Module: BioMart

Defined in:
lib/rbbt/sources/biomart.rb

Overview

This module interacts with BioMart. It performs queries to BioMart and synthesises a hash with the results. Note that this module connects to the online BioMart WS using the Open in ‘rbbt/util/open’ module which offers caching by default. To obtain up to date results you may need to clear the cache from previous queries.

Defined Under Namespace

Classes: QueryError

Constant Summary collapse

BIOMART_URL = 
'ensembl.org/biomart/martservice'
MISSING_IN_ARCHIVE = 
Rbbt.etc.biomart.missing_in_archive.exists? ? Rbbt.etc.biomart.missing_in_archive.find.yaml : {}

Class Method Summary collapse

Class Method Details

.query(database, main, attrs = nil, filters = nil, data = nil, open_options = {}) ⇒ Object

This method performs a query in biomart for a datasets and a given set of attributes, there must be a main attribute that will be used as the key in the result hash, optionally there may be a list of additional attributes and filters. The data parameter at the end is used internally to incrementally building the result, due to a limitation of the BioMart WS that only allows 3 external arguments, users normally should leave it unspecified or nil. The result is a hash, where the keys are the different values for the main attribute, and the value is a hash with every other attribute as key, and as value and array with all possible values (Note that for a given value of the main attribute, there may be more than one value for another attribute). If filters is left a nil it adds a filter to the BioMart query to remove results with the main attribute empty, this may cause an error if the BioMart WS does not allow filtering with that attribute.



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
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
 
# File 'lib/rbbt/sources/biomart.rb', line 161

def self.query(database, main, attrs = nil, filters = nil, data = nil, open_options = {})
  IndiferentHash.setup(open_options)
  open_options = Misc.add_defaults open_options, :nocache => false, :filename => nil, :field_names => nil, :by_chr => false
  filename, field_names, by_chr = Misc.process_options open_options, :filename, :field_names, :by_chr
  attrs   ||= []
  open_options = Misc.add_defaults open_options, :keep_empty => false, :merge => true

  IndiferentHash.setup(open_options)

  Log.low "BioMart query: '#{main}' [#{(attrs || []) * ', '}] [#{Log.fingerprint filters}] #{open_options.inspect}"

  max_items = 1
  chunks = []
  chunk = []
  attrs.each{|a|
    chunk << a
    if chunk.length == max_items
      chunks << chunk
      chunk = []
    end
  }

  chunks << chunk if chunk.any?

  chunks << [] if chunks.empty?

  Log.low "Chunks: #{chunks.length}"
  if chunks.any?
    chunks.each_with_index{|chunk,i|
      Log.low "Chunk #{ i + 1 } / #{chunks.length}: [#{chunk * ", "}]"
      data = get(database, main, chunk, filters, data, open_options.dup)
    }
  else
    data = get(database, main, [], filters, data, open_options.dup)
  end

  open_options[:filename] = "BioMart[#{main}+#{attrs.length}]"

  if filename.nil?
    results = TSV.open data, open_options
    results.key_field = main
    results.fields = attrs
    results
  else
    Open.write(filename) do |f|
      f.puts "#: " << Misc.hash2string(TSV.annotations{|key| [key, open_options[key]]})
      if field_names.nil?
        f.puts "#" << [main, attrs].flatten * "\t"
      else
        f.puts "#" << field_names * "\t"
      end
      f.write Open.read(data)
    end
    FileUtils.rm data
    filename
  end
end

.tsv(database, main, attrs = nil, filters = nil, data = nil, open_options = {}) ⇒ Object 



219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
 
# File 'lib/rbbt/sources/biomart.rb', line 219

def self.tsv(database, main, attrs = nil, filters = nil, data = nil, open_options = {})
  attrs ||= []

  current_archive = Thread.current['archive'] 
  missing = MISSING_IN_ARCHIVE['all'] || []
  missing += MISSING_IN_ARCHIVE[current_archive] || [] if current_archive

  MISSING_IN_ARCHIVE.each do |k,v|
    if k =~ /^<(.*)/ 
      t = $1.strip
      missing+=v if Organism.compare_archives(current_archive, t) == -1
    elsif k=~ /^>(.*)/ 
      t = $1.strip
      missing+=v if Organism.compare_archives(current_archive, t) == 1
    end
  end
  attrs = attrs.uniq.reject{|attr| missing.include? attr[1]}
  changes = {}
  missing.select{|m| m.include? "~" }.each do |str|
    orig,_sep, new = str.partition "~"
    if orig.include?(":")
      target_db, _sep, orig = orig.partition(":")
      if target_db[0] == "-"
        next if database == target_db[1..-1]
      else
        next unless database == target_db
      end
      changes[orig] = new
    else
      changes[orig] = new
    end
  end
  changed = true
  while changed
    new_attrs = attrs.collect{|n,k| [n, changes[k] || k] }
    changed = new_attrs != attrs
    attrs = new_attrs
  end


  codes = attrs.collect{|attr| attr[1]}
  if open_options[:filename].nil?
    tsv = query(database, main.last, codes, filters, data, open_options)
    tsv.key_field = main.first
    tsv.fields    = attrs.collect{|attr| attr.first} 
    tsv
  else
    query(database, main.last, codes, filters, data, open_options.merge(:field_names => [main.first, attrs.collect{|attr| attr.first}].flatten))
  end
end