rbfam 0.0.2

data/lib/modules/alignment.rb

@@ -0,0 +1,60 @@
+module Rbfam
+  class Alignment
+    LINE_REGEXP = /^([\w\.]+)\/(\d+)\-(\d+)\s+([AUGC\.]+)$/
+    
+    attr_reader :family, :seed
+    
+    def initialize(family)
+      @family = family
+    end
+    
+    def entries(alignment = :seed)
+      @parsed_entries ||= pull_from_server(alignment).split(/\n/).reject do |line| 
+        line =~ /^#/
+      end.select do |line| 
+        line =~ LINE_REGEXP
+      end.map(&method(:parse_line)).tap do
+        @seed = alignment == :seed
+      end
+    end
+    
+    def save_entries!
+      entries.each { |sequence| sequence.save!(seed: seed) }
+    end
+    
+    def load_entries!(options = {})
+      Rbfam.script("sequences_in_mysql")
+      
+      @parsed_entries = SequenceTable.where({ family: family.family_name }.merge(options)).map do |entry|
+        entry.to_rbfam_sequence(family)
+      end
+    end
+    
+    private
+    
+    def pull_from_server(alignment)
+      url = "http://rfam.sanger.ac.uk/family/alignment/download/format?acc=%s&alnType=%s&nseLabels=1&format=pfam&download=0" % [
+        family.family_name,
+        alignment
+      ]
+      puts "GET: %s" % url unless @reponse
+      
+      @response ||= if (party = HTTParty.get(url)).response.code == "200"
+        puts "RESPONSE: 200 OK"
+        party.parsed_response
+      else
+        raise RuntimeError.new("HTTParty raised the following error when retrieving family %s: %s %s" % [
+          family_name,
+          party.response.code,
+          party.response.message
+        ])
+      end
+    end
+    
+    def parse_line(line)
+      line_match = line.match(LINE_REGEXP)
+      
+      Rbfam::Sequence.new(family, line_match[1], line_match[2].to_i, line_match[3].to_i, autoload: { length: 300, extend: 3 })
+    end
+  end
+end

data/lib/modules/family.rb

@@ -0,0 +1,60 @@
+module Rbfam
+  class Family
+    attr_reader :family_name
+    
+    class << self
+      def purine; new("RF00167"); end
+      def tpp;    new("RF00059"); end
+    end
+    
+    def initialize(family_name)
+      @family_name = family_name
+    end
+    
+    def alignment
+      Rbfam::Alignment.new(self)
+    end
+    
+    def entries
+      @parsed_entries ||= pull_from_server.split(/\n/).reject { |line| line =~ /^#/ }.map(&method(:parse_line))
+    end
+    
+    def load_entries!(options = {})
+      Rbfam.script("sequences_in_mysql")
+      
+      @parsed_entries = SequenceTable.where({ family: family_name }.merge(options)).map do |entry|
+        entry.to_rbfam_sequence(self)
+      end
+    end
+    
+    def save_entries!
+      entries.each(&:save!)
+    end
+    
+    private
+    
+    def pull_from_server
+      # It isn't the greatest design pattern to memoize a block where a branch has unmanaged exception raising, but for my uses that should never
+      # happen and needs to blow up hard if it does.
+      url = "http://rfam.sanger.ac.uk/family/regions?entry=%s" % family_name
+      puts "GET: %s" % url unless @reponse
+      
+      @reponse ||= if (party = HTTParty.get(url)).response.code == "200"
+        puts "RESPONSE: 200 OK"
+        party.parsed_response
+      else
+        raise RuntimeError.new("HTTParty raised the following error when retrieving family %s: %s %s" % [
+          family_name,
+          party.response.code,
+          party.response.message
+        ])
+      end
+    end
+    
+    def parse_line(line)
+      split_line = line.split(/\t/)
+      
+      Rbfam::Sequence.new(self, split_line[0], split_line[2].to_i, split_line[3].to_i, autoload: { length: 300, extend: 3 })
+    end
+  end
+end

data/lib/modules/sequence.rb

@@ -0,0 +1,95 @@
+module Rbfam
+  class Sequence
+    attr_reader :family, :accession, :from, :to
+
+    def initialize(family, accession, from, to, options = {})
+      @family, @accession, @from, @to = family, accession, from, to
+      
+      if options[:sequence]
+        @raw_sequence = options[:sequence]
+      end
+      
+      if options[:autoload]
+        sequence(options[:autoload].is_a?(Hash) ? options[:autoload] : {})
+      end
+    end
+    
+    def save!(options = {})
+      Rbfam.script("sequences_in_mysql")
+      
+      SequenceTable.create({ 
+        family:          family.family_name, 
+        accession:       accession, 
+        sequence:        sequence, 
+        sequence_length: sequence.length, 
+        from:            from, 
+        to:              to, 
+        seq_from:        up_coord + coord_window({ length: 300, extend: 3 }).min, 
+        seq_to:          up_coord + coord_window({ length: 300, extend: 3 }).max,
+        seed:            options[:seed]
+      })
+    end
+    
+    def up_coord
+      [from, to].min
+    end
+
+    def down_coord
+      [from, to].max
+    end
+    
+    def strand
+      plus_strand? ? :plus : :minus
+    end
+
+    def plus_strand?
+      to > from
+    end
+
+    def minus_strand?
+      !plus_strand?
+    end
+
+    def sequence(options = {})
+      @raw_sequence ||= Rbfam::Utils.rna_sequence_from_entrez(accession, up_coord, coord_window(options))
+      @raw_sequence   = minus_strand? ? @raw_sequence.complement : @raw_sequence
+    end
+    
+    alias :seq :sequence
+    
+    def mfe_structure
+      @mfe_structure ||= ViennaRna::Fold.run(seq).structure
+    end
+    
+    def fftbor
+      @fftbor ||= ViennaRna::Fftbor.run(seq: seq, str: mfe_structure)
+    end
+    
+    def coord_window(options = {})
+      range = 0..(down_coord - up_coord)
+      
+      if options[:length] && options[:extend]
+        if range.count < options[:length]
+          length_difference = options[:length] - range.count
+          
+          case [options[:extend], strand]
+          when [3, :plus], [5, :minus] then Range.new(range.min, range.max + length_difference)
+          when [5, :plus], [3, :minus] then Range.new(range.min - length_difference, range.max)
+          else puts "WARNING: value for :extend key in sequence retreival needs to be one of 5, 3 - found (%s)" % options[:extend]
+          end
+        else
+          puts "WARNING: %s %d-%d (%s) is length %d, but only %d nt. have been requested. Providing the full sequence anyways." % [
+            accession,
+            from,
+            to,
+            strand,
+            range.count,
+            options[:length]
+          ]
+        end
+      else
+        range
+      end
+    end
+  end
+end

data/lib/modules/utils.rb

@@ -0,0 +1,34 @@
+require "entrez"
+
+module Rbfam
+  module Utils
+    class << self
+      def rna_sequence_from_entrez(id, position, window, buffer_size = 0)
+        na_sequence_from_entrez(id, position, window, buffer_size).rna
+      end
+      
+      def na_sequence_from_entrez(id, position, window, buffer_size = 0)
+        Bio::Sequence::NA.new(sequence_from_entrez(id, position, Range.new(window.min - buffer_size, window.max + buffer_size)).seq)
+      end
+
+      def aa_sequence_from_entrez(id, position, window)
+        Bio::Sequence::AA.new(sequence_from_entrez(id, position, window).seq)
+      end
+
+      def sequence_from_entrez(id, position, window)
+        puts "Retrieving sequence from Entrez: using nuccore DB (id: #{id}, seq_start: #{position + window.min}, seq_stop: #{position + window.max})"
+        puts "> True starting position: #{position} with window #{window.min} to #{window.max}"
+
+        fasta = Entrez.EFetch("nuccore", {
+          id:        id, 
+          seq_start: position + window.min, 
+          seq_stop:  position + window.max, 
+          retmode:   :fasta, 
+          rettype:   :text
+        }).response.body
+
+        Bio::FastaFormat.new(fasta)
+      end
+    end
+  end
+end

data/lib/rbfam.rb

@@ -0,0 +1,14 @@
+require "bio"
+require "vienna_rna"
+require "httparty"
+require "active_support/inflector"
+
+module Rbfam
+  Dir[File.join(File.dirname(__FILE__), "/modules/*")].each do |file|
+    autoload File.basename(file, ".rb").camelize.to_sym, file
+  end
+  
+  def self.script(name)
+    require File.dirname(__FILE__) + "/scripts/#{File.basename(name, '.rb')}.rb"
+  end
+end

data/lib/scripts/sequences_in_mysql.rb

@@ -0,0 +1,49 @@
+require "mysql2"
+require "active_record"
+
+class Object; def this; self; end; end
+
+class SequenceTable < ActiveRecord::Base
+  self.table_name = "sequences"
+  
+  validates_uniqueness_of :accession, scope: [:seq_from, :seq_to]
+  
+  def self.connect
+    ActiveRecord::Base.establish_connection(config = { adapter: "mysql2", username: "root", reconnect: true })
+
+    unless ActiveRecord::Base.connection.execute("show databases").map { |i| i }.flatten.include?("rbfam")
+      ActiveRecord::Base.connection.create_database("rbfam")
+    end
+
+    ActiveRecord::Base.establish_connection(config.merge(database: "rbfam"))
+    
+    inline_rails if defined?(inline_rails)
+  end
+  
+  def to_rbfam_sequence(family)
+    Rbfam::Sequence.new(family, accession, from, to, sequence: sequence)
+  end
+end
+
+SequenceTable.connect
+
+class BuildSequence < ActiveRecord::Migration
+  def self.up
+    create_table :sequences do |table|
+      table.string  :family
+      table.string  :accession
+      table.text    :sequence
+      table.integer :sequence_length
+      table.integer :from
+      table.integer :to
+      table.integer :seq_from
+      table.integer :seq_to
+      table.boolean :seed, default: false
+      table.timestamps
+    end 
+  end
+end
+
+unless ActiveRecord::Base.connection.execute("show tables").map(&:this).flatten.include?("sequences")
+  BuildSequence.up
+end

metadata

@@ -0,0 +1,115 @@
+--- !ruby/object:Gem::Specification
+name: rbfam
+version: !ruby/object:Gem::Version
+  version: 0.0.2
+  prerelease: 
+platform: ruby
+authors:
+- Evan Senter
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2012-07-06 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: bio
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.4.2
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.4.2
+- !ruby/object:Gem::Dependency
+  name: entrez
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.5.8.1
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.5.8.1
+- !ruby/object:Gem::Dependency
+  name: httparty
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.8.3
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.8.3
+- !ruby/object:Gem::Dependency
+  name: vienna_rna
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.1.3
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.1.3
+description: Light wrapper for RFam data in Ruby.
+email: evansenter@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- ./lib/modules/alignment.rb
+- ./lib/modules/family.rb
+- ./lib/modules/sequence.rb
+- ./lib/modules/utils.rb
+- ./lib/rbfam.rb
+- ./lib/scripts/sequences_in_mysql.rb
+homepage: http://rubygems.org/gems/rbfam
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.24
+signing_key: 
+specification_version: 3
+summary: Bindings to Rfam.
+test_files: []
+has_rdoc: