bio-ngs 0.4.2.alpha.01 → 0.4.4.alpha.01

data/Gemfile

@@ -4,7 +4,6 @@ source "http://rubygems.org"
 #   gem "activesupport", ">= 2.3.5"
 
 gem "bio", ">= 1.4.2"
-gem "bio-bwa", ">= 0.2.2"
 gem "bio-samtools", ">= 0.3.2"
 gem "thor", ">= 0.14.6"
 gem "rubyvis", ">= 0.5.0"
@@ -15,13 +14,15 @@ gem "sqlite3", ">= 1.3.3"
 gem "bio-blastxmlparser"
 gem "progressbar",">= 0.9.0"
 gem "rake", "0.9.2.2"
+gem "json"
+gem "parallel"
 
 # Add dependencies to develop your gem here.
 # Include everything needed to run rake, tests, features, etc.
 group :development do
   gem "shoulda", ">= 0"
-  gem "bundler", "~> 1.0.0"
-  gem "jeweler", "~> 1.5.2"
+  gem "bundler", "~> 1.1.0"
+  gem "jeweler", "~> 1.8.3"
   gem "rcov", "~> 0.9.11"
   gem "bio", ">= 1.4.2"
   gem "thor", ">= 0.14.6"
@@ -31,11 +32,11 @@ group :development do
   gem "daemons", ">= 1.1.0"
   gem "bio-samtools", ">= 0.3.2"
   gem "ruby-ensembl-api", ">= 1.0.1"
-  gem 'bio-bwa', ">= 0.2.2"
   gem "activerecord",">= 3.0.5"
   gem "sqlite3", ">= 1.3.3"
   gem "bio-blastxmlparser"
   gem "progressbar",">= 0.9.0"
   gem "json"
   gem "rake", "0.9.2.2"
+  gem "parallel"
 end

data/Gemfile.lock

@@ -1,81 +1,83 @@
 GEM
   remote: http://rubygems.org/
   specs:
-    activemodel (3.2.1)
-      activesupport (= 3.2.1)
+    activemodel (3.2.2)
+      activesupport (= 3.2.2)
       builder (~> 3.0.0)
-    activerecord (3.2.1)
-      activemodel (= 3.2.1)
-      activesupport (= 3.2.1)
-      arel (~> 3.0.0)
+    activerecord (3.2.2)
+      activemodel (= 3.2.2)
+      activesupport (= 3.2.2)
+      arel (~> 3.0.2)
       tzinfo (~> 0.3.29)
-    activesupport (3.2.1)
+    activesupport (3.2.2)
       i18n (~> 0.6)
       multi_json (~> 1.0)
-    arel (3.0.0)
+    arel (3.0.2)
     bio (1.4.2)
     bio-blastxmlparser (1.0.1)
       bio-logger (>= 1.0.0)
       nokogiri (>= 1.5.0)
-    bio-bwa (0.2.2)
     bio-logger (1.0.0)
       log4r (>= 1.1.9)
-    bio-samtools (0.5.0)
+    bio-samtools (0.5.2)
       bio (>= 1.4.2)
       ffi
     builder (3.0.0)
     daemons (1.1.8)
     diff-lcs (1.1.3)
     ffi (1.0.11)
-    ffi (1.0.11-java)
     git (1.2.5)
     i18n (0.6.0)
-    jeweler (1.5.2)
-      bundler (~> 1.0.0)
+    jeweler (1.8.3)
+      bundler (~> 1.0)
       git (>= 1.2.5)
       rake
+      rdoc
     json (1.6.5)
-    json (1.6.5-java)
     log4r (1.1.10)
-    multi_json (1.0.4)
-    nokogiri (1.5.0)
-    nokogiri (1.5.0-java)
+    multi_json (1.1.0)
+    nokogiri (1.5.2)
+    parallel (0.5.16)
     progressbar (0.10.0)
     rake (0.9.2.2)
     rcov (0.9.11)
-    rcov (0.9.11-java)
-    rspec (2.8.0)
-      rspec-core (~> 2.8.0)
-      rspec-expectations (~> 2.8.0)
-      rspec-mocks (~> 2.8.0)
-    rspec-core (2.8.0)
-    rspec-expectations (2.8.0)
-      diff-lcs (~> 1.1.2)
-    rspec-mocks (2.8.0)
+    rdoc (3.12)
+      json (~> 1.4)
+    rspec (2.9.0)
+      rspec-core (~> 2.9.0)
+      rspec-expectations (~> 2.9.0)
+      rspec-mocks (~> 2.9.0)
+    rspec-core (2.9.0)
+    rspec-expectations (2.9.0)
+      diff-lcs (~> 1.1.3)
+    rspec-mocks (2.9.0)
     ruby-ensembl-api (1.0.1)
       activerecord
       bio (>= 1)
     rubyvis (0.5.1)
-    shoulda (2.11.3)
+    shoulda (3.0.1)
+      shoulda-context (~> 1.0.0)
+      shoulda-matchers (~> 1.0.0)
+    shoulda-context (1.0.0)
+    shoulda-matchers (1.0.0)
     sqlite3 (1.3.5)
     thor (0.14.6)
-    tzinfo (0.3.31)
+    tzinfo (0.3.32)
 
 PLATFORMS
-  java
   ruby
 
 DEPENDENCIES
   activerecord (>= 3.0.5)
   bio (>= 1.4.2)
   bio-blastxmlparser
-  bio-bwa (>= 0.2.2)
   bio-samtools (>= 0.3.2)
-  bundler (~> 1.0.0)
+  bundler (~> 1.1.0)
   daemons (>= 1.1.0)
   ffi (>= 1.0.6)
-  jeweler (~> 1.5.2)
+  jeweler (~> 1.8.3)
   json
+  parallel
   progressbar (>= 0.9.0)
   rake (= 0.9.2.2)
   rcov (~> 0.9.11)

data/README.rdoc

@@ -34,6 +34,9 @@ Most of this tasks create sub-processes to speed up conversions
 * biongs convert:qseq:fastq:by_lane_index LANE INDEX OUTPUT      # Convert the qseq from a line and index in a fastq file
 * biongs convert:qseq:fastq:samples_by_lane SAMPLES LANE OUTPUT  # Convert the qseqs for each sample in a specific lane. SAMPLES is an array of index codes separated by commas lane is an integer
 
+=== filter
+* biongs filter:by_list QUANTIFICATION LIST  # Extract from QUANTIFICATION rows with a key in LIST
+
 === quality
 
 * biongs quality:boxplot FASTQ_QUALITY_STATS                   # plot reads quality as boxplot

data/VERSION

@@ -1 +1 @@
-0.4.2.alpha.01
+0.4.4.alpha.01

data/bin/biongs

@@ -6,11 +6,6 @@ module Bio
     path = File.expand_path(File.dirname(__FILE__))
     $:<<File.join(File.dirname(path),"lib")
     require 'bio-ngs'
-
-    path = File.expand_path(File.dirname(__FILE__))
-    Dir.glob(File.join(File.dirname(path),"lib","tasks","*.thor")) do |thorfile|
-      Thor::Util.load_thorfile(thorfile)
-    end
     
     class Runner < Thor::Runner
       

data/bio-ngs.gemspec

@@ -9,7 +9,7 @@ Gem::Specification.new do |s|
 
   s.required_rubygems_version = Gem::Requirement.new("> 1.3.1") if s.respond_to? :required_rubygems_version=
   s.authors = ["Francesco Strozzi", "Raoul J.P. Bonnal"]
-  s.date = "2012-02-13"
+  s.date = "2012-03-19"
   s.description = "bio-ngs provides a framework for handling NGS data with BioRuby"
   s.email = "francesco.strozzi@gmail.com"
   s.executables = ["biongs"]
@@ -33,6 +33,7 @@ Gem::Specification.new do |s|
     "lib/bio/appl/ngs/bcl2qseq.rb",
     "lib/bio/appl/ngs/blast.rb",
     "lib/bio/appl/ngs/bowtie-inspect.rb",
+    "lib/bio/appl/ngs/bwa.rb",
     "lib/bio/appl/ngs/cufflinks.rb",
     "lib/bio/appl/ngs/cufflinks/iterators.rb",
     "lib/bio/appl/ngs/fastx.rb",
@@ -102,28 +103,12 @@ Gem::Specification.new do |s|
   s.require_paths = ["lib"]
   s.rubygems_version = "1.8.10"
   s.summary = "bio-ngs provides a framework for handling NGS data with BioRuby"
-  s.test_files = [
-    "spec/converter_qseq_spec.rb",
-    "spec/quality_spec.rb",
-    "spec/sff_extract_spec.rb",
-    "spec/spec_helper.rb",
-    "spec/tophat_spec.rb",
-    "spec/utils_spec.rb",
-    "test/helper.rb",
-    "test/test_bio-ngs.rb",
-    "test/test_db.rb",
-    "test/test_homology.rb",
-    "test/test_ngs.rb",
-    "test/test_ontology.rb",
-    "test/test_utils.rb"
-  ]
 
   if s.respond_to? :specification_version then
     s.specification_version = 3
 
     if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
       s.add_runtime_dependency(%q<bio>, [">= 1.4.2"])
-      s.add_runtime_dependency(%q<bio-bwa>, [">= 0.2.2"])
       s.add_runtime_dependency(%q<bio-samtools>, [">= 0.3.2"])
       s.add_runtime_dependency(%q<thor>, [">= 0.14.6"])
       s.add_runtime_dependency(%q<rubyvis>, [">= 0.5.0"])
@@ -134,9 +119,11 @@ Gem::Specification.new do |s|
       s.add_runtime_dependency(%q<bio-blastxmlparser>, [">= 0"])
       s.add_runtime_dependency(%q<progressbar>, [">= 0.9.0"])
       s.add_runtime_dependency(%q<rake>, ["= 0.9.2.2"])
+      s.add_runtime_dependency(%q<json>, [">= 0"])
+      s.add_runtime_dependency(%q<parallel>, [">= 0"])
       s.add_development_dependency(%q<shoulda>, [">= 0"])
-      s.add_development_dependency(%q<bundler>, ["~> 1.0.0"])
-      s.add_development_dependency(%q<jeweler>, ["~> 1.5.2"])
+      s.add_development_dependency(%q<bundler>, ["~> 1.1.0"])
+      s.add_development_dependency(%q<jeweler>, ["~> 1.8.3"])
       s.add_development_dependency(%q<rcov>, ["~> 0.9.11"])
       s.add_development_dependency(%q<bio>, [">= 1.4.2"])
       s.add_development_dependency(%q<thor>, [">= 0.14.6"])
@@ -146,16 +133,15 @@ Gem::Specification.new do |s|
       s.add_development_dependency(%q<daemons>, [">= 1.1.0"])
       s.add_development_dependency(%q<bio-samtools>, [">= 0.3.2"])
       s.add_development_dependency(%q<ruby-ensembl-api>, [">= 1.0.1"])
-      s.add_development_dependency(%q<bio-bwa>, [">= 0.2.2"])
       s.add_development_dependency(%q<activerecord>, [">= 3.0.5"])
       s.add_development_dependency(%q<sqlite3>, [">= 1.3.3"])
       s.add_development_dependency(%q<bio-blastxmlparser>, [">= 0"])
       s.add_development_dependency(%q<progressbar>, [">= 0.9.0"])
       s.add_development_dependency(%q<json>, [">= 0"])
       s.add_development_dependency(%q<rake>, ["= 0.9.2.2"])
+      s.add_development_dependency(%q<parallel>, [">= 0"])
     else
       s.add_dependency(%q<bio>, [">= 1.4.2"])
-      s.add_dependency(%q<bio-bwa>, [">= 0.2.2"])
       s.add_dependency(%q<bio-samtools>, [">= 0.3.2"])
       s.add_dependency(%q<thor>, [">= 0.14.6"])
       s.add_dependency(%q<rubyvis>, [">= 0.5.0"])
@@ -166,9 +152,11 @@ Gem::Specification.new do |s|
       s.add_dependency(%q<bio-blastxmlparser>, [">= 0"])
       s.add_dependency(%q<progressbar>, [">= 0.9.0"])
       s.add_dependency(%q<rake>, ["= 0.9.2.2"])
+      s.add_dependency(%q<json>, [">= 0"])
+      s.add_dependency(%q<parallel>, [">= 0"])
       s.add_dependency(%q<shoulda>, [">= 0"])
-      s.add_dependency(%q<bundler>, ["~> 1.0.0"])
-      s.add_dependency(%q<jeweler>, ["~> 1.5.2"])
+      s.add_dependency(%q<bundler>, ["~> 1.1.0"])
+      s.add_dependency(%q<jeweler>, ["~> 1.8.3"])
       s.add_dependency(%q<rcov>, ["~> 0.9.11"])
       s.add_dependency(%q<bio>, [">= 1.4.2"])
       s.add_dependency(%q<thor>, [">= 0.14.6"])
@@ -178,17 +166,16 @@ Gem::Specification.new do |s|
       s.add_dependency(%q<daemons>, [">= 1.1.0"])
       s.add_dependency(%q<bio-samtools>, [">= 0.3.2"])
       s.add_dependency(%q<ruby-ensembl-api>, [">= 1.0.1"])
-      s.add_dependency(%q<bio-bwa>, [">= 0.2.2"])
       s.add_dependency(%q<activerecord>, [">= 3.0.5"])
       s.add_dependency(%q<sqlite3>, [">= 1.3.3"])
       s.add_dependency(%q<bio-blastxmlparser>, [">= 0"])
       s.add_dependency(%q<progressbar>, [">= 0.9.0"])
       s.add_dependency(%q<json>, [">= 0"])
       s.add_dependency(%q<rake>, ["= 0.9.2.2"])
+      s.add_dependency(%q<parallel>, [">= 0"])
     end
   else
     s.add_dependency(%q<bio>, [">= 1.4.2"])
-    s.add_dependency(%q<bio-bwa>, [">= 0.2.2"])
     s.add_dependency(%q<bio-samtools>, [">= 0.3.2"])
     s.add_dependency(%q<thor>, [">= 0.14.6"])
     s.add_dependency(%q<rubyvis>, [">= 0.5.0"])
@@ -199,9 +186,11 @@ Gem::Specification.new do |s|
     s.add_dependency(%q<bio-blastxmlparser>, [">= 0"])
     s.add_dependency(%q<progressbar>, [">= 0.9.0"])
     s.add_dependency(%q<rake>, ["= 0.9.2.2"])
+    s.add_dependency(%q<json>, [">= 0"])
+    s.add_dependency(%q<parallel>, [">= 0"])
     s.add_dependency(%q<shoulda>, [">= 0"])
-    s.add_dependency(%q<bundler>, ["~> 1.0.0"])
-    s.add_dependency(%q<jeweler>, ["~> 1.5.2"])
+    s.add_dependency(%q<bundler>, ["~> 1.1.0"])
+    s.add_dependency(%q<jeweler>, ["~> 1.8.3"])
     s.add_dependency(%q<rcov>, ["~> 0.9.11"])
     s.add_dependency(%q<bio>, [">= 1.4.2"])
     s.add_dependency(%q<thor>, [">= 0.14.6"])
@@ -211,13 +200,13 @@ Gem::Specification.new do |s|
     s.add_dependency(%q<daemons>, [">= 1.1.0"])
     s.add_dependency(%q<bio-samtools>, [">= 0.3.2"])
     s.add_dependency(%q<ruby-ensembl-api>, [">= 1.0.1"])
-    s.add_dependency(%q<bio-bwa>, [">= 0.2.2"])
     s.add_dependency(%q<activerecord>, [">= 3.0.5"])
     s.add_dependency(%q<sqlite3>, [">= 1.3.3"])
     s.add_dependency(%q<bio-blastxmlparser>, [">= 0"])
     s.add_dependency(%q<progressbar>, [">= 0.9.0"])
     s.add_dependency(%q<json>, [">= 0"])
     s.add_dependency(%q<rake>, ["= 0.9.2.2"])
+    s.add_dependency(%q<parallel>, [">= 0"])
   end
 end
 

data/ext/mkrf_conf.rb

@@ -58,7 +58,11 @@ path_binary = File.join(path_external,"bin","common")
       ["common", Bio::Ngs::Utils.os_type].each do |kind_software|
         path_binary = File.join("#{path_external}", 'bin', kind_software)
         versions[kind_software].each do |tool, info|
-          Bio::Ngs::Utils.compile_source(tool, info, "#{path_external}", path_binary) if info["type"]=="source"
+          if info["type"]=="source"
+            Bio::Ngs::Utils.compile_source(tool, info, "#{path_external}", path_binary)
+          elsif info["type"]=="make"
+            Bio::Ngs::Utils.just_make(tool, info, "#{path_external}", path_binary)
+          end
         end #versions
       end   
     end #compile

data/lib/bio-ngs.rb

@@ -11,6 +11,7 @@
 
 
 # External gems
+require 'parallel'
 require 'yaml'
 require 'json'
 require 'daemons'
@@ -20,7 +21,6 @@ require 'thor/group'
 require 'thor/runner'
 require 'bio-blastxmlparser'
 require 'bio'
-require 'bio-bwa'
 require 'active_record'
 require 'sqlite3'
 
@@ -49,7 +49,14 @@ require 'bio/appl/ngs/cufflinks'
 require 'bio/appl/ngs/samtools'
 require 'bio/appl/ngs/fastx'
 require 'bio/appl/ngs/blast'
+require 'bio/appl/ngs/bwa'
 
 # history 
 Bio::Ngs::HISTORY_FILE = Dir.pwd+"/.task-history.yml"
-Bio::Ngs::Utils.extend_system_path
+Bio::Ngs::Utils.extend_system_path
+
+# loading Tasks
+path = File.expand_path(File.dirname(__FILE__))
+Dir.glob(File.join(path,"tasks","*.thor")) do |thorfile|
+  Thor::Util.load_thorfile(thorfile)
+end

data/lib/bio/appl/ngs/bcftools.rb

@@ -0,0 +1,100 @@
+module Bio
+	module Ngs
+		module Bcftools
+
+			#Usage: bcftools view [options] <in.bcf> [reg]
+
+			#Input/output options:
+
+			 #      -A        keep all possible alternate alleles at variant sites
+			 #     -b        output BCF instead of VCF
+			 #      -D FILE   sequence dictionary for VCF->BCF conversion [null]
+			 #      -F        PL generated by r921 or before (which generate old ordering)
+			 #      -G        suppress all individual genotype information
+			 #      -l FILE   list of sites (chr pos) or regions (BED) to output [all sites]
+			 #      -L        calculate LD for adjacent sites
+			 #      -N        skip sites where REF is not A/C/G/T
+			 #      -Q        output the QCALL likelihood format
+			 #      -s FILE   list of samples to use [all samples]
+			 #      -S        input is VCF
+			 #      -u        uncompressed BCF output (force -b)
+
+			#Consensus/variant calling options:
+
+			 #      -c        SNP calling (force -e)
+			 #      -d FLOAT  skip loci where less than FLOAT fraction of samples covered [0]
+			 #      -e        likelihood based analyses
+			 #      -g        call genotypes at variant sites (force -c)
+			 #      -i FLOAT  indel-to-substitution ratio [-1]
+			 #      -I        skip indels
+			 #      -p FLOAT  variant if P(ref|D)<FLOAT [0.5]
+			 #      -P STR    type of prior: full, cond2, flat [full]
+			 #      -t FLOAT  scaled substitution mutation rate [0.001]
+			 #      -T STR    constrained calling; STR can be: pair, trioauto, trioxd and trioxs (see manual) [null]
+			 #      -v        output potential variant sites only (force -c)
+
+			#Contrast calling and association test options:
+
+			 #      -1 INT    number of group-1 samples [0]
+			 #      -C FLOAT  posterior constrast for LRT<FLOAT and P(ref|D)<0.5 [1]
+			 #      -U INT    number of permutations for association testing (effective with -1) [0]
+			 #      -X FLOAT  only perform permutations for P(chi^2)<FLOAT [0.01]
+			 class View
+			 	include Bio::Command::Wrapper	
+			 	set_program Bio::Ngs::Utils.binary("bcftools")
+			 	set_sub_program "view"
+			 	use_aliases
+			 	add_option :alternate, :type => :boolean, :aliases => "-A", :desc => "keep all possible alternate alleles at variant sites"
+			 	add_option :bcfout, :type => :boolean, :aliases => "-b", :desc => "output BCF instead of VCF"
+			 	add_option :dictionary, :type => :string, :aliases => "-D", :desc => "sequence dictionary for VCF->BCF conversion [null]"
+			 	add_option :suppress_genotype, :type => :boolean, :aliases => "-G", :desc => "suppress all individual genotype information"
+			 	add_option :positions, :type => :string, :aliases => "-l", :desc => "list of sites (chr pos) or regions (BED) to output [all sites]"
+			 	add_option :ld, :type => :boolean, :aliases => "-L", :desc => "calculate LD for adjacent sites"
+			 	add_option :skipN, :type => :boolean, :aliases => "-N", :desc => "skip sites where REF is not A/C/G/T"
+			 	add_option :qcall, :type => :boolean, :aliases => "-Q", :desc => "output the QCALL likelihood format"
+			 	add_option :samples, :type => :string, :aliases => "-s", :desc => "list of samples to use [all samples]"
+			 	add_option :vcfinput, :type => :boolean, :aliases => "-S", :desc => "input is VCF"
+			 	add_option :uncompressed, :type => :boolean, :aliases => "-u", :desc => "uncompressed BCF output (force -b)"
+			 	add_option :snpcall, :type => :boolean, :aliases => "-c", :desc => "SNP calling (force -e)"
+			 	add_option :skiploci, :type => :numeric, :aliases => "-d", :desc => "skip loci where less than FLOAT fraction of samples covered [0]"
+			 	add_option :likelihood, :type => :boolean, :aliases => "-e", :desc => "likelihood based analyses"
+			 	add_option :callgenotype, :type => :boolean, :aliases => "-g", :desc => "call genotypes at variant sites (force -c)"
+			 	add_option :indelratio, :type => :numeric, :aliases => "-i", :desc => "indel-to-substitution ratio [-1]"
+			 	add_option :skipindels, :type => :boolean, :aliases => "-I", :desc => "skip indels"
+			 	add_option :variant_probability, :type => :numeric, :aliases => "-p", :desc => "variant if P(ref|D)<FLOAT [0.5]"
+			 	add_option :prior_type, :type => :string, :aliases => "-P", :desc => "type of prior: full, cond2, flat [full]"
+			 	add_option :substitution_mutation_rate, :type => :numeric, :aliases => "-t", :desc => "scaled substitution mutation rate [0.001]"
+			 	add_option :contrained_calling, :type => :string, :aliases => "-T", :desc => "constrained calling; STR can be: pair, trioauto, trioxd and trioxs (see manual) [null]"
+			 	add_option :variant_sites, :type => :boolean, :aliases => "-v", :desc => "output potential variant sites only (force -c)"
+			 	add_option :group1, :type => :numeric, :aliases => "-1", :desc => "number of group-1 samples [0]"
+			 	add_option :posterior_contrast, :type => :numeric, :aliases => "-C", :desc => "posterior constrast for LRT<FLOAT and P(ref|D)<0.5 [1]"
+			 	add_option :association_permutations, :type => :numeric, :aliases => "-U", :desc => "number of permutations for association testing (effective with -1) [0]"
+			 	add_option :permutations_cutoff, :type => :numeric, :aliases => "-X", :desc => "only perform permutations for P(chi^2)<FLOAT [0.01]"
+			 end
+
+			 class Index
+			 	include Bio::Command::Wrapper
+			 	set_program Bio::Ngs::Utils.binary("bcftools")
+			 	set_sub_program "index"
+			 end
+
+			 class Cat
+			 	include Bio::Command::Wrapper
+			 	set_program Bio::Ngs::Utils.binary("bcftools")
+			 	set_sub_program "cat"
+			 end
+
+			 class Ld
+			 	include Bio::Command::Wrapper
+			 	set_program Bio::Ngs::Utils.binary("bcftools")
+			 	set_sub_program "ld"
+			 end
+
+			 class Ldpair
+			 	include Bio::Command::Wrapper
+			 	set_program Bio::Ngs::Utils.binary("bcftools")
+			 	set_sub_program "ldpair"
+			 end
+		end
+	end
+end