-# Copyright (C) 2007-2011 Martin A. Hansen.
+# Copyright (C) 2007-2012 Martin A. Hansen.
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
require 'maasha/patternmatcher'
require 'maasha/bits'
require 'maasha/backtrack'
-require 'maasha/digest'
+require 'maasha/seq/digest'
+require 'maasha/seq/trim'
+require 'narray'
#require 'maasha/patscan'
# Residue alphabets
INDELS = %w[. - _ ~]
# Quality scores bases
-SCORE_PHRED = 33
-SCORE_ILLUMINA = 64
-SCORE_MIN = 0
-SCORE_MAX = 40
+SCORE_BASE = 64
+SCORE_MIN = 0
+SCORE_MAX = 40
# Error class for all exceptions to do with Seq.
class SeqError < StandardError; end
include PatternMatcher
include BackTrack
include Digest
+ include Trim
attr_accessor :seq_name, :seq, :type, :qual
self.seq.scan(regex).size
end
+ # Method to remove indels from seq and qual if qual.
+ def indels_remove
+ if self.qual.nil?
+ self.seq.delete!(Regexp.escape(INDELS.join('')))
+ else
+ na_seq = NArray.to_na(self.seq, "byte")
+ na_qual = NArray.to_na(self.qual, "byte")
+ mask = NArray.byte(self.length)
+
+ INDELS.each do |c|
+ mask += na_seq.eq(c.ord)
+ end
+
+ mask = mask.eq(0)
+
+ self.seq = na_seq[mask].to_s
+ self.qual = na_qual[mask].to_s
+ end
+
+ self
+ end
+
# Method that returns true is a given sequence type is DNA.
def is_dna?
self.type == 'dna'
# and of a given length.
def subseq(start, length = self.length - start)
raise SeqError, "subsequence start: #{start} < 0" if start < 0
- raise SeqError, "subsequence length: #{length} < 1" if length <= 0
+ raise SeqError, "subsequence length: #{length} < 0" if length < 0
raise SeqError, "subsequence start + length > Seq.length: #{start} + #{length} > #{self.length}" if start + length > self.length
- stop = start + length - 1
- seq = self.seq[start .. stop]
- qual = self.qual[start .. stop] unless self.qual.nil?
+ if length == 0
+ seq = ""
+ qual = "" unless self.qual.nil?
+ else
+ stop = start + length - 1
+
+ seq = self.seq[start .. stop]
+ qual = self.qual[start .. stop] unless self.qual.nil?
+ end
+
Seq.new(self.seq_name, seq, self.type, qual)
end
# and of a given length.
def subseq!(start, length = self.length - start)
raise SeqError, "subsequence start: #{start} < 0" if start < 0
- raise SeqError, "subsequence length: #{length} < 1" if length <= 0
+ raise SeqError, "subsequence length: #{length} < 0" if length < 0
raise SeqError, "subsequence start + length > Seq.length: #{start} + #{length} > #{self.length}" if start + length > self.length
- stop = start + length - 1
+ if length == 0
+ self.seq = ""
+ self.qual = "" unless self.qual.nil?
+ else
+ stop = start + length - 1
+
+ self.seq = self.seq[start .. stop]
+ self.qual = self.qual[start .. stop] unless self.qual.nil?
+ end
- self.seq = self.seq[start .. stop]
- self.qual = self.qual[start .. stop] unless self.qual.nil?
+ self
end
# Method that returns a subsequence of a given length
self.subseq(start, length)
end
- def quality_trim_right(min)
- raise SeqError, "no sequence" if self.seq.nil?
- raise SeqError, "no quality score" if self.qual.nil?
- raise SeqError, "minimum value: #{min} out of range #{SCORE_MIN} .. #{SCORE_MAX}" unless (SCORE_MIN .. SCORE_MAX).include? min
-
- regex_right = Regexp.new("[#{(SCORE_ILLUMINA).chr}-#{(SCORE_ILLUMINA + min).chr}]+$")
-
- self.qual.match(regex_right) do |m|
- self.subseq!(0, $`.length) if $`.length > 0
- end
-
- self
- end
-
- def quality_trim_left(min)
- raise SeqError, "no sequence" if self.seq.nil?
- raise SeqError, "no quality score" if self.qual.nil?
- raise SeqError, "minimum value: #{min} out of range #{SCORE_MIN} .. #{SCORE_MAX}" unless (SCORE_MIN .. SCORE_MAX).include? min
-
- regex_left = Regexp.new("^[#{(SCORE_ILLUMINA).chr}-#{(SCORE_ILLUMINA + min).chr}]+")
-
- self.qual.match(regex_left) do |m|
- self.subseq!(m.to_s.length, self.length - m.to_s.length) if self.length - m.to_s.length > 0
- end
-
- self
- end
-
- def quality_trim(min)
- self.quality_trim_right(min)
- self.quality_trim_left(min)
- self
- end
-
# Method that returns the residue compositions of a sequence in
# a hash where the key is the residue and the value is the residue
# count.
((self.seq.scan(/[a-z]/).size.to_f / (self.len - self.indels).to_f) * 100).round(2)
end
- # Method to convert the quality scores from a specified base
- # to another base.
- def convert_phred2illumina!
- self.qual.gsub!(/./) do |score|
- score_phred = score.ord - SCORE_PHRED
- raise SeqError, "Bad Phred score: #{score} (#{score_phred})" unless (0 .. 41).include? score_phred
- score_illumina = score_phred + SCORE_ILLUMINA
- score = score_illumina.chr
+ # Hard masks sequence residues where the corresponding quality score
+ # is below a given cutoff.
+ def mask_seq_hard_old(cutoff)
+ seq = self.seq.upcase
+ scores = self.qual
+ i = 0
+
+ scores.each_char do |score|
+ seq[i] = 'N' if score.ord - SCORE_BASE < cutoff
+ i += 1
end
+
+ self.seq = seq
end
- # Method to convert the quality scores from Solexa odd/ratio to
- # Illumina format.
- def convert_solexa2illumina!
- self.qual.gsub!(/./) do |score|
- score = solexa_char2illumina_char(score)
- end
+ # Hard masks sequence residues where the corresponding quality score
+ # is below a given cutoff.
+ def mask_seq_hard!(cutoff)
+ raise SeqError, "seq is nil" if self.seq.nil?
+ raise SeqError, "qual is nil" if self.qual.nil?
+ raise SeqError, "cufoff value: #{cutoff} out of range #{SCORE_MIN} .. #{SCORE_MAX}" unless (SCORE_MIN .. SCORE_MAX).include? cutoff
+
+ na_seq = NArray.to_na(self.seq, "byte")
+ na_qual = NArray.to_na(self.qual, "byte")
+ mask = (na_qual - SCORE_BASE) < cutoff
+ mask *= na_seq.ne("-".ord)
+
+ na_seq[mask] = 'N'.ord
+
+ self.seq = na_seq.to_s
+
+ self
end
- private
+ # Soft masks sequence residues where the corresponding quality score
+ # is below a given cutoff.
+ def mask_seq_soft!(cutoff)
+ raise SeqError, "seq is nil" if self.seq.nil?
+ raise SeqError, "qual is nil" if self.qual.nil?
+ raise SeqError, "cufoff value: #{cutoff} out of range #{SCORE_MIN} .. #{SCORE_MAX}" unless (SCORE_MIN .. SCORE_MAX).include? cutoff
+
+ na_seq = NArray.to_na(self.seq, "byte")
+ na_qual = NArray.to_na(self.qual, "byte")
+ mask = (na_qual - SCORE_BASE) < cutoff
+ mask *= na_seq.ne("-".ord)
+
+ na_seq[mask] ^= ' '.ord
- # Method to convert a Solexa score (odd ratio) to
- # a phred (probability) integer score.
- def solexa2phred(score)
- (10.0 * Math.log(10.0 ** (score / 10.0) + 1.0, 10)).to_i
+ self.seq = na_seq.to_s
+
+ self
end
- # Method to convert a Solexa score encoded using base
- # 64 ASCII to a Phred score encoded using base 64 ASCII.
- def solexa_char2illumina_char(char)
- score_solexa = char.ord - 64
- score_phred = solexa2phred(score_solexa)
- (score_phred + 64).chr
+ # Method to convert quality scores inbetween formats.
+ # Sanger base 33, range 0-40
+ # Solexa base 64, range -5-40
+ # Illumina13 base 64, range 0-40
+ # Illumina15 base 64, range 3-40
+ # Illumina18 base 33, range 0-41
+ def convert_scores!(from, to)
+ unless from == to
+ na_qual = NArray.to_na(self.qual, "byte")
+
+ case from.downcase
+ when "sanger" then na_qual -= 33
+ when "solexa" then na_qual -= 64
+ when "illumina13" then na_qual -= 64
+ when "illumina15" then na_qual -= 64
+ when "illumina18" then na_qual -= 33
+ else raise SeqError, "unknown quality score encoding: #{from}"
+ end
+
+ case to.downcase
+ when "sanger" then na_qual += 33
+ when "solexa" then na_qual += 64
+ when "illumina13" then na_qual += 64
+ when "illumina15" then na_qual += 64
+ when "illumina18" then na_qual += 33
+ else raise SeqError, "unknown quality score encoding: #{from}"
+ end
+
+ self.qual = na_qual.to_s
+ end
+
+ self
end
end