Class: GeneValidator::AlignmentValidation

Inherits:
ValidationTest show all
Defined in:
lib/genevalidator/validation_alignment.rb

Overview

This class contains the methods necessary for validations based on multiple alignment

Instance Attribute Summary collapse

Attributes inherited from ValidationTest

#cli_name, #description, #header, #hits, #prediction, #running_time, #short_header, #type, #validation_report

Instance Method Summary collapse

Constructor Details

#initialize(type, prediction, hits, filename, raw_seq_file, index_file_name, raw_seq_file_load, db, num_threads) ⇒ AlignmentValidation

Initilizes the object Params: type: type of the predicted sequence (:nucleotide or :protein) prediction: a Sequence object representing the blast query hits: a vector of Sequence objects (representing blast hits) filename: name of the fasta file mafft_path: path of the MAFFT installation raw_seq_file: name of the fasta file with raw sequences index_file_name: name of the fasta index file raw_seq_file_load: String - loaded content of the index file



101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/genevalidator/validation_alignment.rb', line 101

def initialize(type, prediction, hits, filename, raw_seq_file,
               index_file_name, raw_seq_file_load, db, num_threads)
  super
  @short_header       = 'MA'
  @header             = 'Missing/Extra sequences'
  @description        = 'Finds missing and extra sequences in the' \
                        ' prediction, based on the multiple alignment of' \
                        ' the best hits. Also counts the percentage of' \
                        ' the conserved regions that appear in the' \
                        ' prediction.'
  @filename           = filename
  @raw_seq_file       = raw_seq_file
  @index_file_name    = index_file_name
  @raw_seq_file_load  = raw_seq_file_load
  @db                 = db
  @multiple_alignment = []
  @cli_name           = 'align'
  @num_threads        = num_threads
end

Instance Attribute Details

#filename ⇒ Object (readonly)

Returns the value of attribute filename.



84
85
86
 
# File 'lib/genevalidator/validation_alignment.rb', line 84

def filename
  @filename
end

#index_file_name ⇒ Object (readonly)

Returns the value of attribute index_file_name.



87
88
89
 
# File 'lib/genevalidator/validation_alignment.rb', line 87

def index_file_name
  @index_file_name
end

#multiple_alignment ⇒ Object (readonly)

Returns the value of attribute multiple_alignment.



85
86
87
 
# File 'lib/genevalidator/validation_alignment.rb', line 85

def multiple_alignment
  @multiple_alignment
end

#raw_seq_file ⇒ Object (readonly)

Returns the value of attribute raw_seq_file.



86
87
88
 
# File 'lib/genevalidator/validation_alignment.rb', line 86

def raw_seq_file
  @raw_seq_file
end

#raw_seq_file_load ⇒ Object (readonly)

Returns the value of attribute raw_seq_file_load.



88
89
90
 
# File 'lib/genevalidator/validation_alignment.rb', line 88

def raw_seq_file_load
  @raw_seq_file_load
end

Instance Method Details

#array_to_ranges(ar) ⇒ Object

converts an array of integers into array of ranges



405
406
407
408
409
410
411
412
413
414
 
# File 'lib/genevalidator/validation_alignment.rb', line 405

def array_to_ranges(ar)
  prev = ar[0]

  ranges = ar.slice_before { |e|
    prev, prev2 = e, prev
    prev2 + 1 != e
  }.map { |a| a[0]..a[-1] }

  ranges
end

#consensus_validation(prediction_raw, consensus) ⇒ Object

Returns the percentage of consesnsus residues from the ma that are in the prediction Params: prediction_raw: String corresponding to the prediction sequence consensus: String corresponding to the statistical model Output: Fixnum with the score



322
323
324
325
326
327
328
329
330
331
332
333
 
# File 'lib/genevalidator/validation_alignment.rb', line 322

def consensus_validation(prediction_raw, consensus)
  return 1 if prediction_raw.length != consensus.length
  # no of conserved residues among the hits
  no_conserved_residues = consensus.length - consensus.scan(/[\?-]/).length

  return 1 if no_conserved_residues == 0

  # no of conserved residues from the hita that appear in the prediction
  no_conserved_pred = consensus.split(//).each_index.select { |j| consensus[j] != '-' && consensus[j] != '?' && consensus[j] == prediction_raw[j] }.length

  no_conserved_pred / (no_conserved_residues + 0.0)
end

#extra_sequence_validation(prediction_raw, sm) ⇒ Object

Returns the percentage of extra sequences in the prediction with respect to the statistical model Params: prediction: String corresponding to the prediction sequence sm: String corresponding to the statistical model Output: Fixnum with the score



303
304
305
306
307
308
309
310
311
312
 
# File 'lib/genevalidator/validation_alignment.rb', line 303

def extra_sequence_validation(prediction_raw, sm)
  return 1 if prediction_raw.length != sm.length
  # find residues that are in the prediction
  # but not in the statistical model
  no_insertions = 0
  (0..sm.length - 1).each do |i|
    no_insertions += 1 if prediction_raw[i] != '-' && sm[i] == '-'
  end
  no_insertions / (sm.length + 0.0)
end

#gap_validation(prediction_raw, sm) ⇒ Object

Returns the percentage of gaps in the prediction with respect to the statistical model Params: prediction: String corresponding to the prediction sequence sm: String corresponding to the statistical model Output: Fixnum with the score



284
285
286
287
288
289
290
291
292
293
 
# File 'lib/genevalidator/validation_alignment.rb', line 284

def gap_validation(prediction_raw, sm)
  return 1 if prediction_raw.length != sm.length
  # find gaps in the prediction and
  # not in the statistical model
  no_gaps = 0
  (0..sm.length - 1).each do |i|
    no_gaps += 1 if prediction_raw[i] == '-' && sm[i] != '-'
  end
  no_gaps / (sm.length + 0.0)
end

#get_consensus(ma = @multiple_alignment) ⇒ Object

Returns the consensus regions among a set of multiple aligned sequences i.e positions where there is the same element in all sequences Params: ma: array of Strings, corresponding to the multiple aligned sequences Output: String with the consensus regions



271
272
273
274
 
# File 'lib/genevalidator/validation_alignment.rb', line 271

def get_consensus(ma = @multiple_alignment)
  align = Bio::Alignment.new(ma)
  align.consensus
end

#get_sm_pssm(ma = @multiple_alignment, threshold = 0.7) ⇒ Object

Builds a statistical model from a set of multiple aligned sequences based on PSSM (Position Specific Matrix) Params: ma: array of Strings, corresponding to the multiple aligned sequences threshold: percentage of genes that are considered in statistical model Output: String representing the statistical model Array with the maximum frequeny of the majoritary residue for each position



345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
 
# File 'lib/genevalidator/validation_alignment.rb', line 345

def get_sm_pssm(ma = @multiple_alignment, threshold = 0.7)
  sm = ''
  freq = []
  (0..ma[0].length - 1).each do |i|
    freqs = Hash.new(0)
    ma.map { |seq| seq[i] }.each { |res| freqs[res] += 1 }
    # get the residue with the highest frequency
    max_freq = freqs.map { |_res, n| n }.max
    residue = (freqs.map { |res, n| n == max_freq ? res : [] }.flatten)[0]

    if residue == '-'
      freq.push(0)
    else
      freq.push(max_freq / (ma.length + 0.0))
    end

    if max_freq / (ma.length + 0.0) >= threshold
      sm << residue
    else
      sm << '?'
    end
  end
  [sm, freq]
end

#isalpha(str) ⇒ Object

Returns true if the string contains only letters and false otherwise



399
400
401
 
# File 'lib/genevalidator/validation_alignment.rb', line 399

def isalpha(str)
  !str.match(/[^A-Za-z]/)
end

#multiple_align_mafft(prediction = @prediction, hits = @hits) ⇒ Object

Builds the multiple alignment between all the hits and the prediction using MAFFT tool Also creates a fasta file with the alignment Params: prediction: a Sequence object representing the blast query hits: a vector of Sequience objects (usually representing blast hits) path: path of mafft installation Output: Array of Strings, corresponding to the multiple aligned sequences the prediction is the last sequence in the vector



240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
 
# File 'lib/genevalidator/validation_alignment.rb', line 240

def multiple_align_mafft(prediction = @prediction, hits = @hits)
  fail Exception unless prediction.is_a?(Sequence) && hits[0].is_a?(Sequence)

  options = ['--maxiterate', '1000', '--localpair', '--anysymbol',
             '--quiet', '--thread', "#{@num_threads}"]
  mafft = Bio::MAFFT.new('mafft', options)
  sequences = hits.map(&:raw_sequence)
  sequences.push(prediction.protein_translation)

  report = mafft.query_align(sequences)
  # Accesses the actual alignment.
  align = report.alignment

  align.each_with_index do |s, _i|
    @multiple_alignment.push(s.to_s)
  end

  @multiple_alignment
rescue Exception
  raise NoMafftInstallationError
end

#plot_alignment(freq, output = "#{@filename}_ma.json", ma = @multiple_alignment) ⇒ Object

Generates a json file cotaining data used for plotting lines for multiple hits alignment, prediction and statistical model Params: freq: String residue frequency from the statistical model output: filename of the json file ma: String array with the multiple alignmened hits and prediction



422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
 
# File 'lib/genevalidator/validation_alignment.rb', line 422

def plot_alignment(freq, output = "#{@filename}_ma.json", ma = @multiple_alignment)
  # get indeces of consensus in the multiple alignment
  consensus = get_consensus(@multiple_alignment[0..@multiple_alignment.length - 2])
  consensus_idxs = consensus.split(//).each_index.select { |j| isalpha(consensus[j]) }
  consensus_ranges = array_to_ranges(consensus_idxs)

  consensus_all = get_consensus(@multiple_alignment)
  consensus_all_idxs = consensus_all.split(//).each_index.select { |j| isalpha(consensus_all[j]) }
  consensus_all_ranges = array_to_ranges(consensus_all_idxs)

  match_alignment = ma[0..ma.length - 2].each_with_index.map { |seq, _j| seq.split(//).each_index.select { |j| isalpha(seq[j]) } }
  match_alignment_ranges = []
  match_alignment.each { |arr| match_alignment_ranges << array_to_ranges(arr) }

  query_alignment = ma[ma.length - 1].split(//).each_index.select { |j| isalpha(ma[ma.length - 1][j]) }
  query_alignment_ranges = array_to_ranges(query_alignment)

  len = ma[0].length

  f = File.open(output, 'w')
  f.write((
  # plot statistical model
  freq.each_with_index.map { |f, j| { 'y' => ma.length, 'start' => j, 'stop' => j + 1, 'color' => 'orange', 'height' => f } } +
  # hits
  match_alignment_ranges.each_with_index.map { |ranges, j| ranges.map { |range| { 'y' => ma.length - j - 1, 'start' => range.first, 'stop' => range.last, 'color' => 'red', 'height' => -1 } } }.flatten +
  ma[0..ma.length - 2].each_with_index.map { |_seq, j|
    consensus_ranges.map { |range| { 'y' => j + 1, 'start' => range.first, 'stop' => range.last, 'color' => 'yellow', 'height' => -1 } }
  }.flatten +
  # plot prediction
  [{ 'y' => 0, 'start' => 0, 'stop' => len, 'color' => 'gray', 'height' => -1 }] +
  query_alignment_ranges.map { |range| { 'y' => 0, 'start' => range.first, 'stop' => range.last, 'color' => 'red', 'height' => -1 } }.flatten +

  # plot consensus
  consensus_all_ranges.map { |range| { 'y' => 0, 'start' => range.first, 'stop' => range.last, 'color' => 'yellow', 'height' => -1 } }.flatten).to_json)

  f.close

  yAxisValues = 'Prediction'
  (1..ma.length - 1).each do |i|
    yAxisValues << ", hit&nbsp;#{i}"
  end

  yAxisValues << ', Statistical Model'

  Plot.new(output.scan(%r{([^/]+)$})[0][0],
           :align,
           'Missing/Extra sequences Validation: Multiple Align. & Statistical model of hits',
           'Conserved Region, Yellow',
           'Offset in the Alignment',
           '',
           ma.length + 1,
           yAxisValues)
end

#remove_isolated_residues(seq, len = 2) ⇒ Object

Remove isolated residues inside long gaps from a given sequence Params: seq:String: sequence of residues len:Fixnum: number of isolated residues to be removed Output: String: the new sequence



378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
 
# File 'lib/genevalidator/validation_alignment.rb', line 378

def remove_isolated_residues(seq, len = 2)
  gap_starts = seq.to_enum(:scan, /(-\w{1,#{len}}-)/i).map { |_m| $`.size + 1 }
  # remove isolated residues
  gap_starts.each do |i|
    (i..i + len - 1).each do |j|
      seq[j] = '-' if isalpha(seq[j])
    end
  end
  # remove isolated gaps
  res_starts = seq.to_enum(:scan, /([?\w]-{1,2}[?\w])/i).map { |_m| $`.size + 1 }
  res_starts.each do |i|
    (i..i + len - 1).each do |j|
      seq[j] = '?' if seq[j] == '-'
    end
  end
  seq
end

#run(n = 10) ⇒ Object

Find gaps/extra regions based on the multiple alignment of the first n hits Output: AlignmentValidationOutput object



126
127
128
129
130
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
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
 
# File 'lib/genevalidator/validation_alignment.rb', line 126

def run(n = 10)
  n = 50 if n > 50

  fail NotEnoughHitsError unless hits.length >= n
  fail Exception unless prediction.is_a?(Sequence) &&
                        hits[0].is_a?(Sequence)
  start = Time.new
  # get the first n hits
  less_hits    = @hits[0..[n - 1, @hits.length].min]
  useless_hits = []

  # get raw sequences for less_hits
  less_hits.map do |hit|
    # get gene by accession number
    next unless hit.raw_sequence.nil?

    hit.get_sequence_from_index_file(@raw_seq_file, @index_file_name,
                                     hit.identifier, @raw_seq_file_load)

    if hit.raw_sequence.nil? || hit.raw_sequence.empty?
      seq_type = (hit.type == :protein) ? 'protein' : 'nucleotide'
      hit.get_sequence_by_accession_no(hit.accession_no, seq_type, @db)
    end

    useless_hits.push(hit) if hit.raw_sequence.nil?
    useless_hits.push(hit) if hit.raw_sequence.empty?
  end

  useless_hits.each { |hit| less_hits.delete(hit) }

  fail NoInternetError if less_hits.length == 0
  # in case of nucleotide prediction sequence translate into protein
  # translate with the reading frame of all hits considered for alignment
  reading_frames = less_hits.map(&:reading_frame).uniq
  fail ReadingFrameError if reading_frames.length != 1

  if @type == :nucleotide
    s = Bio::Sequence::NA.new(prediction.raw_sequence)
    prediction.protein_translation = s.translate(reading_frames[0])
  end

  # multiple align sequences from less_hits with the prediction
  # the prediction is the last sequence in the vector
  multiple_align_mafft(prediction, less_hits)

  out = get_sm_pssm(@multiple_alignment[0..@multiple_alignment.length - 2])
  sm = out[0]
  freq = out[1]

  # remove isolated residues from the predicted sequence
  index          = @multiple_alignment.length - 1
  prediction_raw = remove_isolated_residues(@multiple_alignment[index])
  # remove isolated residues from the statistical model
  sm = remove_isolated_residues(sm)

  a1 = get_consensus(@multiple_alignment[0..@multiple_alignment.length - 2])

  plot1     = plot_alignment(freq)
  gaps      = gap_validation(prediction_raw, sm)
  extra_seq = extra_sequence_validation(prediction_raw, sm)
  consensus = consensus_validation(prediction_raw, a1)

  @validation_report = AlignmentValidationOutput.new(@short_header, @header,
                                                     @description, gaps,
                                                     extra_seq, consensus)
  @validation_report.plot_files.push(plot1)
  @validation_report.running_time = Time.now - start
  @validation_report

rescue NotEnoughHitsError
  @validation_report = ValidationReport.new('Not enough evidence',
                                            :warning, @short_header,
                                            @header, @description,
                                            @approach, @explanation,
                                            @conclusion)
rescue NoMafftInstallationError
  @validation_report = ValidationReport.new('Mafft error', :error,
                                            @short_header, @header,
                                            @description, @approach,
                                            @explanation, @conclusion)
  @validation_report.errors.push NoMafftInstallationError
rescue NoInternetError
  @validation_report = ValidationReport.new('Internet error', :error,
                                            @short_header, @header,
                                            @description, @approach,
                                            @explanation, @conclusion)
  @validation_report.errors.push NoInternetError
rescue ReadingFrameError
  @validation_report = ValidationReport.new('Multiple reading frames',
                                            :error, @short_header,
                                            @header, @description,
                                            @approach, @explanation,
                                            @conclusion)
  @validation_report.errors.push 'Multiple reading frames Error'
rescue Exception
  @validation_report = ValidationReport.new('Unexpected error', :error,
                                            @short_header, @header,
                                            @description, @approach,
                                            @explanation, @conclusion)
  @validation_report.errors.push 'Unexpected Error'
end