Class: Transrate::ReadMetrics

Inherits:
Object
  • Object
show all
Defined in:
lib/transrate/read_metrics.rb,
ext/transrate/transrate.c

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(assembly) ⇒ ReadMetrics 



13
14
15
16
17
18
19
20
21
 
# File 'lib/transrate/read_metrics.rb', line 13

def initialize assembly
  @assembly = assembly # Transrate::Assembly
  @mapper = Snap.new
  @salmon = Salmon.new
  self.initial_values

  load_executables
  @read_length = 100
end

Instance Attribute Details

#bad ⇒ Object (readonly)

Returns the value of attribute bad.



8
9
10
 
# File 'lib/transrate/read_metrics.rb', line 8

def bad
  @bad
end

#fragments ⇒ Object (readonly)

Returns the value of attribute fragments.



7
8
9
 
# File 'lib/transrate/read_metrics.rb', line 7

def fragments
  @fragments
end

#fragments_mapping ⇒ Object (readonly)

Returns the value of attribute fragments_mapping.



7
8
9
 
# File 'lib/transrate/read_metrics.rb', line 7

def fragments_mapping
  @fragments_mapping
end

#good ⇒ Object (readonly)

Returns the value of attribute good.



8
9
10
 
# File 'lib/transrate/read_metrics.rb', line 8

def good
  @good
end

#has_run ⇒ Object (readonly)

Returns the value of attribute has_run.



10
11
12
 
# File 'lib/transrate/read_metrics.rb', line 10

def has_run
  @has_run
end

#p_good_mapping ⇒ Object (readonly)

Returns the value of attribute p_good_mapping.



7
8
9
 
# File 'lib/transrate/read_metrics.rb', line 7

def p_good_mapping
  @p_good_mapping
end

#read_length ⇒ Object (readonly)

Returns the value of attribute read_length.



11
12
13
 
# File 'lib/transrate/read_metrics.rb', line 11

def read_length
  @read_length
end

#supported_bridges ⇒ Object (readonly)

Returns the value of attribute supported_bridges.



9
10
11
 
# File 'lib/transrate/read_metrics.rb', line 9

def supported_bridges
  @supported_bridges
end

Instance Method Details

#analyse_bam(bamfile, csv_output) ⇒ Object 



184
185
186
187
188
189
190
191
192
193
194
195
 
# File 'lib/transrate/read_metrics.rb', line 184

def analyse_bam bamfile, csv_output
  if !File.exist?(csv_output)
    cmd = "#{@bam_reader} #{bamfile} #{csv_output} 0.7"
    reader = Cmd.new cmd
    reader.run
    if !reader.status.success?
      msg = "Couldn't get information from bam file: #{bamfile}\n"
      msg << "#{reader.stdout}\n#{reader.stderr}"
      raise TransrateError.new msg
    end
  end
end

#analyse_expression(salmon_output) ⇒ Object 



123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
 
# File 'lib/transrate/read_metrics.rb', line 123

def analyse_expression salmon_output
  salmon_output.each_pair do |name, expr|
    contig_name = Bio::FastaDefline.new(name.to_s).entry_id
    contig_name.gsub!(/;$/, '') # trim trailing semicolon
    contig = @assembly[contig_name]
    if expr[:eff_len]==0
      coverage = 0
    else
      coverage = expr[:eff_count] * @read_length / expr[:eff_len]
    end
    @contigs_uncovered += 1 if coverage < 1
    @contigs_lowcovered += 1 if coverage < 10
    contig.coverage = coverage.round(2)
    contig.eff_length = expr[:eff_len]
    contig.eff_count = expr[:eff_count]
    contig.tpm = expr[:tpm]
  end
end

#analyse_read_mappings(bamfile) ⇒ Object 



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
 
# File 'lib/transrate/read_metrics.rb', line 142

def analyse_read_mappings bamfile
  if File.exist?(bamfile) && File.size(bamfile) > 0
    csv_output = "#{File.basename(@assembly.file)}_bam_info.csv"
    csv_output = File.expand_path(csv_output)

    analyse_bam bamfile, csv_output
    # open output csv file
    @potential_bridges = 0

    CSV.foreach(csv_output, :headers => true,
                            :header_converters => :symbol,
                            :converters => :all) do |row|
      populate_contig_data row
    end
    @bad = @fragments_mapped - @good
  else
    raise TransrateError.new "couldn't find bamfile: #{bamfile}"
  end
  salmon_results = "#{File.basename @assembly.file}_quant.sf"

  if File.exist?(salmon_results)
    analyse_expression(@salmon.load_expression(salmon_results))
  else
    abort "Can't find #{salmon_results}"
  end
  update_proportions
end

#assign_and_quantify(bamfile, threads) ⇒ Object 



119
120
121
 
# File 'lib/transrate/read_metrics.rb', line 119

def assign_and_quantify(bamfile, threads)
  @salmon.run(@assembly, bamfile, threads)
end

#get_bin_path(bin) ⇒ Object 



27
28
29
30
31
32
33
34
 
# File 'lib/transrate/read_metrics.rb', line 27

def get_bin_path bin
  which_bin = Cmd.new("which #{bin}")
  which_bin.run
  if !which_bin.status.success?
    raise TransrateIOError.new("ReadMetrics: could not find #{bin} in path")
  end
  which_bin.stdout.split("\n").first
end

#get_read_length(left, right) ⇒ Object 



100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
 
# File 'lib/transrate/read_metrics.rb', line 100

def get_read_length(left, right)
  count=0
  file = File.open(left.split(",").first)
  name = file.readline.chomp
  seq = file.readline.chomp
  na = file.readline.chomp
  qual = file.readline.chomp
  read_length = 0
  while name and count < 5000 # get max read length from first 5000 reads
    read_length = [read_length, seq.length].max
    name = file.readline.chomp rescue nil
    seq = file.readline.chomp rescue nil
    na = file.readline.chomp rescue nil
    qual = file.readline.chomp rescue nil
    count+=1
  end
  read_length
end

#initial_values ⇒ Object 



223
224
225
226
227
228
229
230
231
232
233
 
# File 'lib/transrate/read_metrics.rb', line 223

def initial_values
  @fragments = 0
  @fragments_mapped = 0
  @good = 0
  @bad = 0
  @bases_uncovered = 0
  @contigs_uncovbase = 0 # any base cov < 1
  @contigs_uncovered = 0 # mean cov < 1
  @contigs_lowcovered = 0 # mean cov < 10
  @contigs_segmented = 0 # p_not_segmented < 0.5
end

#load_executables ⇒ Object 



23
24
25
 
# File 'lib/transrate/read_metrics.rb', line 23

def load_executables
  @bam_reader = get_bin_path 'bam-read'
end

#populate_contig_data(row) ⇒ Object 



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
 
# File 'lib/transrate/read_metrics.rb', line 197

def populate_contig_data row
  name = Bio::FastaDefline.new(row[:name].to_s).entry_id
  name.gsub!(/;$/, '') # trim trailing semicolon
  contig = @assembly[name]
  contig.p_seq_true = row[:p_seq_true]
  contig.uncovered_bases = row[:bases_uncovered]
  @bases_uncovered += contig.uncovered_bases
  if row[:fragments_mapped] and row[:fragments_mapped] > 1
    contig.p_good = row[:good]/row[:fragments_mapped].to_f
  end
  contig.p_not_segmented = row[:p_not_segmented]
  if contig.p_not_segmented < 0.5
    @contigs_segmented += 1
  end
  contig.in_bridges = row[:bridges]
  if row[:bridges] > 1
    @potential_bridges += 1
  end
  @fragments_mapped += row[:fragments_mapped]
  contig.good = row[:good]
  @good += row[:good]
  if row[:bases_uncovered] > 0
    @contigs_uncovbase += 1
  end
end

#read_stats ⇒ Object 



78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
 
# File 'lib/transrate/read_metrics.rb', line 78

def read_stats
  {
    :fragments => @fragments,
    :fragments_mapped => @fragments_mapped,
    :p_fragments_mapped => @p_fragments_mapped,
    :good_mappings => @good,
    :p_good_mapping => @p_good_mapping,
    :bad_mappings => @bad,
    :potential_bridges => @potential_bridges,
    :bases_uncovered => @bases_uncovered,
    :p_bases_uncovered => @p_bases_uncovered,
    :contigs_uncovbase => @contigs_uncovbase,
    :p_contigs_uncovbase => @p_contigs_uncovbase,
    :contigs_uncovered => @contigs_uncovered,
    :p_contigs_uncovered => @p_contigs_uncovered,
    :contigs_lowcovered => @contigs_lowcovered,
    :p_contigs_lowcovered => @p_contigs_lowcovered,
    :contigs_segmented => @contigs_segmented,
    :p_contigs_segmented => @p_contigs_segmented
  }
end

#run(left, right, threads: 8) ⇒ Object 



36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
 
# File 'lib/transrate/read_metrics.rb', line 36

def run left, right, threads:8
  # check all read files exist
  [left, right].each do |readfile|
    raise TransrateIOError.new "Read file is nil" if readfile.nil?
    readfile.split(",").each do |file|
      unless File.exist? file
        raise TransrateIOError.new "ReadMetrics: read file does not " +
                                   "exist: #{file}"
      end
    end
  end

  # estimate max read length
  @read_length = get_read_length(left, right)

  # map reads
  @mapper.build_index(@assembly.file, threads)
  bamfile = @mapper.map_reads(@assembly.file, left, right,
                              threads: threads)
  @fragments = @mapper.read_count

  assigned_bam = "postSample.bam"
  final_bam = "#{File.basename(bamfile, '.bam')}.assigned.bam"

  # check for latest files first and create what is needed
  if !File.exist?(final_bam)
    if !File.exist?(assigned_bam)
      assigned_bam = assign_and_quantify(bamfile, threads)
    end
    if File.exist?(assigned_bam)
      File.rename(assigned_bam, final_bam)
    else
      logger.error "Couldn't find #{assigned_bam} to rename"
      raise ReadMetricsError
    end
  end
  # analyse the final mappings
  analyse_read_mappings final_bam

  @has_run = true
end

#update_proportions ⇒ Object 



170
171
172
173
174
175
176
177
178
179
180
181
182
 
# File 'lib/transrate/read_metrics.rb', line 170

def update_proportions
  nbases = @assembly.n_bases.to_f
  ncontigs = @assembly.size.to_f

  @p_bases_uncovered = @bases_uncovered / nbases
  @p_contigs_uncovbase = @contigs_uncovbase / ncontigs
  @p_contigs_uncovered = @contigs_uncovered / ncontigs
  @p_contigs_lowcovered = @contigs_lowcovered / ncontigs
  @p_contigs_segmented = @contigs_segmented / ncontigs

  @p_good_mapping = @good.to_f / @fragments.to_f
  @p_fragments_mapped = @fragments_mapped / @fragments.to_f
end