macroape 3.3.1 → 3.3.2

data/lib/macroape/exec/eval_alignment.rb

@@ -114,25 +114,12 @@ begin
   
   pwm_second.reverse_complement!  if reverse
   
-  first_pwm_alignment = '.' * [-shift, 0].max + '>' * pwm_first.length
-  second_pwm_alignment = '.' * [shift, 0].max + (orientation == :direct ? '>' : '<') * pwm_second.length
-  overlap = [pwm_first.length + [-shift,0].max, pwm_second.length + [shift,0].max].min - shift.abs
-  alignment_length = [first_pwm_alignment.length, second_pwm_alignment.length].max
-  (first_pwm_alignment.length...alignment_length).each{|i| first_pwm_alignment[i] = '.'}
-  (second_pwm_alignment.length...alignment_length).each{|i| second_pwm_alignment[i] = '.'}
-
-  cmp = Macroape::PWMCompareAligned.new(pwm_first.left_augment([-shift,0].max), 
-                              pwm_second.left_augment([shift,0].max))
-
+  cmp = Macroape::PWMCompareAligned.new(pwm_first, pwm_second, shift, orientation)
+ 
   first_threshold = pwm_first.threshold(pvalue)
   second_threshold = pwm_second.threshold(pvalue)
 
-  info = cmp.jaccard(first_threshold, second_threshold).merge(
-    text: "#{first_pwm_alignment}\n#{second_pwm_alignment}",
-    shift: shift,
-    orientation: orientation,
-    overlap: overlap,
-    alignment_length: alignment_length)
+  info = cmp.alignment_infos.merge( cmp.jaccard(first_threshold, second_threshold) )
   
   puts "#{info[:similarity]}\n#{info[:recognized_by_both]}\t#{info[:alignment_length]}\n#{info[:text]}\n#{info[:shift]}\t#{info[:orientation]}"
   

data/lib/macroape/exec/preprocess_collection.rb

@@ -74,18 +74,23 @@ begin
   Dir.glob(File.join(folder,'*')) do |filename|
     STDERR.puts filename unless silent
     pwm = Bioinform::PWM.new(File.read(filename))
+    pwm.name ||= File.basename(filename, File.extname(filename))
+    
+    # When support of onefile collections is introduced - then here should be check if name exists. 
+    # Otherwise it should skip motif and tell you about this
+    # Also two command line options to fail on skipping or to skip silently should be included
+    
     info = {rough: {}, precise: {}}
-    output = `ruby "#{File.join current_dir,'find_threshold.rb'}" #{filename} -p #{pvalues.join(' ')} -b #{background.join(' ')} -d #{rough_discretization}`.split("\n")
-    output.each do |line|
-      pvalue, threshold, real_pvalue = line.split.map(&:to_f)
-      info[:rough][pvalue] = threshold
-    end
+    pwm.background(background)
     
-    output = `ruby "#{File.join current_dir,'find_threshold.rb'}" #{filename} -p #{pvalues.join(' ')} -b #{background.join(' ')} -d #{precise_discretization}`.split("\n")
-    output.each do |line|
-      pvalue, threshold, real_pvalue = line.split.map(&:to_f)
-      info[:precise][pvalue] = threshold
+    pwm.discrete(rough_discretization).thresholds(*pvalues) do |pvalue, threshold, real_pvalue|
+      info[:rough][pvalue] = threshold / rough_discretization
     end
+
+    pwm.discrete(precise_discretization).thresholds(*pvalues) do |pvalue, threshold, real_pvalue|
+      info[:precise][pvalue] = threshold / precise_discretization
+    end
+    
     collection.add_pwm(pwm, info)
   end
   File.open(output_file,'w') do |f|

data/lib/macroape/exec/scan_collection.rb

@@ -94,7 +94,6 @@ begin
   
   similarities = {}
   precision_file_mode = {}
-  unnamed_index = 0
   
   collection.pwms.each_key do |name|
     pwm = collection.pwms[name]
@@ -102,7 +101,6 @@ begin
     STDERR.puts pwm.name unless silent
     cmp = Macroape::PWMCompare.new(query_pwm_rough, pwm.background(collection.background).discrete(collection.rough_discretization))
     info = cmp.jaccard(threshold, pwm_info[:rough][pvalue] * collection.rough_discretization)
-    name = pwm.name || "Unnamed #{unnamed_index += 1}"
     precision_file_mode[name] = :rough
     
     if precision_mode == :precise and info[:similarity] >= minimal_similarity

data/lib/macroape/pwm_compare.rb

@@ -7,35 +7,20 @@ module Macroape
     end
 
     def jaccard(threshold_first, threshold_second)
-      self.map_each_align do |align, alignment_info|
-        align.jaccard(threshold_first, threshold_second).merge(alignment_info)
-      end.max_by {|alignment_info| alignment_info[:similarity]}
+      self.map_each_alignment do |alignment|
+        alignment.alignment_infos.merge( alignment.jaccard(threshold_first, threshold_second) )
+      end.max_by {|alignment_infos| alignment_infos[:similarity] }
     end
-    
-    
-    def each
+
+    def each_alignment
       second_rc = second.reverse_complement
       (-second.length..first.length).to_a.product([:direct,:revcomp]) do |shift, orientation|
-        first_pwm_alignment = '.' * [-shift, 0].max + '>' * first.length
-        second_pwm_alignment = '.' * [shift, 0].max + (orientation == :direct ? '>' : '<') * second.length
-        overlap = [first.length + [-shift,0].max, second.length + [shift,0].max].min - shift.abs
-        alignment_length = [first_pwm_alignment.length, second_pwm_alignment.length].max
-        (first_pwm_alignment.length...alignment_length).each{|i| first_pwm_alignment[i] = '.'}
-        (second_pwm_alignment.length...alignment_length).each{|i| second_pwm_alignment[i] = '.'}
-        
-        yield(PWMCompareAligned.new(first.left_augment([-shift,0].max), 
-                                    (orientation == :direct ? second : second_rc).left_augment([shift,0].max)),
-              text: "#{first_pwm_alignment}\n#{second_pwm_alignment}",
-              shift: shift,
-              orientation: orientation,
-              overlap: overlap,
-              alignment_length: alignment_length
-              )
+        yield PWMCompareAligned.new(first, (orientation == :direct ? second : second_rc), shift, orientation)
       end
     end
-    include Enumerable
-    alias :each_align :each
-    alias :map_each_align :map
     
+    include Enumerable
+    alias_method :each, :each_alignment
+    alias_method :map_each_alignment, :map
   end
 end

data/lib/macroape/pwm_compare_aligned.rb

@@ -2,16 +2,89 @@ require 'macroape/aligned_pair_intersection'
 
 module Macroape
   class PWMCompareAligned
-    attr_reader :first, :second, :length
-    def initialize(first, second)
+    attr_reader :first, :second, :length, :shift, :orientation, :unaligned_first, :unaligned_second
+    def initialize(first, second, shift, orientation)
+      @unaligned_first, @unaligned_second = first, second
+      @shift, @orientation = shift, orientation
+      if shift > 0
+        first, second = first, second.left_augment(shift)
+      else
+        first, second = first.left_augment(-shift), second
+      end
       @length = [first.length, second.length].max
       @first = first.right_augment(@length - first.length)
       @second = second.right_augment(@length - second.length)
     end
     
-    #def discrete(rate)
-    #  PWMCompareAligned.new(first.discrete(rate), second.discrete(rate))
-    #end
+    def direct?
+      orientation == :direct
+    end
+    def revcomp?
+      orientation == :revcomp
+    end
+    
+    def overlap
+      length.times.count{|pos| first_overlaps?(pos) && second_overlaps?(pos) }
+    end
+    
+    def first_pwm_alignment
+      length.times.map do |pos|
+        if first_overlaps?(pos)
+          '>'
+        else
+          '.'
+        end
+      end.join
+    end
+    
+    def second_pwm_alignment
+      length.times.map do |pos|
+        if second_overlaps?(pos)
+          direct? ? '>' : '<'
+        else
+          '.'
+        end
+      end.join
+    end
+    
+    def alignment_infos
+      {shift: shift,
+      orientation: orientation,
+      text: "#{first_pwm_alignment}\n#{second_pwm_alignment}",
+      overlap: overlap,
+      alignment_length: length}
+    end
+    
+    def first_length
+      unaligned_first.length
+    end
+    def second_length
+      unaligned_second.length
+    end
+    
+    # whether first matrix overlap specified position
+    def first_overlaps?(pos)
+      return false unless pos >= 0 && pos < length
+      if shift > 0
+        pos < first_length
+      else
+        pos >= -shift && pos < -shift + first_length
+      end
+    end
+    
+    def second_overlaps?(pos)
+      return false unless pos >= 0 && pos < length
+      if shift > 0
+        pos >= shift && pos < shift + second_length
+      else
+        pos < second_length
+      end
+    end
+    
+=begin    
+    def discrete(rate)
+      PWMCompareAligned.new(first.discrete(rate), second.discrete(rate))
+    end
 
     def sort_pair_of_matrices_by(&block)
       mat = first.pwm.zip(second.pwm).sort_by(&block).transpose
@@ -26,6 +99,7 @@ module Macroape
     def permute_columns(permutation_index)
       PWMCompareAligned.new(first.permute(permutation_index), second.permute(permutation_index))
     end
+=end
 
     def jaccard(first_threshold, second_threshold)
       f = first.counts_by_thresholds(first_threshold).first

data/lib/macroape/threshold_by_pvalue.rb

@@ -56,6 +56,10 @@ module Bioinform
       scores
     end
     
+    def count_distribution
+      count_distribution_after_threshold(worst_score)
+    end
+
     def recalc_score_hash(scores, column, least_sufficient)
       new_scores = Hash.new(0)
       scores.each do |score, count|

data/lib/macroape/version.rb

@@ -1,3 +1,3 @@
 module Macroape
-  VERSION = "3.3.1"
+  VERSION = "3.3.2"
 end

data/macroape.gemspec

@@ -15,5 +15,5 @@ Gem::Specification.new do |gem|
   gem.require_paths = ["lib"]
   gem.version       = Macroape::VERSION
   
-  gem.add_dependency('bioinform', '>= 0.1.1')
+  gem.add_dependency('bioinform', '>= 0.1.2')
 end

data/test/data/test_collection/KLF4_f2.pat

@@ -1,4 +1,3 @@
-KLF4_f2
 0.30861857265872605 -2.254321000121579 0.13505703522674192 0.3285194224375633 
 -1.227018967707036 -4.814127713368663 1.3059890687390967 -4.908681463544344 
 -2.443469374521196 -4.648238485031404 1.3588686548279805 -4.441801801188402 

data/test/preprocess_collection_test.rb

@@ -5,7 +5,9 @@ require 'macroape'
 class TestPreprocessCollection < Test::Unit::TestCase
   def test_multipvalue_preproceessing
     system(Helpers.exec_cmd('preprocess_collection','test/data/test_collection -o test/data/test_collection.yaml.tmp -p 0.0005 0.0001 0.00005 --silent'))
-    assert_equal YAML.load_file('test/data/test_collection.yaml'), YAML.load_file('test/data/test_collection.yaml.tmp')
+    # Don't use YAML.load_file() instead of YAML.load(File.read()) because in ruby before v1.93 p194 
+    # it doesn't immediately release file descriptor (if I understood error right way) so File.delete fails
+    assert_equal YAML.load(File.read('test/data/test_collection.yaml')), YAML.load(File.read('test/data/test_collection.yaml.tmp'))
     File.delete 'test/data/test_collection.yaml.tmp'
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: macroape
 version: !ruby/object:Gem::Version
-  version: 3.3.1
+  version: 3.3.2
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-06-27 00:00:00.000000000 Z
+date: 2012-07-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bioinform
@@ -18,7 +18,7 @@ dependencies:
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
-        version: 0.1.1
+        version: 0.1.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -26,7 +26,7 @@ dependencies:
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
-        version: 0.1.1
+        version: 0.1.2
 description: Macroape is an abbreviation for MAtrix CompaRisOn by Approximate P-value
   Estimation. It's a bioinformatic tool for evaluating similarity measure and best
   alignment between a pair of Position Weight Matrices(PWM), finding thresholds by