macroape 3.3.2 → 3.3.3

data/lib/macroape/exec/eval_alignment.rb

@@ -16,10 +16,10 @@ Output format:
   <aligned 1st matrix>
   <aligned 2nd matrix>
   <shift> <orientation>
- 
-Examples:  
+
+Examples:
   ruby eval_alignment.rb motifs/KLF4_f2.pat motifs/SP1_f1.pat -1 direct -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
-     or on windows 
+     or on windows
   type motifs/SP1.pat | ruby eval_alignment.rb motifs/KLF4.pat .stdin 0 revcomp -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
      or in linux
   cat motifs/KLF4.pat motifs/SP1.pat | ruby eval_alignment.rb .stdin .stdin 3 direct -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
@@ -38,20 +38,20 @@ discretization = 10
 first_background = [1,1,1,1]
 second_background = [1,1,1,1]
 
-begin 
+begin
   first_file = ARGV.shift
   second_file = ARGV.shift
-  
+
   shift = ARGV.shift
   orientation = ARGV.shift
-  
+
   raise "You'd specify two input sources (each is filename or .stdin)" unless first_file and second_file
   raise 'You\'d specify shift' unless shift
   raise 'You\'d specify orientation' unless orientation
-  
+
   shift = shift.to_i
   orientation = orientation.to_sym
-  
+
   case orientation
     when :direct
       reverse = false
@@ -60,7 +60,7 @@ begin
     else
       raise 'Unknown orientation(direct/revcomp)'
   end
-  
+
 
   until ARGV.empty?
     case ARGV.shift
@@ -80,19 +80,19 @@ begin
         second_background = ARGV.shift(4).map(&:to_f)
     end
   end
-  raise 'background should be symmetric' unless first_background == first_background.reverse
+  raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless first_background == first_background.reverse
   raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless second_background == second_background.reverse
 
-  
+
   Macroape::MaxHashSizeSingle = 1000000 unless defined? Macroape::MaxHashSizeSingle
   Macroape::MaxHashSizeDouble = 1000 unless defined? Macroape::MaxHashSizeDouble
-  
+
 #  if first_file == '.stdin' || second_file == '.stdin'
 #    r_stream, w_stream = IO.pipe
 #    STDIN.readlines.each{|line| w_stream.write(line)}
 #    w_stream.close
 #  end
-  
+
   if first_file == '.stdin'
 #    r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
 #    pwm_first = Macroape::SingleMatrix.load_from_line_array(extracted_pwm).with_background(first_background).discrete(discretization)
@@ -100,7 +100,7 @@ begin
     raise "Error! File #{first_file} don't exist" unless File.exist?(first_file)
     pwm_first = Bioinform::PWM.new(File.read(first_file)).background(first_background).discrete(discretization)
   end
-  
+
   if second_file == '.stdin'
 #    r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
 #    pwm_second = Macroape::SingleMatrix.load_from_line_array(extracted_pwm).with_background(second_background).discrete(discretization)
@@ -108,21 +108,18 @@ begin
     raise "Error! File #{second_file} don't exist" unless File.exist?(second_file)
     pwm_second = Bioinform::PWM.new(File.read(second_file)).background(second_background).discrete(discretization)
   end
-  
+
 #  r_stream.close if first_file == '.stdin' || second_file == '.stdin'
-  
-  
-  pwm_second.reverse_complement!  if reverse
-  
+
   cmp = Macroape::PWMCompareAligned.new(pwm_first, pwm_second, shift, orientation)
- 
+
   first_threshold = pwm_first.threshold(pvalue)
   second_threshold = pwm_second.threshold(pvalue)
 
   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]}"
-  
+
 rescue => err
   STDERR.puts "\n#{err}\n#{err.backtrace.first(5).join("\n")}\n\nUse -help option for help\n"
 end

data/lib/macroape/exec/eval_similarity.rb

@@ -17,10 +17,10 @@ Output has format:
   <optimal alignment, the 1st matrix>
   <optimal alignment, the 2nd matrix>
   <shift> <orientation>
- 
-Examples:  
+
+Examples:
   ruby eval_similarity.rb motifs/KLF4.pat motifs/SP1.pat -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
-     or on windows 
+     or on windows
   type motifs/SP1.pat | ruby eval_similarity.rb motifs/KLF4.pat .stdin -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
      or in linux
   cat motifs/KLF4.pat motifs/SP1.pat | ruby eval_similarity.rb .stdin .stdin -p 0.0005 -d 100 -b 0.4 0.3 0.2 0.1
@@ -40,7 +40,7 @@ discretization = 10
 first_background = [1,1,1,1]
 second_background = [1,1,1,1]
 
-begin 
+begin
   first_file = ARGV.shift
   second_file = ARGV.shift
   raise "You'd specify two input sources (each is filename or .stdin)" unless first_file and second_file
@@ -60,22 +60,22 @@ begin
       when '-b1'
         first_background = ARGV.shift(4).map(&:to_f)
       when '-b2'
-        second_background = ARGV.shift(4).map(&:to_f)      
+        second_background = ARGV.shift(4).map(&:to_f)
     end
   end
   raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless first_background == first_background.reverse
   raise 'background should be symmetric: p(A)=p(T) and p(G) = p(C)' unless second_background == second_background.reverse
-  
+
   Macroape::MaxHashSizeSingle = 1000000 unless defined? Macroape::MaxHashSizeSingle
   Macroape::MaxHashSizeDouble = 1000 unless defined? Macroape::MaxHashSizeDouble
-  
-  
+
+
 #  if first_file == '.stdin' || second_file == '.stdin'
 #    r_stream, w_stream = IO.pipe
 #    STDIN.readlines.each{|line| w_stream.write(line)}
 #    w_stream.close
 #  end
-  
+
   if first_file == '.stdin'
 #    r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
 #    pwm_first = Macroape::SingleMatrix.load_from_line_array(extracted_pwm).with_background(first_background).discrete(discretization)
@@ -83,7 +83,7 @@ begin
     raise "Error! File #{first_file} don't exist" unless File.exist?(first_file)
     pwm_first = Bioinform::PWM.new(File.read(first_file)).background(first_background).discrete(discretization)
   end
-  
+
   if second_file == '.stdin'
 #    r_stream, w_stream, extracted_pwm = extract_pwm(r_stream, w_stream)
 #    pwm_second = Macroape::SingleMatrix.load_from_line_array(extracted_pwm).with_background(second_background).discrete(discretization)
@@ -91,7 +91,7 @@ begin
     raise "Error! File #{second_file} don't exist" unless File.exist?(second_file)
     pwm_second = Bioinform::PWM.new(File.read(second_file)).background(second_background).discrete(discretization)
   end
-  
+
   r_stream.close if first_file == '.stdin' || second_file == '.stdin'
 
   cmp = Macroape::PWMCompare.new(pwm_first, pwm_second)
@@ -100,9 +100,9 @@ begin
   second_threshold = pwm_second.threshold(pvalue)
 
   info = 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]}"
-  
+
 rescue => err
   STDERR.puts "\n#{err}\n#{err.backtrace.first(5).join("\n")}\n\nUse -help option for help\n"
 end

data/lib/macroape/exec/find_pvalue.rb

@@ -1,7 +1,7 @@
 help_string = %q{
 Command-line format:
 ruby find_pvalue.rb <pat-file> <threshold list> [options]
-        	or in linux 
+        	or in linux
 cat <pat-file> | ruby find_pvalue.rb .stdin <threshold> [options]
  	or on windows
 type <pat-file> | ruby find_pvalue.rb .stdin <threshold> [options]
@@ -16,7 +16,7 @@ Output format:
 	threshold_3 count_3  pvalue_3
 The results are printed out in the same order as in the given threshold list.
 
-Examples:  
+Examples:
   ruby find_pvalue.rb motifs/KLF4.pat 7.32 -d 1000 -b 0.2 0.3 0.2 0.3
            or on windows
   type motifs/KLF4.pat | ruby find_pvalue.rb .stdin 7.32 4.31 5.42
@@ -35,7 +35,7 @@ end
 discretization = 10000
 background = [1,1,1,1]
 thresholds = []
-begin 
+begin
   filename = ARGV.shift
 
   loop do
@@ -46,10 +46,10 @@ begin
       raise StopIteration
     end
   end
-  
+
   raise "No input. You'd specify input source: filename or .stdin" unless filename
   raise 'You should specify at least one threshold' if thresholds.empty?
-  
+
   until ARGV.empty?
     case ARGV.shift
       when '-b'
@@ -61,8 +61,8 @@ begin
     end
   end
   Macroape::MaxHashSizeSingle = 1000000 unless defined? Macroape::MaxHashSizeSingle
-  
-  
+
+
   if filename == '.stdin'
 #  TODO
   else

data/lib/macroape/exec/find_threshold.rb

@@ -14,9 +14,9 @@ Options:
 Output format:
  	requested_pvalue_1 threshold_1 achieved_pvalue_1
  	requested_pvalue_2 threshold_2 achieved_pvalue_2
-	 
-	 
-Example:  
+
+
+Example:
   ruby find_threshold.rb motifs/KLF4.pat -p 0.001 0.0001 0.0005 -d 1000 -b 0.4 0.3 0.2 0.1
 }
 
@@ -35,7 +35,7 @@ discretization = 10000
 begin
   filename = ARGV.shift
   raise "No input. You'd specify input source: filename or .stdin" unless filename
-  
+
   pvalues = []
   until ARGV.empty?
     case ARGV.shift
@@ -57,18 +57,18 @@ begin
       end
   end
   pvalues = default_pvalues if pvalues.empty?
-  
+
   Macroape::MaxHashSizeSingle = 1000000 unless defined? Macroape::MaxHashSizeSingle
-  
+
   if filename == '.stdin'
 ##  TODO
   else
     raise "Error! File #{filename} doesn't exist" unless File.exist?(filename)
     pwm = Bioinform::PWM.new( File.read(filename) )
   end
-  
+
   pwm.background(background)
-  
+
   pwm.discrete(discretization).thresholds(*pvalues) do |pvalue, threshold, real_pvalue|
     puts "#{pvalue}\t#{threshold / discretization}\t#{real_pvalue}"
   end

data/lib/macroape/exec/preprocess_collection.rb

@@ -10,7 +10,7 @@ Options:
   [--silent] - don't show current progress information during scan (by default this information's written into stderr)
 
 The tool stores preprocessed Macroape collection to the specified YAML-file.
- 
+
 Example:
   ruby preprocess_collection.rb ./motifs -p 0.001 0.0005 0.0001 -d 1 10 -b 0.2 0.3 0.2 0.3 -o collection.yaml
 }
@@ -34,7 +34,7 @@ begin
   folder = ARGV.shift
   raise "No input. You'd specify folder with pat-files" unless folder
   raise "Error! Folder #{folder} doesn't exist" unless Dir.exist?(folder)
-  
+
   pvalues = []
   silent = false
   until ARGV.empty?
@@ -64,7 +64,7 @@ begin
       end
   end
   pvalues = default_pvalues if pvalues.empty?
-  
+
   Macroape::MaxHashSizeSingle = 1000000 unless defined? Macroape::MaxHashSizeSingle
   Macroape::MaxHashSizeDouble = 1000 unless defined? Macroape::MaxHashSizeDouble
 
@@ -75,14 +75,14 @@ begin
     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. 
+
+    # 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: {}}
     pwm.background(background)
-    
+
     pwm.discrete(rough_discretization).thresholds(*pvalues) do |pvalue, threshold, real_pvalue|
       info[:rough][pvalue] = threshold / rough_discretization
     end
@@ -90,7 +90,7 @@ begin
     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

@@ -11,13 +11,13 @@ Options:
   [-c <similarity cutoff (minimal similarity to be included in output)> ] or [--all], '-c 0.05' by default
   [--precise [<level, minimal similarity to check on a more precise discretization level on the second pass>]], off by default, '--precise 0.01' if level is not set
   [--silent] - don't show current progress information during scan (by default this information's written into stderr)
-  
+
 Output format:
  <name> <similarity jaccard index> <shift> <overlap> <orientation> * [in case that result calculated on the second pass(in precise mode)]
-    Attention! Name can contain whitespace characters. 
+    Attention! Name can contain whitespace characters.
     Attention! The shift and orientation are reported for the collection matrix relative to the query matrix.
-	 
-Example:  
+
+Example:
   ruby scan_collection.rb motifs/KLF4.pat collection.yaml -p 0.005
             or in linux
   cat motifs/KLF4.pat | ruby scan_collection.rb .stdin collection.yaml -p 0.005 --precise 0.03
@@ -38,7 +38,7 @@ begin
   raise "No input. You'd specify input source for pat: filename or .stdin" unless filename
   raise "No input. You'd specify input file with collection" unless collection_file
   raise "Collection file #{collection_file} doesn't exist" unless File.exist?(collection_file)
-  
+
   pvalue = 0.0005
   cutoff = 0.05 # minimal similarity to output
   collection = YAML.load_file(collection_file)
@@ -65,7 +65,7 @@ begin
         silent = true
       when '--precise'
         precision_mode = :precise
-        begin 
+        begin
           Float(ARGV.first)
           minimal_similarity = ARGV.shift.to_f
         rescue
@@ -75,26 +75,26 @@ begin
   end
   Macroape::MaxHashSizeSingle = 1000000 unless defined? Macroape::MaxHashSizeSingle
   Macroape::MaxHashSizeDouble = 1000 unless defined? Macroape::MaxHashSizeDouble
-  
+
   raise "Thresholds for pvalue #{pvalue} aren't presented in collection (#{collection.pvalues.join(', ')}). Use one of listed pvalues or recalculate the collection with needed pvalue" unless collection.pvalues.include? pvalue
-  
+
   if filename == '.stdin'
 #    query_pwm = Macroape::SingleMatrix.load_from_stdin(STDIN)
   else
     raise "Error! File #{filename} doesn't exist" unless File.exist?(filename)
     query_pwm = Bioinform::PWM.new(File.read(filename))
   end
-  
-  
+
+
   query_pwm_rough = query_pwm.background(background_query).discrete(collection.rough_discretization)
   query_pwm_precise = query_pwm.background(background_query).discrete(collection.precise_discretization)
-  
+
   threshold = query_pwm_rough.threshold(pvalue)
   threshold_precise = query_pwm_precise.threshold(pvalue)
-  
+
   similarities = {}
   precision_file_mode = {}
-  
+
   collection.pwms.each_key do |name|
     pwm = collection.pwms[name]
     pwm_info = collection.infos[name]
@@ -102,7 +102,7 @@ begin
     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)
     precision_file_mode[name] = :rough
-    
+
     if precision_mode == :precise and info[:similarity] >= minimal_similarity
       cmp = Macroape::PWMCompare.new(query_pwm_precise, pwm.background(collection.background).discrete(collection.precise_discretization))
       info = cmp.jaccard(threshold_precise, pwm_info[:precise][pvalue] * collection.precise_discretization)
@@ -110,7 +110,7 @@ begin
     end
     similarities[name] = info
   end
-  
+
   puts "#pwm\tsimilarity\tshift\toverlap\torientation"
   similarities.sort_by do |name, info|
     info[:similarity]
@@ -118,7 +118,7 @@ begin
     precision_text = (precision_file_mode[name] == :precise) ? "\t*" : ""
     puts "#{name}\t#{info[:similarity]}\t#{info[:shift]}\t#{info[:overlap]}\t#{info[:orientation]}#{precision_text}" if info[:similarity] >= cutoff
   end
-  
+
 rescue => err
   STDERR.puts "\n#{err}\n#{err.backtrace.first(5).join("\n")}\n\nUse -help option for help\n"
 end

data/lib/macroape/pwm_compare.rb

@@ -13,12 +13,11 @@ module Macroape
     end
 
     def each_alignment
-      second_rc = second.reverse_complement
       (-second.length..first.length).to_a.product([:direct,:revcomp]) do |shift, orientation|
-        yield PWMCompareAligned.new(first, (orientation == :direct ? second : second_rc), shift, orientation)
+        yield PWMCompareAligned.new(first, second, shift, orientation)
       end
     end
-    
+
     include Enumerable
     alias_method :each, :each_alignment
     alias_method :map_each_alignment, :map

data/lib/macroape/pwm_compare_aligned.rb

@@ -2,31 +2,37 @@ require 'macroape/aligned_pair_intersection'
 
 module Macroape
   class PWMCompareAligned
-    attr_reader :first, :second, :length, :shift, :orientation, :unaligned_first, :unaligned_second
-    def initialize(first, second, shift, orientation)
-      @unaligned_first, @unaligned_second = first, second
+    attr_reader :first, :second, :length, :shift, :orientation, :first_length, :second_length
+    def initialize(first_unaligned, second_unaligned, shift, orientation)
       @shift, @orientation = shift, orientation
+
+      @first_length, @second_length = first_unaligned.length, second_unaligned.length
+      @length = self.class.calculate_alignment_length(@first_length, @second_length, @shift)
+
+      first, second = first_unaligned, second_unaligned
+      second = second.reverse_complement  if revcomp?
+      
       if shift > 0
-        first, second = first, second.left_augment(shift)
+        second = second.left_augment(shift)
       else
-        first, second = first.left_augment(-shift), second
+        first = first.left_augment(-shift)
       end
-      @length = [first.length, second.length].max
+
       @first = first.right_augment(@length - first.length)
       @second = second.right_augment(@length - second.length)
     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)
@@ -36,7 +42,7 @@ module Macroape
         end
       end.join
     end
-    
+
     def second_pwm_alignment
       length.times.map do |pos|
         if second_overlaps?(pos)
@@ -46,7 +52,7 @@ module Macroape
         end
       end.join
     end
-    
+
     def alignment_infos
       {shift: shift,
       orientation: orientation,
@@ -54,15 +60,8 @@ module Macroape
       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
+
+    # whether first matrix overlap specified position of alignment
     def first_overlaps?(pos)
       return false unless pos >= 0 && pos < length
       if shift > 0
@@ -71,7 +70,7 @@ module Macroape
         pos >= -shift && pos < -shift + first_length
       end
     end
-    
+
     def second_overlaps?(pos)
       return false unless pos >= 0 && pos < length
       if shift > 0
@@ -80,8 +79,8 @@ module Macroape
         pos < second_length
       end
     end
-    
-=begin    
+
+=begin
     def discrete(rate)
       PWMCompareAligned.new(first.discrete(rate), second.discrete(rate))
     end
@@ -104,13 +103,13 @@ module Macroape
     def jaccard(first_threshold, second_threshold)
       f = first.counts_by_thresholds(first_threshold).first
       s = second.counts_by_thresholds(second_threshold).first
-      if f == 0 or s == 0
+      if f == 0 || s == 0
         return {similarity: -1, tanimoto: -1, recognized_by_both: 0,
               recognized_by_first: f,
               recognized_by_second: s,
             }
       end
-      
+
       intersect = counts_for_two_matrices(first_threshold, second_threshold)
       intersect = Math.sqrt(intersect[0] * intersect[1])
       union = f + s - intersect
@@ -118,6 +117,15 @@ module Macroape
       { similarity: similarity,  tanimoto: 1.0 - similarity,  recognized_by_both: intersect,
         recognized_by_first: f,  recognized_by_second: s }
     end
-    
+
+    def self.calculate_alignment_length(first_len, second_len, shift)
+      if shift > 0
+        [first_len, second_len + shift].max
+      else
+        [first_len - shift, second_len].max
+      end
+    end
+
   end
+
 end