identifiers(org).tsv options

  tsv.attach index, :fields => [:key], :persist_input => true
end

.entrez_taxid_organism(taxid) ⇒ Object 



177
178
179
180
181
182
183
184
185
 
# File 'lib/rbbt/sources/organism.rb', line 177

def self.entrez_taxid_organism(taxid)
  all_organisms = Organism.installable_organisms

  all_organisms.each do |organism|
    return organism if Organism.entrez_taxids(organism).read.split("\n").include? taxid.to_s
  end

  raise "No organism identified for taxid #{taxid}. Supported organism are: #{all_organisms * ", "}"
end

.gene_list_bases(genes, organism = nil) ⇒ Object 



187
188
189
190
191
192
193
194
195
196
197
198
199
200
 
# File 'lib/rbbt/sources/organism.rb', line 187

def self.gene_list_bases(genes, organism = nil)
  if genes.respond_to? :orgnanism
    organism = genes.organism if organism.nil?
    genes = genes.clean_annotations
  end

  organism ||= "Hsa"

  @@gene_start_end ||= {}
  gene_start_end = @@gene_start_end[organism] ||= Organism.gene_positions(organism).tsv(:persist => true, :key_field => "Ensembl Gene ID", :fields => ["Gene Start", "Gene End"], :type => :list, :cast => :to_i, :unmamed => true)

  ranges = genes.collect{|gene| start, eend = gene_start_end[gene]; (start..eend) }
  Misc.total_length(ranges)
end

.gene_list_exon_bases(genes, organism = nil) ⇒ Object 



202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
 
# File 'lib/rbbt/sources/organism.rb', line 202

def self.gene_list_exon_bases(genes, organism = nil)
  if genes.respond_to? :orgnanism
    organism = genes.organism if organism.nil?
    genes = genes.clean_annotations
  end

  organism ||= "Hsa"

  @@gene_exons_tsv ||= {}
  gene_exons = @@gene_exons_tsv[organism] ||= Organism.exons(organism).tsv(:persist => true, :key_field => "Ensembl Gene ID", :fields => ["Ensembl Exon ID"], :type => :flat, :merge => true, :unnamed => true)

  @@exon_range_tsv ||= {}
  exon_ranges = @@exon_range_tsv[organism] ||= Organism.exons(organism).tsv(:persist => true, :fields => ["Exon Chr Start", "Exon Chr End"], :type => :list, :cast => :to_i, :unnamed => true)

  exons = gene_exons.values_at(*genes).compact.flatten.uniq

  exon_ranges = exons.collect{|exon|
    Log.low "Exon #{ exon } does not have range" and next if not exon_ranges.include? exon
    pos = exon_ranges[exon]
    (pos.first..pos.last)
  }.compact
  
  Misc.total_length(exon_ranges)
end

.guess_id(org, values) ⇒ Object 



148
149
150
151
 
# File 'lib/rbbt/sources/organism.rb', line 148

def self.guess_id(org, values)
  field_matches = TSV.field_match_counts(Organism.identifiers(org).find, values)
  field_matches.sort_by{|field, count| count.to_i}.last
end

.hg_build(organism) ⇒ Object 



41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
 
# File 'lib/rbbt/sources/organism.rb', line 41

def self.hg_build(organism)
  require 'rbbt/sources/ensembl_ftp'

  raise "Only organism 'Hsa' (Homo sapiens) supported" unless organism =~ /^Hsa/

  return 'hg20' unless organism =~ /\//
  date = organism.split("/")[1] 

  release = Ensembl.releases[date]

  release_number = release.sub(/.*-/,'').to_i
  if release_number <= 54 
    'hg18'
  elsif release_number <= 70
    'hg19'
  else
    'hg20'
  end
end

.installable_organisms ⇒ Object 



23
24
25
 
# File 'lib/rbbt/sources/organism.rb', line 23

def self.installable_organisms
  self.installed_organisms
end

.installed_organisms ⇒ Object 



19
20
21
 
# File 'lib/rbbt/sources/organism.rb', line 19

def self.installed_organisms
  Rbbt.share.install.Organism.find.glob('???').collect{|f| File.basename(f)}
end

.known_ids(name) ⇒ Object 



173
174
175
 
# File 'lib/rbbt/sources/organism.rb', line 173

def self.known_ids(name)
  TSV::Parser.new(Organism.identifiers(name).open).all_fields
end

.liftOver(positions, source, target) ⇒ Object 



61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
 
# File 'lib/rbbt/sources/organism.rb', line 61

def self.liftOver(positions, source, target)

  source_hg = hg_build(source)
  target_hg = hg_build(target)

  case
  when (source_hg == 'hg19' and target_hg == 'hg18')
    map_url = "http://hgdownload.cse.ucsc.edu/goldenPath/hg19/liftOver/hg19ToHg18.over.chain.gz" 
  when (source_hg == 'hg18' and target_hg == 'hg19')
    map_url = "http://hgdownload.cse.ucsc.edu/goldenPath/hg18/liftOver/hg18ToHg19.over.chain.gz" 
  else
    return positions
  end

  positions_bed = positions.collect{|position| 
    chr, pos = position.split(":").values_at(0,1)
    ["chr" << chr, pos.to_i-1, pos, position] * "\t"
  } * "\n" + "\n"

  new_positions = {}

  TmpFile.with_file(positions_bed) do |source_bed|
    TmpFile.with_file() do |unmapped_file|
      TmpFile.with_file() do |map_file|


        Open.write(map_file, Open.read(map_url))
        new_mutations = TmpFile.with_file() do |target_bed|
          FileUtils.chmod(755, Rbbt.software.opt.bin.liftOver.produce.find)
          CMD.cmd("#{Rbbt.software.opt.bin.liftOver.find} '#{source_bed}' '#{map_file}' '#{target_bed}' '#{unmapped_file}'").read
          Open.read(target_bed) do |line|
            chr, position_alt, position, name = line.chomp.split("\t")
            chr.sub! /chr/, ''

            old_chr, old_position, *rest = name.split(":")
            new_positions[name] = ([chr, position].concat rest) * ":"
          end
        end
      end
    end
  end

  positions.collect do |position|
    new_positions[position]
  end
end

.normalize(org, list, target = nil, fields = nil, options = {}) ⇒ Object 



122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
 
# File 'lib/rbbt/sources/organism.rb', line 122

def self.normalize(org, list, target = nil, fields = nil, options = {})
  return [] if list.nil? or list.empty?
  options = Misc.add_defaults options, :persist => true, :case_insensitive => true, :double => false
  double = Misc.process_options options, :double


  options.merge! :target => target unless target.nil?
  options.merge! :fields => fields unless fields.nil?

  index = identifiers(org).index options

  if Array === list
    if double
      index.values_at *list
    else
      index.values_at(*list).collect{|e| Misc.first e}
    end
  else
    if double
      index[list]
    else
      index[list].first
    end
  end
end

.organism(name) ⇒ Object 



161
162
163
164
165
 
# File 'lib/rbbt/sources/organism.rb', line 161

def self.organism(name)
  organisms.select{|organism|
    organism == name or Organism.scientific_name(organism) =~ /#{ name }/i
  }.first
end

.organism_code(name) ⇒ Object 



167
168
169
170
171
 
# File 'lib/rbbt/sources/organism.rb', line 167

def self.organism_code(name)
  organisms.select{|organism|
    organism == name or Organism.scientific_name(organism) =~ /#{ name }/i
  }.first
end

.organism_codes(organism = nil) ⇒ Object 



9
10
11
12
13
14
15
16
17
 
# File 'lib/rbbt/sources/organism.rb', line 9

def self.organism_codes(organism = nil)
  if organism
    Rbbt.etc.allowed_biomart_archives.list.collect{|build| [organism, build] * "/" }
  else
    Rbbt.etc.organisms.list.collect{|organism|
      organism =~ /\// ? organism : Rbbt.etc.allowed_biomart_archives.list.collect{|build| [organism, build] * "/" }
    }.flatten.compact.uniq
  end
end

.organisms ⇒ Object 



153
154
155
 
# File 'lib/rbbt/sources/organism.rb', line 153

def self.organisms
  Dir.glob(File.join(Organism.root.find, '*')).collect{|f| File.basename(f)}
end

.scientific_name(organism) ⇒ Object 



157
158
159
 
# File 'lib/rbbt/sources/organism.rb', line 157

def self.scientific_name(organism)
  Organism[organism]["scientific_name"].produce.read.strip
end