ms-ident 0.0.18 → 0.0.19

data/Rakefile

@@ -8,7 +8,7 @@ Jeweler::Tasks.new do |gem|
   gem.homepage = "http://github.com/jtprince/ms-ident"
   gem.license = "MIT"
   gem.summary = %Q{mspire library for working with mzIdentML and pepxml}
-  gem.description = %Q{mspire library for working with mzIdentML and pepxml}
+  gem.description = %Q{mspire library for working with mzIdentML, pepxml, and related.}
   gem.email = "jtprince@gmail.com"
   gem.authors = ["John T. Prince"]
   gem.rubyforge_project = 'mspire'

data/VERSION

@@ -1 +1 @@
-0.0.18
+0.0.19

data/lib/ms/ident/peptide/db.rb

@@ -5,7 +5,11 @@ module Ms ; end
 module Ms::Ident ; end
 module Ms::Ident::Peptide ; end
 
-module Ms::Ident::Peptide::Db 
+# the object itself is a modified Hash.
+# It is initialized with the database file and a protein array can be
+# retrieved with the #[] method given an amino acid sequence.  All other
+# methods are untested at this time and should be avoided!
+class Ms::Ident::Peptide::Db < Hash
   MAX_NUM_AA_EXPANSION = 3
 
   # the twenty standard amino acids
@@ -165,12 +169,24 @@ module Ms::Ident::Peptide::Db
     to_expand
   end
 
+  def initialize(db_file)
+    self.replace(YAML.load_file(db_file))
+  end
+
+  alias_method :old_bracket, '[]'.to_sym
+
+  # returns the protein id's as an array
+  def [](key)
+    old_bracket(key).chomp.split(PROTEIN_DELIMITER)
+  end
+
   # an object for on disk retrieval of db entries
   # proteins are returned as an array.
   # behaves much like a hash once it is opened.
   class IO
     include Enumerable
     def self.open(filename, &block)
+      raise ArgumentError unless block
       File.open(filename) do |io|
         block.call(self.new(io))
       end
@@ -192,9 +208,11 @@ module Ms::Ident::Peptide::Db
         @index[key] = [start, end_pos-start]
       end
     end
+
     # returns an array of proteins for the given key (peptide aaseq)
     def [](key)
       (start, length) = @index[key]
+      return nil unless start
       @io.seek(start)
       string = @io.read(length)
       string.chomp!

data/lib/ms/ident/peptide_hit/qvalue.rb

@@ -0,0 +1,55 @@
+require 'ms/ident/peptide_hit'
+
+module Ms ; end
+module Ms::Ident ; end
+
+class Ms::Ident::PeptideHit
+  module Qvalue
+    attr_accessor :qvalue
+    FILE_EXTENSION = '.phq.tsv'
+    FILE_DELIMITER = "\t"
+    HEADER = %w(aaseq charge qvalue)
+
+    class << self
+
+      # writes to the file, adding an extension
+      def to_phq(base, hits, qvalues=nil)
+        to_file(base + FILE_EXTENSION, hits)
+      end
+
+      # writes the peptide hits to a phq.tsv file. qvalues is a parallel array
+      # to hits that can provide qvalues if not inherent to the hits
+      # returns the filename.
+      def to_file(filename, hits, qvalues=[])
+        File.open(filename,'w') do |out|
+          out.puts HEADER.join(FILE_DELIMITER)
+          hits.zip(qvalues) do |hit, qvalue|
+            out.puts [hit.aaseq, hit.charge, qvalue || hit.qvalue].join(FILE_DELIMITER)
+          end
+        end
+        filename
+      end
+
+      # returns an array of PeptideHit objects from a phq.tsv
+      def from_file(filename)
+        peptide_hits = []
+        File.open(filename) do |io|
+          header = io.readline.chomp.split(FILE_DELIMITER)
+          raise "bad headers" unless header == HEADER 
+          io.each do |line|
+            line.chomp!
+            (aaseq, charge, qvalue) = line.split(FILE_DELIMITER)
+            ph = Ms::Ident::PeptideHit.new
+            ph.aaseq = aaseq ; ph.charge = charge.to_i ; ph.qvalue = qvalue.to_f
+            peptide_hits << ph
+          end
+        end
+        peptide_hits
+      end
+
+      alias_method :from_phq, :from_file
+
+    end
+  end # Qvalue
+  include Qvalue
+end # Peptide Hit

data/lib/ms/ident/peptide_hit.rb

@@ -0,0 +1,8 @@
+module Ms ; end
+module Ms::Ident ; end
+
+class Ms::Ident::PeptideHit
+  attr_accessor :aaseq
+  attr_accessor :charge
+  attr_accessor :proteins
+end

data/lib/ms/ident/protein.rb

@@ -1,70 +1,12 @@
 module Ms ; end
 module Ms::Ident ; end
 
-require 'set'
 
 module Ms::Ident::Protein
-  
-  class << self
-  end
-
   # gives the information up until the first space or carriage return.
   # Assumes the protein can respond_to? :reference
   def first_entry
     reference.split(/[\s\r]/)[0]
   end
-
-  PRIORITIZE_PROTEINS = lambda do |protein_group_and_peptide_hits|
-    peptide_hits = protein_group_and_peptide_hits.last
-    num_uniq_aaseqs = peptide_hits.map {|hit| hit.aaseq }.uniq.size
-    num_uniq_aaseqs_at_z = peptide_hits.map {|hit| [hit.aaseq, hit.charge] }.uniq.size
-    [num_uniq_aaseqs, num_uniq_aaseqs_at_z, peptide_hits.size]
-  end
-
-
-  module_function
-  # greedy algorithm to map a set of peptide_hits to protein groups.  each
-  # peptide hit should respond to :aaseq, :charge, :proteins if a block is
-  # given, yields a single argument: a doublet of protein_group and peptide
-  # set.  It expects a metric or array to sort by for creating greedy protein
-  # groups (the greediest proteins should sort to the back of the array).  if
-  # no block is given, the groups are sorted by [# uniq aaseqs, # uniq
-  # aaseq+charge, # peptide_hits] (see PRIORITIZE_PROTEINS).  Sets of
-  # peptide_hits and the objects returned by peptide_hit#proteins are used as
-  # hash keys.  As long as each peptide hit has a unique signature (like an
-  # id) then any object will work.  If they are Struct objects, you might
-  # consider redefining the #hash method to be object_id for performance and
-  # accuracy.
-  def peptide_hits_to_protein_groups(peptide_hits, &sort_by)
-    sort_by ||= PRIORITIZE_PROTEINS
-    # note to self: I wrote this in 2011, so I think I know what I'm doing now
-    protein_to_peptides = Hash.new {|h,k| h[k] = Set.new }
-    peptide_hits.each do |peptide_hit|
-      peptide_hit.proteins.each do |protein|
-        protein_to_peptides[protein] << peptide_hit
-      end
-    end
-    peptides_to_protein_group = Hash.new {|h,k| h[k] = [] }
-    protein_to_peptides.each do |protein, peptide_set|
-      peptides_to_protein_group[peptide_set] << protein
-    end
-    protein_group_to_peptides = peptides_to_protein_group.invert
-    greedy_first = protein_group_to_peptides.sort_by(&sort_by).reverse
-    accounted_for = Set.new
-    surviving_protein_groups = []
-    # we are discarding the subsumed sets, but we could get them with
-    # partition
-    greedy_first.select do |group, peptide_set|
-      has_an_unaccounted_peptide = false
-      peptide_set.each do |peptide_hit| 
-        unless accounted_for.include?(peptide_hit) 
-          has_an_unaccounted_peptide = true
-          accounted_for.add(peptide_hit)
-        end
-      end
-      has_an_unaccounted_peptide
-    end
-  end
-
 end
 

data/lib/ms/ident/protein_group.rb

@@ -0,0 +1,72 @@
+require 'set'
+
+module Ms
+  module Ident
+    # represents a group of proteins, typically indistinguishable in the
+    # experiment.
+    class ProteinGroup < Array
+      attr_accessor :peptide_hits
+
+      PRIORITIZE_PROTEINS = lambda do |protein_group_and_peptide_hits|
+        peptide_hits = protein_group_and_peptide_hits.last
+        num_uniq_aaseqs = peptide_hits.map {|hit| hit.aaseq }.uniq.size
+        num_uniq_aaseqs_at_z = peptide_hits.map {|hit| [hit.aaseq, hit.charge] }.uniq.size
+        [num_uniq_aaseqs, num_uniq_aaseqs_at_z, peptide_hits.size]
+      end
+
+      # greedy algorithm to map a set of peptide_hits to protein groups.  each
+      # peptide hit should respond to :aaseq, :charge, :proteins if a block is
+      # given, yields a single argument: a doublet of protein_group and peptide
+      # set.  It expects a metric or array to sort by for creating greedy protein
+      # groups (the greediest proteins should sort to the back of the array).  if
+      # no block is given, the groups are sorted by [# uniq aaseqs, # uniq
+      # aaseq+charge, # peptide_hits] (see PRIORITIZE_PROTEINS).  Sets of
+      # peptide_hits and the objects returned by peptide_hit#proteins are used as
+      # hash keys.  As long as each peptide hit has a unique signature (like an
+      # id) then any object will work.  If they are Struct objects, you might
+      # consider redefining the #hash method to be object_id for performance and
+      # accuracy.
+      #
+      # returns an array of ProteinGroup objects, each set with :peptide_hits
+      def self.peptide_hits_to_protein_groups(peptide_hits, &sort_by)
+        sort_by ||= PRIORITIZE_PROTEINS
+        # note to self: I wrote this in 2011, so I think I know what I'm doing now
+        protein_to_peptides = Hash.new {|h,k| h[k] = Set.new }
+        peptide_hits.each do |peptide_hit|
+          peptide_hit.proteins.each do |protein|
+            protein_to_peptides[protein] << peptide_hit
+          end
+        end
+        peptides_to_protein_group = Hash.new {|h,k| h[k] = [] }
+        protein_to_peptides.each do |protein, peptide_set|
+          peptides_to_protein_group[peptide_set] << protein
+        end
+        peptides_to_protein_group.each do |pephits,ar_of_prots| 
+          pg = Ms::Ident::ProteinGroup.new(ar_of_prots)
+          pg.peptide_hits = pephits
+          peptides_to_protein_group[pephits] = pg
+        end
+
+        protein_group_to_peptides = peptides_to_protein_group.invert
+        greedy_first = protein_group_to_peptides.sort_by(&sort_by).reverse
+
+        accounted_for = Set.new
+        # we are discarding the subsumed sets, but we could get them with
+        # partition
+        greedy_first.select! do |group, peptide_set|
+          has_an_unaccounted_peptide = false
+          peptide_set.each do |peptide_hit| 
+            unless accounted_for.include?(peptide_hit) 
+              has_an_unaccounted_peptide = true
+              accounted_for.add(peptide_hit)
+            end
+          end
+          group.peptide_hits = peptide_set if has_an_unaccounted_peptide
+          has_an_unaccounted_peptide
+        end
+        greedy_first.map(&:first)
+      end
+
+    end
+  end
+end

data/lib/ms/ident/protein_hit.rb

@@ -0,0 +1,17 @@
+
+module Ms ; end
+module Ms::Ident ; end
+
+class Ms::Ident::ProteinHit
+  attr_accessor :id
+  attr_accessor :seq
+  alias_method :sequence, :seq
+  alias_method :sequence=, :seq=
+  attr_accessor :peptide_hits
+
+  def initialize(id=nil)
+    @peptide_hits = []
+    @id = id
+  end
+end
+

data/spec/ms/ident/peptide/db_spec.rb

@@ -82,6 +82,12 @@ describe 'reading a peptide centric database' do
   outfiles = Ms::Ident::Peptide::Db.cmdline([FASTA_FILE])
   @outfile = outfiles.first
 
+  it 'creates a hash that can retrieve peptides as an array' do
+    hash = Ms::Ident::Peptide::Db.new(@outfile)
+    hash["AVTEQGHELSNEER"].enums %w(sp|P31946|1433B_HUMAN	sp|P31946-2|1433B_HUMAN)
+    hash["VRAAR"].enums ["tr|D3DX18|D3DX18_HUMAN"]
+  end
+
   it 'reads the file on disk with random access or is enumerable' do
     Ms::Ident::Peptide::Db::IO.open(@outfile) do |io|
       io["AVTEQGHELSNEER"].enums %w(sp|P31946|1433B_HUMAN	sp|P31946-2|1433B_HUMAN)

data/spec/ms/ident/{protein_spec.rb → protein_group_spec.rb}

@@ -1,6 +1,6 @@
 require 'spec_helper'
 
-require 'ms/ident/protein'
+require 'ms/ident/protein_group'
 
 PeptideHit = Struct.new(:aaseq, :charge, :proteins) do
   def inspect # easier to read output
@@ -36,22 +36,21 @@ describe 'creating minimal protein groups from peptide hits' do
   it 'is a greedy algorithm' do
     @prot_hits.each {|prthit| @prot_hits_hash[prthit.id].each {|pep| pep.proteins << prthit } }
     # big_guy has all the peptides, so it takes them all
-    reply = Ms::Ident::Protein.peptide_hits_to_protein_groups(@pep_hits)
-    reply.first.size.is 2 # the group and the peptide set
-    reply.first.first.size.is 1 # the group
-    reply.first.first.first.id.is 'big_guy'
+    protein_groups = Ms::Ident::ProteinGroup.peptide_hits_to_protein_groups(@pep_hits)
+    protein_groups.first.size.is 1# the group
+    protein_groups.first.first.id.is 'big_guy'
   end
 
   it 'removes proteins accounted for only as little pieces of larger proteins' do
     @prot_hits[1..-1].each {|prthit| @prot_hits_hash[prthit.id].each {|pep| pep.proteins << prthit } }
-    reply = Ms::Ident::Protein.peptide_hits_to_protein_groups(@pep_hits)
+    protein_groups = Ms::Ident::ProteinGroup.peptide_hits_to_protein_groups(@pep_hits)
     # no subsumed_by_medium
-    reply.map(&:first).any? {|protein_list| protein_list.any? {|v| v.id == 'subsumed_by_medium' }}.is false
+    protein_groups.any? {|prot_group| prot_group.any? {|v| v.id == 'subsumed_by_medium' }}.is false
   end
 
   it 'allows alternate sorting algorithms for greediness' do
     @prot_hits.each {|prthit| @prot_hits_hash[prthit.id].each {|pep| pep.proteins << prthit } }
-    reply = Ms::Ident::Protein.peptide_hits_to_protein_groups(@pep_hits) do |prot_and_peptide_hits|
+    prot_groups = Ms::Ident::ProteinGroup.peptide_hits_to_protein_groups(@pep_hits) do |prot_and_peptide_hits|
       # deliberate using a counterintuitive sorting method to give little guys
       # a chance
       -prot_and_peptide_hits.last.size
@@ -61,7 +60,7 @@ describe 'creating minimal protein groups from peptide hits' do
     # to add to the mix.  This demonstrates how proteins can be weighted in
     # different ways based on their peptide hits.
     seen = []
-    reply.each {|pair| pair.first.each {|prot| seen << prot.id } }
+    prot_groups.each {|pg| pg.each {|prot| seen << prot.id } }
     # big guy is completely accounted for in the now prioritized little guy
     # and medium guys, etc.
     seen.sort.is @prot_hits_hash.keys[1..-1].sort

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 0
-  - 18
-  version: 0.0.18
+  - 19
+  version: 0.0.19
 platform: ruby
 authors: 
 - John T. Prince
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-03-28 00:00:00 -06:00
+date: 2011-03-30 00:00:00 -06:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -97,7 +97,7 @@ dependencies:
         version: "0"
   type: :development
   version_requirements: *id006
-description: mspire library for working with mzIdentML and pepxml
+description: mspire library for working with mzIdentML, pepxml, and related.
 email: jtprince@gmail.com
 executables: []
 
@@ -116,6 +116,8 @@ files:
 - lib/ms/ident.rb
 - lib/ms/ident/peptide.rb
 - lib/ms/ident/peptide/db.rb
+- lib/ms/ident/peptide_hit.rb
+- lib/ms/ident/peptide_hit/qvalue.rb
 - lib/ms/ident/pepxml.rb
 - lib/ms/ident/pepxml/modifications.rb
 - lib/ms/ident/pepxml/msms_pipeline_analysis.rb
@@ -131,6 +133,8 @@ files:
 - lib/ms/ident/pepxml/search_summary.rb
 - lib/ms/ident/pepxml/spectrum_query.rb
 - lib/ms/ident/protein.rb
+- lib/ms/ident/protein_group.rb
+- lib/ms/ident/protein_hit.rb
 - lib/ms/ident/search.rb
 - schema/pepXML_v115.xsd
 - schema/pepXML_v19.xsd
@@ -138,7 +142,7 @@ files:
 - spec/ms/ident/pepxml/sample_enzyme_spec.rb
 - spec/ms/ident/pepxml/search_hit/modification_info_spec.rb
 - spec/ms/ident/pepxml_spec.rb
-- spec/ms/ident/protein_spec.rb
+- spec/ms/ident/protein_group_spec.rb
 - spec/spec_helper.rb
 - spec/tfiles/ms/ident/peptide/db/uni_11_sp_tr.fasta
 - spec/tfiles/ms/ident/peptide/db/uni_11_sp_tr.msd_clvg2.min_aaseq4.yml
@@ -179,5 +183,5 @@ test_files:
 - spec/ms/ident/pepxml/sample_enzyme_spec.rb
 - spec/ms/ident/pepxml/search_hit/modification_info_spec.rb
 - spec/ms/ident/pepxml_spec.rb
-- spec/ms/ident/protein_spec.rb
+- spec/ms/ident/protein_group_spec.rb
 - spec/spec_helper.rb