bio-fasta_retrieve 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 5769b466a33fbea95e39d68e59bdcb05b3aa4909
+  data.tar.gz: f2cb672aa9e4cb7abde0dd7a57ac2423bf07a3e0
+SHA512:
+  metadata.gz: 9e01d0a422faf4ea23bedbe989e2db0c8e2fbe974106a49e4aa99ed294b430ed5c032c8964dbba5b02c894daf31bbf8047a22bef6d604315c859374ec3199b69
+  data.tar.gz: 4074fb0903cb3efbbd9baa6bc06cd7c3f05a6f34216b0fabefbfe11d0db5444787d7283cc71d2d98699d3698ac91f06ae6393a6cc8892b08af6194cd6462f3ea

data/lib/bio-fasta_retrieve.rb

@@ -0,0 +1,47 @@
+# Copyright::   Copyright (C) 2015
+#               Andrei Rozanski <rozanski.andrei@gmail.com>
+# License::     The Ruby License
+# == Description
+# This file containts DNA sequence retriever from UCSC that outputs Bio::Sequence object
+
+#Define module FastaRetriever
+module FastaRetriever
+  class Retriever
+
+	require 'bio' # For creation of fasta object.
+	require 'nokogiri' # For sequence retrieve and parse.
+ 	require 'open-uri' # For sequence retrieve.
+
+    def initialize;end
+
+    #input organism, chromosome, start, end separated by space (i.e. hg19 chr10 1000 2000)
+    def self.retrieve(organism,chr,start_coord,end_coord)
+      address="http://genome.ucsc.edu/cgi-bin/das/#{organism}/dna?segment=#{chr}:#{start_coord},#{end_coord}"
+      xml=ucsc_connect(address)
+      seq=compose_fasta(xml,organism,chr,start_coord,end_coord)       
+      puts seq
+    end # End retrieve
+
+    #connects to ucsc and retrieve xml with DNA sequence
+    def self.ucsc_connect(address)
+      xml = Nokogiri::XML(open(address))
+      xml_res=parse_xml(xml)
+      return xml_res
+    end # End ucsc_connect
+
+    #parse sequence from xml
+    def self.parse_xml(xml)
+      xml_res=xml.xpath("//DNA").text.tr("\n","")
+      return xml_res
+    end # End parse_xml
+
+    def self.compose_fasta(sequence,organism,chr,start_coord,end_coord)
+      seq=Bio::Sequence::NA.new(sequence)
+      seq_res=seq.to_fasta("#{organism}_#{chr}_#{start_coord}_#{end_coord}")
+      return seq_res
+    end # End compose_fasta
+
+  end # End Class Retriever
+end # End Module FastaRetriever
+
+ 

metadata

@@ -0,0 +1,44 @@
+--- !ruby/object:Gem::Specification
+name: bio-fasta_retrieve
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+platform: ruby
+authors:
+- Andrei Rozanski
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2015-08-13 00:00:00.000000000 Z
+dependencies: []
+description: Gem that aims to retrieve DNA sequence from UCSC
+email: rozanski.andrei@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/bio-fasta_retrieve.rb
+homepage: http://rubygems.org/gems/bio-fasta_retrieve.rb
+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: []
+rubyforge_project: 
+rubygems_version: 2.2.3
+signing_key: 
+specification_version: 4
+summary: Using DAS from UCSC, retrieve DNA sequences from given coordinates.
+test_files: []