gene-matcher 0.1.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA256:
+  metadata.gz: 829cbe42a19bf6eb809358fe514714377cc7a8a856b1d2a21070760962c60d02
+  data.tar.gz: 04545fb06b196d235531d125575e246e268c7e526f9684db5fb37079b66b9844
+SHA512:
+  metadata.gz: '0854fc5cac30e888a57dda0fbda709145fe206be6acea8a29493e793075e14633a33d5ee4fa57b6b84ea4e3baa660f7a869fc14a1d51a22fc930f16f377ed2ac'
+  data.tar.gz: 23e94a7632e246589b347ad9c87445780046b3d2bf1e44ae94691ea065fe22141a7d3cb12f3b1d05b855c9f2c76682577aeeef92bad2e6ccabad6b135ef00285

data/lib/alignment.rb

@@ -0,0 +1,62 @@
+class Alignment
+  attr_accessor :score, :alignmentI, :alignmentJ, :startI, :startJ, :endI, :endJ, :reversed, :aside, :source
+
+  BLANK = "-"
+
+  def initialize
+    # スコア
+    @score = 0
+    # 検索対象（データベースに入っていた）配列
+    @alignmentI = ""
+    # 検索した（クエリ=入力された）配列
+    @alignmentJ = ""
+    # start （元の配列中の）アライメント開始位置
+    # end（元の配列中の）アライメント終了位置
+    @startI = 0
+    @startJ = 0
+    endI = 0
+    endJ = 0 
+    # 前後の反転
+    @reversed = false
+    # 逆の鎖
+    @aside = false
+    # アライメント対象配列の取得先。egtcの場合、クローンテーブルまたはアクセッションテーブル
+    @source = ""
+  end
+
+  # 二つの配列の一致部分と不一致部分を表した文字列を返す。
+  # ex. AGTCAAAAAAAAA- :...:::::::::. AT-TAAAAAAAAAG
+  # 
+  # 戻り値 :と.からなる文字列
+  def alignment
+    len = @alignmentI.length
+    len = @alignmentJ.length if @alignmentJ.length < len
+    buf = ""
+    len.times do |i|
+      ii = @alignmentI[i]
+      jj = @alignmentJ[i]
+      if ii == jj
+        buf += ":"
+      else
+        buf += "."
+      end
+    end
+    return buf
+  end
+
+  # 二つの配列の一致部分の長さを返す
+  def alignment_count
+    count = 0
+    alignment.each_char do |c|
+      count += 1 if c == ":"
+    end
+    return count  
+  end
+
+  def number_of_blank
+  end
+
+  def to_s
+    "score=#{@score} I=" + @alignmentI + " J=" + @alignmentJ;
+  end
+end

data/lib/gene-matcher.rb

@@ -0,0 +1,20 @@
+
+require_relative 'smith-waterman'
+
+class Matcher
+  attr_accessor :input_sequence, :limit
+  attr_reader :alignments
+
+  def initialize(input_sequence, limit = 0.6)
+    @limit = limit
+    @input_sequence = input_sequence
+    @alignments = []
+  end
+
+  def scan(target_sequence)
+    sw = SmithWaterman.instance
+    a = sw.alignment(target_sequence, @input_sequence)
+    @alignments += [a] if a.score >= @limit
+  end
+end
+

data/lib/smith-waterman.rb

@@ -0,0 +1,119 @@
+# Calculate the similarity of two sequences.
+# Originally created in Java language on 2004/08/04
+# Migrated to Ruby on 2023/07/10
+require_relative 'alignment'
+require 'singleton'
+
+class SmithWaterman
+  include Singleton
+
+  def alignment(target,input)
+    alignments = [] * 2
+    alignments[0] = alignment_local(target,input)
+    alignments[1] = alignment_local(target,aside_sequence(input))
+    alignments[1].aside = true
+
+    # TODO: wholeLengthI, wholeLengthJを設定する
+    # a[1].startJ = a[1].wholeLengthJ -a[1].startJ -1;
+    # a[1].endJ = a[1].wholeLengthJ -a[1].endJ-1;
+    return max(alignments)
+  end
+
+  def alignment_local(target,input)
+    raise "input is nil or empty" if input.nil? || input.empty?
+    raise "target is nil or empty" if target.nil? || target.empty?
+
+    # 行列の初期化
+
+    matrix = Array.new(target.length) { Array.new(input.length,0) }
+    maxScore = 0; maxI = 0; maxJ = 0
+    target.length.times do |i|
+      ci = target[i] # 検索対象文字
+      input.length.times do |j|
+        cj = input[j] # 検索文字
+        candidates = [0] * 4
+        candidates[0] = 0 # 未使用(常に0)
+        if i > 0 && j > 0
+          candidates[1] = matrix[i-1][j-1] + s(ci,cj)
+        else
+          candidates[1] = s(ci,cj)
+        end
+
+        if i > 0
+          candidates[2] = matrix[i-1][j] - 1
+        end
+        if j > 0
+          candidates[3] = matrix[i][j-1] - 1
+        end
+        matrix[i][j] = candidates.max
+        # スコアの最大点を記憶
+        if matrix[i][j] >= maxScore
+          maxScore = matrix[i][j]
+          maxI = i
+          maxJ = j
+        end
+      end
+      puts ci+" "+matrix[i].join(" ") if ENV["DEBUG"]      
+    end
+    puts "maxScore=#{maxScore} maxI=#{maxI} maxJ=#{maxJ}" if ENV["DEBUG"]
+
+    a = Alignment.new
+    a.endI = maxI
+    a.endJ = maxJ
+    i = maxI; j = maxJ; bufI = target[i]; bufJ = input[j]
+    while i > 0 && j > 0 do
+      dst = [] * 3
+      dst[0] = matrix[i-1][j-1]
+      dst[1] = matrix[i-1][j] if i > 0
+      dst[2] = matrix[i][j-1] if j > 0
+      break if dst.max == 0 # 行き先がなければ終了
+      case dst.index(dst.max)
+      when 0
+        i -= 1
+        j -= 1
+        bufI += target[i]
+        bufJ += input[j]
+      when 1
+        i -= 1
+        bufI += target[i]
+        bufJ += Alignment::BLANK        
+      when 2
+        j -= 1
+        bufI += Alignment::BLANK
+        bufJ += input[j]
+      end
+    end
+    a.alignmentI = bufI.reverse
+    a.alignmentJ = bufJ.reverse
+    a.startI = i
+    a.startJ = j
+
+    if a.alignmentI.length <= 20
+      a.score = 0
+    else
+      a.score = a.alignment_count ** 2 / a.alignmentI.length.to_f
+    end
+    return a
+  end
+
+  private
+    # 与えられたアライメント配列のうちスコアが最大のものを返す。
+    def max(alignments)
+      max_alignment = alignments[0]
+      alignments.each do |a|
+        if a.score > max_alignment.score
+          max_alignment = a
+        end
+      end
+      max_alignment
+    end
+
+    # スコアリング関数。一致したら1、そうでなければ0を返す。
+    def s(a,b)
+      return a == b ? 1 : 0
+    end
+
+    def aside_sequence(seq)
+      seq.reverse.tr("AGTC", "TCAG")
+    end
+end

metadata

@@ -0,0 +1,45 @@
+--- !ruby/object:Gem::Specification
+name: gene-matcher
+version: !ruby/object:Gem::Version
+  version: 0.1.1
+platform: ruby
+authors:
+- ITO Yosei
+autorequire:
+bindir: bin
+cert_chain: []
+date: 2023-08-02 00:00:00.000000000 Z
+dependencies: []
+description: Algorithm for determining similar regions between nucleic acid sequences.
+email: y-itou@lumber-mill.co.jp
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/alignment.rb
+- lib/gene-matcher.rb
+- lib/smith-waterman.rb
+homepage: https://github.com/lumbermill/gene-matcher
+licenses:
+- MIT
+metadata: {}
+post_install_message:
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubygems_version: 3.3.26
+signing_key:
+specification_version: 4
+summary: Algorithm for determining similar regions between nucleic acid sequences.
+test_files: []