Class: BacterialComparator

Inherits:
Object
  • Object
show all
Defined in:
lib/bacterial-comparator.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(options, root) ⇒ BacterialComparator

Initialize BacterialAnnotator options, options, ROOT, options, options)



19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
 
# File 'lib/bacterial-comparator.rb', line 19

def initialize options, root

  @root = root
  @outdir = options[:outdir]
  Dir.mkdir(@outdir) if ! Dir.exists? @outdir
  @genomes_list = options[:genomes_list]
  @proc = options[:proc].to_i
  @phylo_nb_genes = options[:phylo_nb_genes]
  min_cov = options[:min_cov].to_f
  min_pid = options[:pidentity].to_f
  if min_cov > 1
    min_cov = min_cov/100
  end
  if min_pid > 1
    min_pid = min_pid/100
  end

  @ref_prot = get_ref_prot
  @synteny = read_prot_synteny
  @stats = extract_syntenic_fasta min_cov, min_pid

end

Instance Attribute Details

#genomes_list ⇒ Object (readonly)

Returns the value of attribute genomes_list.



15
16
17
 
# File 'lib/bacterial-comparator.rb', line 15

def genomes_list
  @genomes_list
end

#stats ⇒ Object (readonly)

Returns the value of attribute stats.



15
16
17
 
# File 'lib/bacterial-comparator.rb', line 15

def stats
  @stats
end

Instance Method Details

#build_multifasta(ref_prot, synteny) ⇒ Object 



93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
 
# File 'lib/bacterial-comparator.rb', line 93

def build_multifasta ref_prot, synteny

  pep_out_dir = "./#{@outdir}/align-genes-pep"
  dna_out_dir = "./#{@outdir}/align-genes-dna"

  # create multifasta by syntenic proteins (pep)
  if ! File.exists? pep_out_dir+"/#{ref_prot}.pep"
    pep_out = File.open(pep_out_dir+"/#{ref_prot}.pep", "w")
    pep_file = Dir["#{@genomes_list[0]}/*.pep"]
    flatfile = Bio::FlatFile.auto("#{pep_file[0]}")
    pep_out.write(get_sequence_from_flatfile flatfile, ref_prot)
    flatfile.close
    @genomes_list.each_with_index do |g,i|
      flatfile = Bio::FlatFile.auto("#{g}/Proteins.fa")
      pep_out.write(get_sequence_from_flatfile flatfile, synteny[i][:query_prot])
      flatfile.close
    end
    pep_out.close
  end

  # create multifasta by syntenic genes (dna)
  if ! File.exists? dna_out_dir+"/#{ref_prot}.dna"
    dna_out = File.open(dna_out_dir+"/#{ref_prot}.dna", "w")
    # create multifasta by syntenic proteins
    dna_file = Dir["#{@genomes_list[0]}/*.dna"]
    flatfile = Bio::FlatFile.auto("#{dna_file[0]}")
    dna_out.write(get_sequence_from_flatfile flatfile, ref_prot)
    flatfile.close
    @genomes_list.each_with_index do |g,i|
      flatfile = Bio::FlatFile.auto("#{g}/Genes.fa")
      dna_out.write(get_sequence_from_flatfile flatfile, synteny[i][:query_prot])
      flatfile.close
    end
    dna_out.close
  end

end

#concat_alignments(outfile, ref_id) ⇒ Object 



293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
 
# File 'lib/bacterial-comparator.rb', line 293

def concat_alignments outfile, ref_id

  if File.exists?("../#{outfile}") and File.size("../#{outfile}") > 0
    puts "..Alignment concatenated file already exists, skipping."
    return
  end

  fout = File.open("../#{outfile}", "w")

  seq = ""
  Dir["*.aln"].each do |f|
    flat = Bio::FlatFile.auto(f)
    ref_seq = flat.entries[0]
    flat.close
    seq += ref_seq.seq
  end

  bioseq = Bio::Sequence.auto(seq)
  out = bioseq.output_fasta("#{ref_id}", 60)
  fout.write(out)

  for i in 1..@genomes_list.length
    seq = ""
    Dir["*.aln"].each do |f|
      flat = Bio::FlatFile.auto(f)
      j=0
      flat.each_entry do |entry|
        if j<i
          j+=1
          next
        elsif i == j
          seq += entry.seq
          j+=1
        else
          break
        end
      end
      flat.close
    end
    bioseq = Bio::Sequence.auto(seq)
    # get the file name without path prefix and extension
    genome_name = genomes_list[i-1].split("/")[-1].split(".")[0]
    out = bioseq.output_fasta(genome_name,60)
    fout.write(out)
  end

  fout.close

end

#extract_syntenic_fasta(min_cov, min_pid) ⇒ Object 



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
 
# File 'lib/bacterial-comparator.rb', line 132

def extract_syntenic_fasta min_cov, min_pid

  puts "# Extracting Proteins and Genes multifasta.."
  nb_of_syntenic = 0
  stats = {}
  stats[:syntenic] = []
  fout = File.open("#{@outdir}/cds-synteny.tsv", "w")
  fout.write("Gene\t"+@genomes_list.join("\t")+"\n")

  to_build_multifasta = []

  @synteny.each do |k,v|
    is_syntenic = 1
    v.each do |v_|
      if v_[:query_cov] == "-"
        is_syntenic = 0
        break
      elsif v_[:query_cov] > min_cov and
           v_[:ref_cov] > min_cov and
           v_[:pId] > min_pid
      # synteny -> great !
      else
        is_syntenic = 0
        break
      end
    end

    fout.write("#{k}")
    if is_syntenic == 1
      nb_of_syntenic += 1
      # build_multifasta k, v
      to_build_multifasta << [k,v]
      v.each do |x|
        fout.write("\t#{x[:query_prot]}|#{x[:pId]}|#{x[:query_cov]}|#{x[:ref_cov]}")
        stats[:syntenic] << k
      end
      fout.write("\n")
    else
      v.each do |x|
        if x[:pId] == 0.0
          fout.write("\t-")
        else
          fout.write("\t#{x[:query_prot]}|#{x[:pId]}|#{x[:query_cov]}|#{x[:ref_cov]}")
        end
      end
      fout.write("\n")
    end

  end

  fout.close

  pep_out_dir = "./#{@outdir}/align-genes-pep"
  dna_out_dir = "./#{@outdir}/align-genes-dna"
  Dir.mkdir(pep_out_dir) if ! Dir.exists? pep_out_dir
  Dir.mkdir(dna_out_dir) if ! Dir.exists? dna_out_dir

  Parallel.map(to_build_multifasta, in_processes: @proc) { |k,v|
    build_multifasta k, v
  }

  stats[:nb_of_syntenic] = nb_of_syntenic
  puts "Syntenic genes : " + nb_of_syntenic.to_s + " / " + @ref_prot.length.to_s

end

#get_ref_prot ⇒ Object 



67
68
69
70
71
72
73
74
75
76
 
# File 'lib/bacterial-comparator.rb', line 67

def get_ref_prot
  ref_prot = []
  pep_file = Dir["#{@genomes_list[0]}/*.pep"]
  flatfile = Bio::FlatFile.auto("#{pep_file[0]}")
  flatfile.each_entry do |entry|
    ref_prot << entry.definition.split(" ")[0]
  end
  flatfile.close
  ref_prot
end

#get_sequence_from_flatfile(flatfile, name) ⇒ Object 



79
80
81
82
83
84
85
86
87
88
89
90
 
# File 'lib/bacterial-comparator.rb', line 79

def get_sequence_from_flatfile flatfile, name

  out = ""
  flatfile.each_entry do |entry|
    if entry.definition.split(" ")[0] == name
      bioseq = Bio::Sequence.auto(entry.seq)
      out = bioseq.output_fasta("#{name}",60)
    end
  end
  out

end

#mafft_align(f) ⇒ Object 



199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
 
# File 'lib/bacterial-comparator.rb', line 199

def mafft_align f

  trying = 0
  begin
    cmd = system("#{@root}/mafft.linux --quiet #{f} > #{f}.aln")
    if File.size("#{f}.aln") == 0
      puts "File size of 0.. --#{f}--"
      puts "Command used : #{@root}/mafft.linux --quiet #{f} > #{f}.aln"
      fail
    else
      status = "OK"
      status = "FAILED" if cmd != true
      # puts "Alignment #{f} : #{status}"
    end
  rescue
    if trying < 3
      trying += 1
      retry
    end
    status = "FAILED"
    puts "Alignment #{f} : #{status}"
  end

end

#mafft_align_all_dna ⇒ Object 



259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
 
# File 'lib/bacterial-comparator.rb', line 259

def mafft_align_all_dna
  puts "# MAFFT multialign all gene sequences.."
  ori_dir = Dir.pwd
  Dir.chdir("#{@outdir}/align-genes-dna/")

  is_done = 1
  if Dir["*.dna"].length == Dir["*.aln"].length
    Dir["*.aln"].each do |a|
      if File.size(a) == 0
        is_done = 0
      end
    end
  else
    is_done = 0
  end

  if is_done == 0
    Parallel.map(Dir["*.dna"], in_processes: @proc) { |f|
      mafft_align f
    }
  else
    puts "..Gene alignment files already exists, skipping."
  end

  # ugly hack to find out the reference genome
  ref_id = Dir["#{ori_dir}/#{@genomes_list[0]}/*.pep"][0].split('/')[-1].gsub(".pep","")

  concat_alignments "align-genes-dna.all.fasta", ref_id

  Dir.chdir(ori_dir)

end

#mafft_align_all_pep ⇒ Object 



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
 
# File 'lib/bacterial-comparator.rb', line 225

def mafft_align_all_pep

  puts "# MAFFT multialign all protein sequences.."
  ori_dir = Dir.pwd
  Dir.chdir("#{@outdir}/align-genes-pep/")

  is_done = 1
  if Dir["*.pep"].length == Dir["*.aln"].length
    Dir["*.aln"].each do |a|
      if File.size(a) == 0
        is_done = 0
      end
    end
  else
    is_done = 0
  end

  if is_done==0
    Parallel.map(Dir["*.pep"], in_processes: @proc) { |f|
      mafft_align f
    }
  else
    puts "..Prot alignment files already exists, skipping."
  end

  # ugly hack to find out the reference genome
  ref_id = Dir["#{ori_dir}/#{@genomes_list[0]}/*.pep"][0].split('/')[-1].gsub(".pep","")

  concat_alignments "align-genes-pep.all.fasta", ref_id

  Dir.chdir(ori_dir)

end

#mafft_aln(aln_opt) ⇒ Object 



343
344
345
346
347
348
349
350
351
352
353
354
 
# File 'lib/bacterial-comparator.rb', line 343

def mafft_aln aln_opt

  if aln_opt == "both"
    mafft_align_all_pep
    mafft_align_all_dna
  elsif aln_opt == "prot"
    mafft_align_all_pep
  elsif aln_opt == "dna"
    mafft_align_all_dna
  end

end

#raxml_tree(aln_opt, bt) ⇒ Object 



391
392
393
394
395
396
397
398
399
400
401
402
 
# File 'lib/bacterial-comparator.rb', line 391

def raxml_tree aln_opt, bt

  if aln_opt == "both"
    raxml_tree_dna bt
    raxml_tree_pep bt
  elsif aln_opt == "prot"
    raxml_tree_pep bt
  elsif aln_opt == "dna"
    raxml_tree_dna bt
  end

end

#raxml_tree_dna(bt) ⇒ Object 



357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
 
# File 'lib/bacterial-comparator.rb', line 357

def raxml_tree_dna bt

  # DNA tree
  puts "# RAXML DNA tree creation.. "
  ori_dir = Dir.pwd
  Dir.chdir(@outdir)
  Dir.mkdir("tree-genes-dna") if ! Dir.exists?("tree-genes-dna")
  current_dir = Dir.pwd
  tree_dir = "#{current_dir}/tree-genes-dna"
  cmd = system("#{@root}/raxml.linux -T #{@proc} -f d -N #{bt} -s align-genes-dna.all.fasta  -m GTRGAMMA -p 123454321 -n DnaTree -w #{tree_dir}")
  cmd = system("cat #{tree_dir}/RAxML_result.DnaTree.RUN.* >> #{tree_dir}/RAxML_result.BS")
  cmd = system("#{@root}/raxml.linux -T #{@proc} -f b -z #{tree_dir}/RAxML_result.BS -t #{tree_dir}/RAxML_bestTree.DnaTree -m GTRGAMMA -n DNA_BS_TREE -w #{tree_dir}")
  cmd = system("ln -s #{tree_dir}/RAxML_bipartitionsBranchLabels.DNA_BS_TREE #{tree_dir}/../")
  Dir.chdir(ori_dir)
end

#raxml_tree_pep(bt) ⇒ Object 



373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
 
# File 'lib/bacterial-comparator.rb', line 373

def raxml_tree_pep bt

  # Prot tree
  puts "# RAXML Protein tree creation.. "
  ori_dir = Dir.pwd
  Dir.chdir(@outdir)
  Dir.mkdir("tree-genes-pep") if ! Dir.exists?("tree-genes-pep")
  current_dir = Dir.pwd
  tree_dir = "#{current_dir}/tree-genes-pep"
  cmd = system("#{@root}/raxml.linux -T #{@proc} -f d -N #{bt} -s align-genes-pep.all.fasta  -m PROTGAMMAAUTO -p 123454321 -n PepTree -w #{tree_dir}")
  cmd = system("cat #{tree_dir}/RAxML_result.PepTree.RUN.* >> #{tree_dir}/RAxML_result.BS")
  cmd = system("#{@root}/raxml.linux -T #{@proc} -f b -z #{tree_dir}/RAxML_result.BS -t #{tree_dir}/RAxML_bestTree.PepTree -m PROTGAMMAAUTO -n PEP_BS_TREE -w #{tree_dir}")
  cmd = system("ln -s #{tree_dir}/RAxML_bipartitionsBranchLabels.PEP_BS_TREE #{tree_dir}/../")
  Dir.chdir(ori_dir)

end

#read_prot_synteny ⇒ Object 



42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
 
# File 'lib/bacterial-comparator.rb', line 42

def read_prot_synteny
  puts "# Reading genomes synteny files (from genome annotations) .."
  synteny = {}
  @genomes_list.each do |g|
    puts "   #{g}/Prot-Synteny.tsv"
    genome_synteny = []
    file = File.open("#{g}/Prot-Synteny.tsv", "r")
    l = file.gets             # skip header
    while l = file.gets
      # AAK98805.1      spr0001 453     1.0     100.0   ABAC01000005_14 453     1.0
      lA = l.chomp.split("\t")
      synteny[lA[0]] =  [] if ! synteny.has_key? lA[0]
      synteny[lA[0]] << {ref_cov: lA[3].to_f, pId: lA[4].to_f, query_prot: lA[5], query_cov: lA[7].to_f}
      genome_synteny << lA[0]
    end
    @ref_prot.each do |ref_prot|
      if ! genome_synteny.include? ref_prot
        synteny[lA[0]] << {ref_cov: "-", pId: "-", query_prot: "-", query_cov: "-"}
      end
    end
    file.close
  end
  synteny
end