module MiGA::TaxDist

Methods for taxonomy identification based on AAI/ANI values.

Public Class Methods

aai_path(test) click to toggle source

Absolute path to the :intax or :novel data file (determined by test) for AAI.

# File lib/miga/tax_dist.rb, line 15
def self.aai_path(test)
  test = test.downcase.to_sym
  return nil unless [:intax, :novel].include? test
  File.expand_path("../_data/aai-#{test}.tsv.gz", __FILE__)
aai_pvalues(aai, test) click to toggle source

Returns a Hash, where the keys correspond to the taxonomic level (see MiGA::Taxonomy.LONG_RANKS for the meanings), and the values correspond to the p-values of being :intax or :novel, as determined by test.

# File lib/miga/tax_dist.rb, line 24
def self.aai_pvalues(aai, test) do |fh|
    keys = nil
    fh.each_line do |ln|
      row = ln.chomp.split(/\t/)
      if fh.lineno==1
        keys = row[1, row.size-1].map{ |i| i.to_i }
      elsif row.shift.to_f >= aai
        vals = {}
        keys.each do |i|
          v = row.shift
          next if v=="NA"
          rank = i==0 ? :root : MiGA::Taxonomy.KNOWN_RANKS[i-1]
          vals[rank] = v.to_f
        return vals
    end # each_line ln
  end # open fh
aai_taxtest(aai, test) click to toggle source

Determines the degree to which a Float aai value indicates similar taxonomy (with test :intax) or a novel taxon (with test :novel). Returns a Hash with “likelihood” phrases as keys and values as an array with cannonical rank (as in MiGA::Taxonomy) and estimated p-value.

# File lib/miga/tax_dist.rb, line 50
def self.aai_taxtest(aai, test)
  meaning = {most_likely[[0,0.01],probably[[0.01,0.1],possibly_even[[0.1,0.5]}
  pv = aai_pvalues(aai, test)
  out = {}
  meaning.each do |phrase, thresholds|
    lwr, upr = thresholds
    min ={ |v| v < upr }.max
    return out if min.nil?
    if min >= lwr
      v ={ |_,vj| vj==min }
      out[phrase] = (test==:intax ? v.reverse_each : v).first