dna 0.1.2 → 0.2.0

data/.rspec

@@ -1 +1 @@
---color
+--color -f d

data/Gemfile

@@ -1,9 +1,9 @@
 source :rubygems
 
 group :test do
-  gem 'rspec'
+  gem 'rspec', '~> 2.11.0'
 end
 
 group :development do
-  gem 'jeweler'
+  gem 'jeweler', '~> 1.8.4'
 end

data/dna.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = "dna"
-  s.version = "0.1.2"
+  s.version = "0.2.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Austin G. Davis-Richardson"]
-  s.date = "2012-08-17"
+  s.date = "2012-12-14"
   s.description = "Simple FASTA/FASTQ/QSEQ parser library for Ruby."
   s.email = "harekrishna@gmail.com"
   s.executables = ["dna"]
@@ -26,6 +26,7 @@ Gem::Specification.new do |s|
     "lib/dna/parsers/fasta.rb",
     "lib/dna/parsers/fastq.rb",
     "lib/dna/parsers/qseq.rb",
+    "lib/dna/phred.rb",
     "lib/dna/record.rb",
     "lib/dna/version.rb",
     "readme.md",
@@ -35,25 +36,26 @@ Gem::Specification.new do |s|
     "spec/data/test.fastq",
     "spec/data/test.qseq",
     "spec/dna_spec.rb",
+    "spec/phred_spec.rb",
     "spec/record_spec.rb",
     "spec/spec_helper.rb"
   ]
   s.homepage = "http://audy.github.com/dna"
   s.licenses = ["MIT"]
   s.require_paths = ["lib"]
-  s.rubygems_version = "1.8.24"
+  s.rubygems_version = "1.8.23"
   s.summary = "Simple FASTA/FASTQ/QSEQ parser library for Ruby"
 
   if s.respond_to? :specification_version then
     s.specification_version = 3
 
     if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
-      s.add_development_dependency(%q<jeweler>, [">= 0"])
+      s.add_development_dependency(%q<jeweler>, ["~> 1.8.4"])
     else
-      s.add_dependency(%q<jeweler>, [">= 0"])
+      s.add_dependency(%q<jeweler>, ["~> 1.8.4"])
     end
   else
-    s.add_dependency(%q<jeweler>, [">= 0"])
+    s.add_dependency(%q<jeweler>, ["~> 1.8.4"])
   end
 end
 

data/lib/dna/dna.rb

@@ -1,4 +1,3 @@
-
 ##
 # Dna
 #

data/lib/dna/phred.rb

@@ -0,0 +1,50 @@
+#
+# Phred - intended to be used as a mixin to add phred-score
+# conversions to integers with byte-offset taken into account
+# and also for conversion from byte to probability.
+#
+# http://en.wikipedia.org/wiki/FASTQ_format
+#
+module Phred
+
+  formats = {
+    :illumina => { :offset => 33, :prob_func => Proc.new { |q| 10**(q/-10.0) } },
+    :sanger => { :offset => 33, :prob_func => Proc.new { |q| 10**(q/-10.0) } },
+    :solexa => { :offset => 64, :prob_func => Proc.new { |q| (10**(q/-10.0))/(1 + (10**(q/-10.0))) } },
+  }
+
+  #
+  # use metaprogramming to create methods for converting
+  # raw ascii quality string to quality socres and probabilities
+  #
+  # Do this without using method_missing so that the methods
+  # are visable.
+  #
+  formats.each do |type, props|
+    offset = props[:offset]
+    prob_func = props[:prob_func]
+
+    define_method "#{type}_scores" do
+      qualities_with_offset(offset)
+    end
+
+    define_method "#{type}_probabilities" do
+      probabilities_with_offset(offset, prob_func)
+    end
+  end
+
+  extend self
+
+  private
+
+  def qualities_with_offset(offset)
+    @qualities.each_byte.map { |q| q - offset }
+  end
+
+  def probabilities_with_offset(offset, func)
+    qualities_with_offset(offset).map do |q|
+      func.call(q)
+    end
+  end
+
+end

data/lib/dna/record.rb

@@ -1,4 +1,6 @@
 class Record # nucleotide record
+  include Phred
+
   def length
     @sequence.length
   end

data/lib/dna/version.rb

@@ -1,8 +1,8 @@
 class Dna
   class Version
   MAJOR = 0
-  MINOR = 1
-  PATCH = 2
+  MINOR = 2
+  PATCH = 0
 
   STRING = [MAJOR, MINOR, PATCH].join('.')
   end

data/spec/phred_spec.rb

@@ -0,0 +1,75 @@
+class TestRecord
+  include Phred
+  attr_accessor :qualities
+end
+
+##
+# Definition of different FASTQ formats for Phred scores
+# as defined in the wikipedia article for the FASTQ format.
+#
+# Each format definition needs a :string, a :scores and a :probabilities
+# Conversion between these properties is automatically tested by the
+# test suite.
+#
+## Solexa and Illumina <= 1.3 (we just call it Solexa here)
+#
+# - 0-93 with ascii offset 33
+# - Probability of error:
+#     `Q = -10log10(P/1-P)`
+#     Solve for P yourself :P
+#
+## Sanger
+#
+# - 0-62 with ascii ofset 64
+# - Probability of error given by p = 10**(Q/10)
+# - Quality = -10log10(P)
+#
+## Illumina 1.3+
+#
+# Illumina-formatted quality scores are the same as sanger except:
+# - 0 and 1 are not used
+# - 2 is reserved for B which just means 'really bad, do not use'
+#
+## 454
+#
+# - No one uses 454.
+#
+#
+format_definitions = {
+  :solexa => {
+    :string => (0..62).map { |q| (q + 64).chr }.join,
+    :scores => (0..62).to_a,
+    :probabilities => (0..62).map { |q| (10**((q)/-10.0))/(1 + 10**(q/-10.0)) }
+  },
+
+  :illumina => {
+    :string => (3..93).map { |q| (q + 33).chr }.join + 'B'*4,
+    :scores => (3..93).to_a + [33, 33, 33, 33],
+    :probabilities => ((3..93).to_a + [33, 33, 33, 33]).map { |q| 10**((q)/-10.0) }
+  },
+
+  :sanger => {
+    :string => (0..93).map { |q| (q + 33).chr }.join,
+    :scores => (0..93).to_a,
+    :probabilities => (0..93).map { |q| 10**((q)/-10.0) }
+  }
+}
+
+describe Phred do
+
+  let (:test_record) { TestRecord.new }
+
+  format_definitions.each_pair do |type, properties|
+
+    it "should properly convert #{type} raw ascii PHRED scores to integers" do
+      test_record.qualities = properties[:string]
+      test_record.send(:"#{type}_scores").should == properties[:scores]
+    end
+
+    it "should properly convert #{type.to_s} raw ascii PHRED scores to probabilities" do
+      test_record.qualities = properties[:string]
+      test_record.send(:"#{type}_probabilities").should == properties[:probabilities]
+    end
+  end
+
+end

metadata

@@ -1,45 +1,39 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: dna
-version: !ruby/object:Gem::Version 
-  hash: 31
+version: !ruby/object:Gem::Version
+  version: 0.2.0
   prerelease: 
-  segments: 
-  - 0
-  - 1
-  - 2
-  version: 0.1.2
 platform: ruby
-authors: 
+authors:
 - Austin G. Davis-Richardson
 autorequire: 
 bindir: bin
 cert_chain: []
-
-date: 2012-08-17 00:00:00 Z
-dependencies: 
-- !ruby/object:Gem::Dependency 
-  type: :development
-  requirement: &id001 !ruby/object:Gem::Requirement 
-    none: false
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        hash: 3
-        segments: 
-        - 0
-        version: "0"
-  version_requirements: *id001
+date: 2012-12-14 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
   name: jeweler
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.8.4
+  type: :development
   prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.8.4
 description: Simple FASTA/FASTQ/QSEQ parser library for Ruby.
 email: harekrishna@gmail.com
-executables: 
+executables:
 - dna
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
+files:
 - .rspec
 - .rvmrc
 - .travis.yml
@@ -52,6 +46,7 @@ files:
 - lib/dna/parsers/fasta.rb
 - lib/dna/parsers/fastq.rb
 - lib/dna/parsers/qseq.rb
+- lib/dna/phred.rb
 - lib/dna/record.rb
 - lib/dna/version.rb
 - readme.md
@@ -61,40 +56,35 @@ files:
 - spec/data/test.fastq
 - spec/data/test.qseq
 - spec/dna_spec.rb
+- spec/phred_spec.rb
 - spec/record_spec.rb
 - spec/spec_helper.rb
 homepage: http://audy.github.com/dna
-licenses: 
+licenses:
 - MIT
 post_install_message: 
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
+required_ruby_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+      segments:
       - 0
-      version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
+      hash: -74029819579817896
+required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
-      - 0
-      version: "0"
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
 requirements: []
-
 rubyforge_project: 
-rubygems_version: 1.8.24
+rubygems_version: 1.8.23
 signing_key: 
 specification_version: 3
 summary: Simple FASTA/FASTQ/QSEQ parser library for Ruby
 test_files: []
-