Class: Transrate::ScoreOptimiser

Inherits:
Object
  • Object
show all
Defined in:
lib/transrate/score_optimiser.rb

Overview

This class is used to calculate the basic transrate score. It is also used to automatically optimise the score by calculating a cutoff that maximises the number of reads that map while also minimising the number of low scoring contigs.

Instance Method Summary collapse

Constructor Details

#initialize(assembly, read_metrics) ⇒ ScoreOptimiser

Returns a new instance of ScoreOptimiser.



11
12
13
14
15
16
17
 
# File 'lib/transrate/score_optimiser.rb', line 11

def initialize assembly, read_metrics
  @assembly = assembly
  @fragments = read_metrics.fragments
  read_stats = read_metrics.read_stats
  @total = read_stats[:fragments]
  @good = read_stats[:good_mappings]
end

Instance Method Details

#geomean(x) ⇒ Object

Calculate the geometric mean of an array of numbers



62
63
64
65
66
67
 
# File 'lib/transrate/score_optimiser.rb', line 62

def geomean x
  sum = 0.0
  x.each{ |v| sum += Math.log(v) }
  sum /= x.size
  Math.exp sum
end

#optimal_score(prefix = 'assembly') ⇒ Object 



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
 
# File 'lib/transrate/score_optimiser.rb', line 25

def optimal_score(prefix='assembly')
  return [@optimal, @cutoff] unless @optimal.nil?
  product = 0
  good = 0
  @assembly.assembly.each do |key, contig|
    product += Math.log(contig.score)
    good += contig.good
  end
  count = @assembly.size
  cutoffscores = {}
  contigs_sorted = {}
  @assembly.assembly.sort_by { |k,v| v.score }.each do |a,b|
    contigs_sorted[a] = b
  end

  contigs_sorted.each do |key, contig|
    product -= Math.log(contig.score)
    good -= contig.good
    count -= 1
    score = Math.exp(product / count) * (good/@fragments.to_f)
    cutoffscores[contig.score] = score
  end
  @optimal = 0
  @cutoff = 0
  out = CSV.open("#{prefix}_score_optimisation.csv", 'w')
  out << %w[cutoff assembly_score]
  cutoffscores.each do |c, score|
    out << [c.round(5), score.round(5)]
    if score > @optimal
      @optimal = score
      @cutoff = c
    end
  end
  return [@optimal, @cutoff]
end

#raw_score ⇒ Object 



19
20
21
22
23
 
# File 'lib/transrate/score_optimiser.rb', line 19

def raw_score
  scores = @assembly.assembly.values.map{ |c| c.score }
  @contig_score = geomean scores
  @contig_score * (@good / @total.to_f)
end