macroape 3.3.8 → 4.0.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 3037e164fd1b1c23bf40a9fca1d1da39737934a5
+  data.tar.gz: 4d4482d00ce0c76cbb47fe9d9eff53b48c1d2741
+SHA512:
+  metadata.gz: cc4176fe2b1d2f7b5bf4835b3612d316797f378dbf036bf0bba3135effe1cf9cbba1037ef87c7cffc24d0d2ef8d2ebf16fe5d585780445a1492fa795df44cb48
+  data.tar.gz: d86983eeb148235e1470dbfa7f674cf0e395129bd1836e34256b42bba06525f0881961cd8c6ca0e5e09df520786a8735e1ec2e10e76c4e59b8cfddb38b151685

data/README.md

@@ -17,7 +17,7 @@ Or install it yourself as:
     $ gem install macroape
 
 ## Usage
-  For more information read manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw (not last version but comprehensive description of approach)
+  For more information read manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw
 
 ## Basic usage as a command-line tool
   MacroAPE have 7 command line tools:
@@ -31,7 +31,7 @@ Or install it yourself as:
   * eval_alignment \<first PWM file\> \<second PWM file\> \<shift of second matrix\> \<orientation of second matrix(direct|revcomp)\>
   
 ### Tools for looking through collection for the motifs most similar to a query motif
-  * preprocess_collection \<folder with motif files\> [-o \<collection output file\>]
+  * preprocess_collection \<folder with motif files\> \<collection output file\>
   * scan_collection \<query PWM file\> \<collection file\>
   
 ### Tool for finding mutual alignment of several motifs relative to first(leader) motif. It's designed to use with sequence_logo to draw logos of clusters

data/TODO.txt

@@ -17,6 +17,8 @@ ToDo:
 8)(TODO: for theoretically consistency, while making small inconsistences to old calculations)
   When we work with strong threshold, we round matrix up(in order to overrate threshold comparing to real thus taking underrated pvalue) and take upper bound of discrete-thresholds fork.
   When we are estimating lower bound of threshold (weak threshold) we take lower bound of fork of discrete thresholds. But we should ALSO (not done yet) take matrix discreted down! This'd allow us give exact answer on a question in which range real threshold should lay with given P-value, now we correctly estimate only lower bound of threshold(upper bound of P-value)
+9) (may be) Option to specify predefined query motif threshold in scan_collection
+10) Fix Readme!
 
 Specs and tests:
   create spec on use of MaxHashSize, MaxHashSizeDouble

data/lib/macroape/cli/align_motifs.rb

@@ -1,12 +1,12 @@
-require 'docopt'
 require_relative '../../macroape'
+require 'shellwords'
 
 module Macroape
   module CLI
     module AlignMotifs
 
       def self.main(argv)
-        doc = <<-DOCOPT.strip_doc
+        doc = <<-EOS.strip_doc
           Align motifs tool.
           It takes motifs and builds alignment of each motif to the first (leader) motif.
 
@@ -16,38 +16,78 @@ module Macroape
             pwm_file_3  shift_3  orientation_3
 
           Usage:
-            align_motifs [options] <pm-files>...
+            #{run_tool_cmd} [options] <leader pm> <rest pm files>...
+              or
+            ls rest_pms/*.pm | #{run_tool_cmd} [options] <leader pm>
 
           Options:
-            -h --help       Show this screen.
-            --pcm           Use PCMs instead of PWMs as input
-        DOCOPT
+            [-p <P-value>]
+            [-d <discretization level>]
+            [--pcm] - treat the input file as Position Count Matrix. PCM-to-PWM transformation to be done internally.
+            [--boundary lower|upper] Upper boundary (default) means that the obtained P-value is greater than or equal to the requested P-value
+            [-b <background probabilities] ACGT - 4 numbers, comma-delimited(spaces not allowed), sum should be equal to 1, like 0.25,0.24,0.26,0.25
+        EOS
 
-        options = Docopt::docopt(doc, argv: argv)
+        if argv.empty? || ['-h', '--h', '-help', '--help'].any?{|help_option| argv.include?(help_option)}
+          STDERR.puts doc
+          exit
+        end
 
-        data_model = options['--pcm'] ? Bioinform::PCM : Bioinform::PWM
-        motif_files = options['<pm-files>']
-        leader = motif_files.first
-        background = [1,1,1,1]
+        leader_background = [1,1,1,1]
+        rest_motifs_background = [1,1,1,1]
         discretization = 1
         pvalue = 0.0005
+        max_hash_size = 10000000
+        max_pair_hash_size = 10000
+        pvalue_boundary = :upper
+
+        data_model = argv.delete('--pcm') ? Bioinform::PCM : Bioinform::PWM
+
+        while argv.first && argv.first.start_with?('-')
+          case argv.shift
+            when '-p'
+              pvalue = argv.shift.to_f
+            when '-d'
+              discretization = argv.shift.to_f
+            when '--max-hash-size'
+              max_hash_size = argv.shift.to_i
+            when '--max-2d-hash-size'
+              max_pair_hash_size = argv.shift.to_i
+            when '-b'
+              rest_motifs_background = leader_background = argv.shift.split(',').map(&:to_f)
+            when '-b1'
+              leader_background = argv.shift.split(',').map(&:to_f)
+            when '-b2'
+              rest_motifs_background = argv.shift.split(',').map(&:to_f)
+            when '--boundary'
+              pvalue_boundary = argv.shift.to_sym
+              raise 'boundary should be either lower or upper'  unless  pvalue_boundary == :lower || pvalue_boundary == :upper
+          end
+        end
 
-        shifts = {leader => [0,:direct]}
-        pwm_first = data_model.new(File.read(leader)).to_pwm
-        pwm_first.set_parameters(background: background).discrete!(discretization)
-        motif_files[1..-1].each do |motif_name|
+        leader_pwm_file = argv.shift
+        rest_pwms_file = argv
+        rest_pwms_file += $stdin.read.shellsplit  unless $stdin.tty?
+        rest_pwms_file.reject!{|filename| File.expand_path(filename) == File.expand_path(leader_pwm_file)}
+
+        shifts = []
+        shifts << [leader_pwm_file, 0, :direct]
+        pwm_first = data_model.new(File.read(leader_pwm_file)).to_pwm
+        pwm_first.set_parameters(background: leader_background, max_hash_size: max_hash_size).discrete!(discretization)
+
+        rest_pwms_file.each do |motif_name|
           pwm_second = data_model.new(File.read(motif_name)).to_pwm
-          pwm_second.set_parameters(background: background).discrete!(discretization)
-          info = Macroape::PWMCompare.new(pwm_first, pwm_second).jaccard_by_pvalue(pvalue)
-          shifts[motif_name] = [info[:shift], info[:orientation]]
+          pwm_second.set_parameters(background: rest_motifs_background, max_hash_size: max_hash_size).discrete!(discretization)
+          cmp = Macroape::PWMCompare.new(pwm_first, pwm_second).set_parameters(max_pair_hash_size: max_pair_hash_size)
+          info = cmp.jaccard_by_pvalue(pvalue)
+          shifts << [motif_name, info[:shift], info[:orientation]]
         end
 
-        shifts.each do |motif_name, (shift,orientation)|
+        shifts.each do |motif_name, shift,orientation|
           puts "#{motif_name}\t#{shift}\t#{orientation}"
         end
-
-      rescue Docopt::Exit => e
-        puts e.message
+      rescue => err
+        STDERR.puts "\n#{err}\n#{err.backtrace.first(5).join("\n")}\n\nUse --help option for help\n\n#{doc}"
       end
 
     end

data/lib/macroape/version.rb

@@ -1,4 +1,4 @@
 module Macroape
-  VERSION = "3.3.8"
+  VERSION = "4.0.0"
   STANDALONE = false
 end

data/macroape.gemspec

@@ -4,7 +4,7 @@ require File.expand_path('../lib/macroape/version', __FILE__)
 Gem::Specification.new do |gem|
   gem.authors       = ["Ilya Vorontsov"]
   gem.email         = ["prijutme4ty@gmail.com"]
-  gem.description   = %q{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 P-values and inside out and even searching a collection of motifs for the most similar ones. Used approach and application described in manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw}
+  gem.description   = %q{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 P-values and vice versa and even searching a collection of motifs for the most similar ones. Used approach and application described in manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw}
   gem.summary       = %q{PWM comparison tool using MACROAPE approach}
   gem.homepage      = "http://autosome.ru/macroape/"
 
@@ -15,6 +15,5 @@ Gem::Specification.new do |gem|
   gem.require_paths = ["lib"]
   gem.version       = Macroape::VERSION
   
-  gem.add_dependency('bioinform', '= 0.1.9')
-  gem.add_dependency('docopt', '= 0.5.0')
+  gem.add_dependency('bioinform', '~> 0.1.10')
 end

data/test/align_motifs_test.rb

@@ -21,4 +21,17 @@ class TestAlignmotifs < Test::Unit::TestCase
                   %w[SP1_f1_revcomp.pcm -1 revcomp]],
       Helpers.align_motifs_output('--pcm KLF4_f2.pcm  KLF3_f1.pcm  SP1_f1_revcomp.pcm')
   end
+  def test_names_from_stdin
+    assert_equal [%w[KLF4_f2.pwm 0 direct],
+                  %w[KLF3_f1.pwm -4 direct],
+                  %w[SP1_f1_revcomp.pwm -1 revcomp]],
+      Helpers.provide_stdin('KLF3_f1.pwm  SP1_f1_revcomp.pwm'){ Helpers.align_motifs_output('KLF4_f2.pwm') }
+  end
+  def test_names_from_stdin_duplicate_leader
+    assert_equal [%w[KLF4_f2.pwm 0 direct],
+                  %w[KLF3_f1.pwm -4 direct],
+                  %w[SP1_f1_revcomp.pwm -1 revcomp]],
+      Helpers.provide_stdin('KLF3_f1.pwm KLF4_f2.pwm SP1_f1_revcomp.pwm'){ Helpers.align_motifs_output('KLF4_f2.pwm') }
+  end
+
 end

metadata

@@ -1,52 +1,33 @@
 --- !ruby/object:Gem::Specification
 name: macroape
 version: !ruby/object:Gem::Version
-  version: 3.3.8
-  prerelease: 
+  version: 4.0.0
 platform: ruby
 authors:
 - Ilya Vorontsov
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-12-07 00:00:00.000000000 Z
+date: 2013-04-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bioinform
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
-    - - '='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: 0.1.9
+        version: 0.1.10
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
-    - - '='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: 0.1.9
-- !ruby/object:Gem::Dependency
-  name: docopt
-  requirement: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - '='
-      - !ruby/object:Gem::Version
-        version: 0.5.0
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - '='
-      - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 0.1.10
 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
-  P-values and inside out and even searching a collection of motifs for the most similar
+  P-values and vice versa and even searching a collection of motifs for the most similar
   ones. Used approach and application described in manual at https://docs.google.com/document/pub?id=1_jsxhMNzMzy4d2d_byAd3n6Szg5gEcqG_Sf7w9tEqWw
 email:
 - prijutme4ty@gmail.com
@@ -130,27 +111,26 @@ files:
 - test/test_helper.rb
 homepage: http://autosome.ru/macroape/
 licenses: []
+metadata: {}
 post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
-  none: false
   requirements:
-  - - ! '>='
+  - - '>='
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
-  none: false
   requirements:
-  - - ! '>='
+  - - '>='
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.24
+rubygems_version: 2.0.3
 signing_key: 
-specification_version: 3
+specification_version: 4
 summary: PWM comparison tool using MACROAPE approach
 test_files:
 - spec/count_distribution_spec.rb
@@ -190,3 +170,4 @@ test_files:
 - test/preprocess_collection_test.rb
 - test/scan_collection_test.rb
 - test/test_helper.rb
+has_rdoc: