ms-sequest 0.0.17 → 0.0.18

data/spec/ms/sequest/pepxml_spec.rb

@@ -241,71 +241,6 @@ describe Sequest::PepXML, " created from large bioworks.xml" do
 end
 
 
-
-describe Sequest::PepXML::Modifications do
-  before(:each) do
-    tf_params = Tfiles + "/bioworks32.params"
-    @params = Sequest::Params.new(tf_params)
-    # The params object here is completely unnecessary for this test, except
-    # that it sets up the mass table
-    @obj = Sequest::PepXML::Modifications.new(@params, "(M* +15.90000) (M# +29.00000) (S@ +80.00000) (C^ +12.00000) (ct[ +12.33000) (nt] +14.20000) ")
-  end
-  it 'creates a mod_symbols_hash' do
-    answ = {[:C, 12.0]=>"^", [:S, 80.0]=>"@", [:M, 29.0]=>"#", [:M, 15.9]=>"*", [:ct, 12.33]=>"[", [:nt, 14.2]=>"]"}
-    @obj.mod_symbols_hash.should == answ
-    ## need more here
-  end
-
-  it 'creates a ModificationInfo object given a special peptide sequence' do
-    mod_string = "(M* +15.90000) (M# +29.00000) (S@ +80.00000) (C^ +12.00000) (ct[ +12.33000) (nt] +14.20000) "
-    @params.diff_search_options = "15.90000 M 29.00000 M 80.00000 S 12.00000 C"
-    @params.term_diff_search_options = "14.20000 12.33000"
-    mod = Sequest::PepXML::Modifications.new(@params, mod_string)
-    ## no mods
-    peptide = "PEPTIDE"
-    mod.modification_info(peptide).should be_nil
-    peptide = "]M*EC^S@IDM#M*EMSCM["
-    modinfo = mod.modification_info(peptide)
-    modinfo.modified_peptide.should == peptide
-    modinfo.mod_nterm_mass.should be_close(146.40054, 0.000001)
-    modinfo.mod_cterm_mass.should be_close(160.52994, 0.000001)
-  end
-
-end
-
-describe Sequest::PepXML::SearchHit::ModificationInfo do
-
-  before(:each) do
-    modaaobjs = [[3, 150.3], [6, 345.2]].map do |ar| 
-      Sequest::PepXML::SearchHit::ModificationInfo::ModAminoacidMass.new(ar)
-    end
-    hash = {
-      :mod_nterm_mass => 520.2,
-      :modified_peptide => "MOD*IFI^E&D",
-      :mod_aminoacid_masses => modaaobjs,
-    }
-    #answ = "<modification_info mod_nterm_mass=\"520.2\" modified_peptide=\"MOD*IFI^E&amp;D\">\n\t<mod_aminoacid_mass position=\"3\" mass=\"150.3\"/>\n\t<mod_aminoacid_mass position=\"6\" mass=\"345.2\"/>\n</modification_info>\n"
-    @obj = Sequest::PepXML::SearchHit::ModificationInfo.new(hash)
-  end
-
-  def _re(st)
-    /#{Regexp.escape(st)}/
-  end
-  
-  it 'can produce pepxml' do
-    answ = @obj.to_pepxml
-    answ.should =~ _re('<modification_info')
-    answ.should =~ _re(" mod_nterm_mass=\"520.2\"")
-    answ.should =~ _re(" modified_peptide=\"MOD*IFI^E&amp;D\"")
-    answ.should =~ _re("<mod_aminoacid_mass")
-    answ.should =~ _re(" position=\"3\"")
-    answ.should =~ _re(" mass=\"150.3\"")
-    answ.should =~ _re(" position=\"6\"") 
-    answ.should =~ _re(" mass=\"345.2\"")
-    answ.should =~ _re("</modification_info>")
-  end
-end
-
 describe 'bioworks file with modifications transformed into pepxml' do
 
   spec_large do

data/spec/ms/sequest/srf/pepxml_spec.rb

@@ -0,0 +1,84 @@
+require 'spec_helper'
+
+require 'ms/sequest/srf/pepxml'
+
+describe 'an Ms::Ident::Pepxml object from an srf file with modifications' do
+  before do
+    FileUtils.mkdir @out_path unless File.exist?(@out_path)
+  end
+  after do
+    FileUtils.rm_rf @out_path
+  end
+
+  @srf_file = SEQUEST_DIR + '/opd1_2runs_2mods/sequest331/020.srf'
+  @out_path = TESTFILES + '/tmp'
+  @srf = Ms::Sequest::Srf.new(@srf_file)
+
+  it 'produces xml with all the expected parts' do
+    tags = %w(msms_pipeline_analysis msms_run_summary sample_enzyme specificity search_summary search_database enzymatic_search_constraint aminoacid_modification parameter spectrum_query search_result search_hit modification_info mod_aminoacid_mass search_score)
+    pepxml = @srf.to_pepxml(:verbose => false)
+    xml_string = pepxml.to_xml
+    tags.each do |tag|
+      xml_string.matches %r{<#{tag}}
+    end
+  end
+
+  # takes an xml string of attributes (' key="val" key2="val2" ') and a xml
+  # node that is expected to have those attributes
+  def has_attributes(node, string)
+    if node.nil?
+      raise "your xml node is nil!!!"
+    end
+    if node == []
+      raise "you gave me an empty array instead of a node"
+    end
+    # strips the tail end quote mark, also
+    string.strip!
+    string.chomp!('"')
+    string.split(/"\s+/).each do |str|
+      (key,val) = str.split('=',2)
+      val=val[1..-1] if val[0,1] == '"' 
+      if node[key] != val
+        puts "FAILING"
+        puts "EXPECT: #{key} => #{val} ACTUAL => #{val}"
+        puts "NODE KEYS: "
+        p node.keys
+        puts "NODE VALUES: "
+        p node.values
+      end
+      node[key].is val
+    end
+  end
+
+  it 'gets everything right' do
+    xml_string = @srf.to_pepxml(:verbose => false).to_xml
+    doc = Nokogiri::XML.parse(xml_string, nil, nil, Nokogiri::XML::ParseOptions::DEFAULT_XML | Nokogiri::XML::ParseOptions::NOBLANKS)
+
+    root = doc.root
+
+    root.name.is "msms_pipeline_analysis"
+    has_attributes( root, 'schemaLocation="http://regis-web.systemsbiology.net/pepXML /tools/bin/TPP/tpp/schema/pepXML_v115.xsd"' )
+    root['date'].nil?.is false
+    root['summary_xml'].matches "020.xml"
+    root.namespaces.is( {"xmlns" => "http://regis-web.systemsbiology.net/pepXML" } )
+
+    mrs_node = root.child
+    mrs_node.name.is 'msms_run_summary'
+    has_attributes( mrs_node, 'msManufacturer="Thermo" msModel="LCQ Deca XP" msIonization="ESI" msMassAnalyzer="Ion Trap" msDetector="UNKNOWN" raw_data=".mzXML"' )
+    se_node = mrs_node.child
+    se_node.name.is 'sample_enzyme'
+    has_attributes se_node, 'name="Trypsin"'
+    specificity_node = se_node.child
+    specificity_node.name.is 'specificity'
+    has_attributes specificity_node, 'cut="KR" no_cut="P" sense="C"'
+    search_summary_node = se_node.next_sibling
+    search_summary_node.name.is 'search_summary'
+    has_attributes search_summary_node, 'search_engine="SEQUEST" precursor_mass_type="average" fragment_mass_type="average" search_id="1"'
+    search_summary_node['base_name'].matches %r{sequest/opd1_2runs_2mods/sequest331/020$}
+    # TODO: expand the search summary check!
+    # TODO: finish testing other guys for accurcy
+  
+  end
+end
+
+

data/spec/ms/sequest/srf_spec.rb

@@ -1,5 +1,5 @@
-require File.expand_path( File.dirname(__FILE__) + '/../../spec_helper' )
-require File.expand_path( File.dirname(__FILE__) + '/srf_spec_helper' )
+require 'spec_helper'
+require 'ms/sequest/srf_spec_helper' # in spec/
 
 require 'ms/sequest/srf'
 
@@ -12,7 +12,7 @@ class Hash
     self.all? do |k,v|
       k = k.to_sym
       retval = 
-        if k == :peaks or k == :hits or k == :prots
+        if k == :peaks or k == :hits or k == :proteins
           obj.send(k).size == v
         elsif v.class == Float
           delta = 

data/spec/ms/sequest/srf_spec_helper.rb

@@ -85,7 +85,7 @@ module SRFHelper
       :rsp=>1,
       :ions_matched=>5,
       :ions_total=>35,
-      :prots=>1,
+      :proteins=>1,
       :deltamass=>-0.00579976654989878,
       :ppm=>5.16938660859491,
       :base_name=>"020",
@@ -109,7 +109,7 @@ module SRFHelper
       :rsp=>11,
       :ions_matched=>6,
       :ions_total=>40,
-      :prots=>1,
+      :proteins=>1,
       :deltamass=>0.00243330985608736,
       :ppm=>1.91215729542523,
       :base_name=>"020",

data/spec/spec_helper.rb

@@ -1,25 +1,19 @@
-
 require 'rubygems'
-require 'spec/more'
-
-Bacon.summary_on_exit
-
-# is this already defined??
-TESTFILES = File.expand_path(File.dirname(__FILE__)) + "/testfiles"
+require 'bundler'
 
 begin
-  require 'ms/testdata'
-rescue(LoadError)
-  puts %Q{
-Tests probably cannot be run because the submodules have
-not been initialized. Use these commands and try again:
- 
-% git submodule init
-% git submodule update
- 
-}
-  raise
+  Bundler.setup(:default, :development)
+rescue Bundler::BundlerError => e
+  $stderr.puts e.message
+  $stderr.puts "Run `bundle install` to install missing gems"
+  exit e.status_code
 end
+require 'ms/testdata'
+require 'spec/more'
+
+$LOAD_PATH.unshift(File.dirname(__FILE__))
+$LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
+
 
 def capture_stderr
   begin
@@ -31,3 +25,8 @@ def capture_stderr
   end
 end
 
+TESTFILES = File.dirname(__FILE__) + '/testfiles'
+SEQUEST_DIR = Ms::TESTDATA + '/sequest' 
+
+
+Bacon.summary_on_exit

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 0
-  - 17
-  version: 0.0.17
+  - 18
+  version: 0.0.18
 platform: ruby
 authors: 
 - John T. Prince
@@ -14,11 +14,11 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-02-24 00:00:00 -07:00
+date: 2011-03-08 00:00:00 -07:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
-  name: ms-core
+  name: ms-ident
   requirement: &id001 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
@@ -27,14 +27,29 @@ dependencies:
         segments: 
         - 0
         - 0
-        - 11
-        version: 0.0.11
+        - 17
+        version: 0.0.17
   type: :runtime
   prerelease: false
   version_requirements: *id001
 - !ruby/object:Gem::Dependency 
-  name: arrayclass
+  name: ms-core
   requirement: &id002 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        - 0
+        - 14
+        version: 0.0.14
+  type: :runtime
+  prerelease: false
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
+  name: arrayclass
+  requirement: &id003 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -46,10 +61,39 @@ dependencies:
         version: 0.1.0
   type: :runtime
   prerelease: false
-  version_requirements: *id002
+  version_requirements: *id003
+- !ruby/object:Gem::Dependency 
+  name: ms-msrun
+  requirement: &id004 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        - 3
+        - 3
+        version: 0.3.3
+  type: :runtime
+  prerelease: false
+  version_requirements: *id004
+- !ruby/object:Gem::Dependency 
+  name: trollop
+  requirement: &id005 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 1
+        - 16
+        version: "1.16"
+  type: :runtime
+  prerelease: false
+  version_requirements: *id005
 - !ruby/object:Gem::Dependency 
   name: ms-testdata
-  requirement: &id003 !ruby/object:Gem::Requirement 
+  requirement: &id006 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -61,10 +105,10 @@ dependencies:
         version: 0.1.1
   type: :development
   prerelease: false
-  version_requirements: *id003
+  version_requirements: *id006
 - !ruby/object:Gem::Dependency 
   name: spec-more
-  requirement: &id004 !ruby/object:Gem::Requirement 
+  requirement: &id007 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -74,10 +118,10 @@ dependencies:
         version: "0"
   type: :development
   prerelease: false
-  version_requirements: *id004
+  version_requirements: *id007
 - !ruby/object:Gem::Dependency 
   name: bundler
-  requirement: &id005 !ruby/object:Gem::Requirement 
+  requirement: &id008 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ~>
@@ -89,10 +133,10 @@ dependencies:
         version: 1.0.0
   type: :development
   prerelease: false
-  version_requirements: *id005
+  version_requirements: *id008
 - !ruby/object:Gem::Dependency 
   name: jeweler
-  requirement: &id006 !ruby/object:Gem::Requirement 
+  requirement: &id009 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ~>
@@ -104,10 +148,10 @@ dependencies:
         version: 1.5.2
   type: :development
   prerelease: false
-  version_requirements: *id006
+  version_requirements: *id009
 - !ruby/object:Gem::Dependency 
   name: rcov
-  requirement: &id007 !ruby/object:Gem::Requirement 
+  requirement: &id010 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -117,10 +161,11 @@ dependencies:
         version: "0"
   type: :development
   prerelease: false
-  version_requirements: *id007
+  version_requirements: *id010
 description: reads .SRF, .SQT and supports conversions
 email: jtprince@gmail.com
 executables: 
+- srf_to_pepxml.rb
 - srf_to_search.rb
 - srf_to_sqt.rb
 extensions: []
@@ -138,22 +183,29 @@ files:
 - README.rdoc
 - Rakefile
 - VERSION
+- bin/srf_to_pepxml.rb
 - bin/srf_to_search.rb
 - bin/srf_to_sqt.rb
 - lib/ms/sequest.rb
 - lib/ms/sequest/bioworks.rb
 - lib/ms/sequest/params.rb
 - lib/ms/sequest/pepxml.rb
+- lib/ms/sequest/pepxml/modifications.rb
+- lib/ms/sequest/pepxml/params.rb
 - lib/ms/sequest/sqt.rb
 - lib/ms/sequest/srf.rb
+- lib/ms/sequest/srf/pepxml.rb
+- lib/ms/sequest/srf/pepxml/sequest.rb
 - lib/ms/sequest/srf/search.rb
 - lib/ms/sequest/srf/sqt.rb
 - script/fasta_ipi_to_ncbi-ish.rb
 - spec/ms/sequest/bioworks_spec.rb
 - spec/ms/sequest/params_spec.rb
+- spec/ms/sequest/pepxml/modifications_spec.rb
 - spec/ms/sequest/pepxml_spec.rb
 - spec/ms/sequest/sqt_spec.rb
 - spec/ms/sequest/sqt_spec_helper.rb
+- spec/ms/sequest/srf/pepxml_spec.rb
 - spec/ms/sequest/srf/search_spec.rb
 - spec/ms/sequest/srf/sqt_spec.rb
 - spec/ms/sequest/srf_spec.rb
@@ -180,7 +232,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      hash: 1512994769580445168
+      hash: -2543561928840978627
       segments: 
       - 0
       version: "0"
@@ -202,9 +254,11 @@ summary: An mspire library supporting SEQUEST, Bioworks, SQT, etc
 test_files: 
 - spec/ms/sequest/bioworks_spec.rb
 - spec/ms/sequest/params_spec.rb
+- spec/ms/sequest/pepxml/modifications_spec.rb
 - spec/ms/sequest/pepxml_spec.rb
 - spec/ms/sequest/sqt_spec.rb
 - spec/ms/sequest/sqt_spec_helper.rb
+- spec/ms/sequest/srf/pepxml_spec.rb
 - spec/ms/sequest/srf/search_spec.rb
 - spec/ms/sequest/srf/sqt_spec.rb
 - spec/ms/sequest/srf_spec.rb