bio-polyploid-tools 0.2.3 → 0.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 019bf8dc15f35de0be9a55567b8041f9b98ac326
-  data.tar.gz: 3e0a76bbefead5c5284c64a01b36645748a70098
+  metadata.gz: e22a969416448c22a39bc4ba2e0b019392a97ed4
+  data.tar.gz: 16fa6b61e54d6a6da47867826435194605ce0d22
 SHA512:
-  metadata.gz: 98e2d6c023ee8d89014efe65da619f0a98808c1540c3773aaef901de9f5c2338a3cc4645bdee1a3cdc430d525587c27d576d1f19e2ac8e59d7724a6efaac5901
-  data.tar.gz: 3d09c9a1972b7538eb160ee89786f9bdd3f8c52fec554da110222241767e7a17f6efd7f0f42a219fb06cd3689037fbee2ec88eeae0cb94e40333a5c491259421
+  metadata.gz: 96f1764086a490876516b7f9cf09c70d825733c518ecf4c21ad185ef24dda889e4398a3fe2a2f8e206feb056e2fe6c474fe7b938d03148c50a1a3ae2e6a6e048
+  data.tar.gz: 4792600da52fc6fdc31f15b2ed4c1e7db294aaf87d3c4538b6242951a24a6f4ba9a359f271a12d66af4c8e51f6eee746f588cb21d65a545216ffeb4bb05f8411

data/Gemfile

@@ -4,7 +4,7 @@ source "http://rubygems.org"
 #   gem "activesupport", ">= 2.3.5"
 
 gem "bio", ">= 1.4.3"
-gem "bio-samtools", ">= 2.0.3"
+gem "bio-samtools", ">= 2.0.4"
 gem "rake"
 gem "jeweler"
 

data/Gemfile.lock

@@ -4,7 +4,7 @@ GEM
     addressable (2.3.6)
     atomic (1.1.16)
     bio (1.4.3.0001)
-    bio-samtools (2.0.3)
+    bio-samtools (2.0.4)
       bio (>= 1.4.2)
       bio-svgenes (>= 0.4.1)
     bio-svgenes (0.4.1)
@@ -61,7 +61,7 @@ PLATFORMS
 
 DEPENDENCIES
   bio (>= 1.4.3)
-  bio-samtools (>= 2.0.3)
+  bio-samtools (>= 2.0.4)
   jeweler
   rake
   systemu (>= 2.5.2)

data/VERSION

@@ -1 +1 @@
-0.2.3
+0.3.1

data/bin/bfr.rb

@@ -55,7 +55,7 @@ OptionParser.new do |opts|
   end
   
   opts.on("-n", "--chunk FILE", "Sorted BAM file with the alginments from bulk1 2 (corresponding to the phenotype of parental 2)") do |o|
-    options[:chunk] = o.to_1
+    options[:chunk] = o.to_i
   end
   
     

data/bin/polymarker.rb

@@ -76,14 +76,22 @@ primer_3_input="#{output_folder}/primer_3_input_temp"
 primer_3_output="#{output_folder}/primer_3_output_temp"
 exons_filename="#{output_folder}/exons_genes_and_contigs.fa"
 output_primers="#{output_folder}/primers.csv"
+@status_file="#{output_folder}/status.txt"
 
 primer_3_config=File.expand_path(File.dirname(__FILE__) + '/../conf/primer3_config')
 model=options[:model] 
 
 
+def write_status(status)
+  f=File.open(@status_file, "a")
+  f.puts "#{Time.now.to_s},#{status}"
+  f.close
+end
+
 min_identity= 90
 snps = Array.new
 
+write_status "Loading Reference"
 #0. Load the fasta index 
 fasta_reference_db = nil
 if fasta_reference
@@ -93,9 +101,12 @@ if fasta_reference
 end
 
 
+
+
 #1. Read all the SNP files 
 #All the SNPs should be on the same chromosome as the first SNP. 
 #chromosome = nil
+write_status "Reading SNPs"
 File.open(test_file) do | f |
   f.each_line do | line |
     # p line.chomp!
@@ -121,7 +132,7 @@ end
 #1.1 Close fasta file
 #fasta_reference_db.close() if fasta_reference_db
 #2. Generate all the fasta files
-
+write_status "Writing sequences to align"
 written_seqs = Set.new
 file = File.open(temp_fasta_query, "w")
 snps.each do |snp|
@@ -135,6 +146,7 @@ file.close
 #3. Run exonerate on each of the possible chromosomes for the SNP
 #puts chromosome
 #chr_group = chromosome[0]
+write_status "Searching markers in genome"
 exo_f = File.open(exonerate_file, "w")
 contigs_f = File.open(temp_contigs, "w")
 filename=path_to_contigs 
@@ -165,6 +177,7 @@ contigs_f.close()
 #4. Load all the results from exonerate and get the input filename for primer3
 #Custom arm selection function that only uses the first two characters. Maybe
 #we want to make it a bit more cleaver
+write_status "Reading best alignment on each chromosome"
 arm_selection_first_two = lambda do | contig_name |
   ret = contig_name[0,2]       
   return ret
@@ -188,6 +201,9 @@ snps.each do |snp|
 end
 container.add_alignments({:exonerate_file=>exonerate_file, :arm_selection=>arm_selection_embl, :min_identity=>min_identity})
 
+
+#4.1 generating primer3 file
+write_status "Running primer3"
 file = File.open(exons_filename, "w")
 container.print_fasta_snp_exones(file)
 file.close
@@ -204,7 +220,9 @@ file.close
 
 Bio::DB::Primer3.run({:in=>primer_3_input, :out=>primer_3_output})
 
+
 #5. Pick the best primer and make the primer3 output
+write_status "Selecting best primers"
 kasp_container=Bio::DB::Primer3::KASPContainer.new
 kasp_container.line_1=snp_in
 kasp_container.line_2=original_name
@@ -217,3 +235,4 @@ kasp_container.add_primers_file(primer_3_output)
 header = "Marker,SNP,RegionSize,chromosome,total_contigs,contig_regions,SNP_type,#{snp_in},#{original_name},common,primer_type,orientation,#{snp_in}_TM,#{original_name}_TM,common_TM,selected_from,product_size"
 File.open(output_primers, 'w') { |f| f.write("#{header}\n#{kasp_container.print_primers}") }
 
+write_status "DONE"

data/bio-polyploid-tools.gemspec

@@ -2,16 +2,16 @@
 # DO NOT EDIT THIS FILE DIRECTLY
 # Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
 # -*- encoding: utf-8 -*-
-# stub: bio-polyploid-tools 0.2.3 ruby lib
+# stub: bio-polyploid-tools 0.3.1 ruby lib
 
 Gem::Specification.new do |s|
   s.name = "bio-polyploid-tools"
-  s.version = "0.2.3"
+  s.version = "0.3.1"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib"]
   s.authors = ["Ricardo H.  Ramirez-Gonzalez"]
-  s.date = "2014-04-27"
+  s.date = "2014-05-08"
   s.description = "Repository of tools developed in TGAC and Crop Genetics in JIC to work with polyploid wheat"
   s.email = "ricardo.ramirez-gonzalez@tgac.ac.uk"
   s.executables = ["bfr.rb", "count_variations.rb", "filter_blat_by_target_coverage.rb", "find_best_blat_hit.rb", "find_best_exonerate.rb", "hexaploid_primers.rb", "homokaryot_primers.rb", "map_markers_to_contigs.rb", "markers_in_region.rb", "polymarker.rb", "snps_between_bams.rb"]
@@ -102,6 +102,7 @@ Gem::Specification.new do |s|
     "test/data/S22380157.fa.fai",
     "test/data/Test3Aspecific.csv",
     "test/data/Test3Aspecific_contigs.fa",
+    "test/data/bfr_out_test.csv",
     "test/data/patological_cases5D.csv",
     "test/data/short_primer_design_test.csv",
     "test/data/test_primer3_error.csv",
@@ -121,20 +122,20 @@ Gem::Specification.new do |s|
 
     if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
       s.add_runtime_dependency(%q<bio>, [">= 1.4.3"])
-      s.add_runtime_dependency(%q<bio-samtools>, [">= 2.0.3"])
+      s.add_runtime_dependency(%q<bio-samtools>, [">= 2.0.4"])
       s.add_runtime_dependency(%q<rake>, [">= 0"])
       s.add_runtime_dependency(%q<jeweler>, [">= 0"])
       s.add_runtime_dependency(%q<systemu>, [">= 2.5.2"])
     else
       s.add_dependency(%q<bio>, [">= 1.4.3"])
-      s.add_dependency(%q<bio-samtools>, [">= 2.0.3"])
+      s.add_dependency(%q<bio-samtools>, [">= 2.0.4"])
       s.add_dependency(%q<rake>, [">= 0"])
       s.add_dependency(%q<jeweler>, [">= 0"])
       s.add_dependency(%q<systemu>, [">= 2.5.2"])
     end
   else
     s.add_dependency(%q<bio>, [">= 1.4.3"])
-    s.add_dependency(%q<bio-samtools>, [">= 2.0.3"])
+    s.add_dependency(%q<bio-samtools>, [">= 2.0.4"])
     s.add_dependency(%q<rake>, [">= 0"])
     s.add_dependency(%q<jeweler>, [">= 0"])
     s.add_dependency(%q<systemu>, [">= 2.5.2"])

data/lib/bio/PolyploidTools/ExonContainer.rb

@@ -81,11 +81,7 @@ module Bio::PolyploidTools
       end
     end
 
-    def primer_3_input_for_snp(snp)
-      gene_region = snp.covered_region
-      local_pos_in_gene = snp.local_position
-      puts ""
-    end
+   
 
     def fasta_string_for_snp(snp)
       gene_region = snp.covered_region
@@ -110,7 +106,6 @@ module Bio::PolyploidTools
         to_print[local_pos_in_gene] = to_print[local_pos_in_gene].upcase
         ret_str << to_print
       end
-      puts ret_str
       ret_str
     end
 

data/lib/bio/PolyploidTools/SNP.rb

@@ -22,15 +22,18 @@ module Bio::PolyploidTools
       reg_str.chomp!
       snp = SNP.new
       snp.gene, snp.original, snp.position, snp.snp, snp.chromosome = reg_str.split(",")
-      snp.position = snp.position.strip!.to_i
-      snp.original.upcase!.strip!
-      snp.snp.upcase!.strip!  
+      snp.position.strip!
+      snp.position =  snp.position.to_i
+      snp.original.upcase!
+      snp.original.strip!
+      snp.snp.upcase!
+      snp.snp.strip!  
       snp.chromosome.strip!
       snp.exon_list = Hash.new()
       snp.use_reference = false
       snp
     end
-    
+
     def initialize
        @genomes_count = 3 #TODO: if we want to use this with other polyploids, me need to set this as a variable in the main script. 
     end

data/lib/bio/PolyploidTools/SNPSequence.rb

@@ -1,5 +1,6 @@
 
 require_relative "SNP"
+require 'bio-samtools'
 module Bio::PolyploidTools
   class SNPSequenceException < RuntimeError 
   end

data/test/data/bfr_out_test.csv

@@ -0,0 +1,5 @@
+LIB1721	LIB1722	LIB1716	LIB1719	gnl|UG|Ta#S22380157	g	210	C	1.06	144	136	LIB1722	0.17	0.18	24	24
+LIB1721	LIB1722	LIB1716	LIB1719	gnl|UG|Ta#S22380157	g	297	A	0.8	172	164	LIB1722	0.2	0.16	34	26
+LIB1721	LIB1722	LIB1716	LIB1719	gnl|UG|Ta#S22380157	t	300	C	0.84	173	168	LIB1722	0.18	0.15	32	26
+LIB1721	LIB1722	LIB1716	LIB1719	gnl|UG|Ta#S22380157	g	645	A	0.4	78	87	LIB1721	0.28	0.71	22	62
+LIB1721	LIB1722	LIB1716	LIB1719	gnl|UG|Ta#S22380157	a	674	G	0.36	49	55	LIB1722	0.45	0.16	22	9

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bio-polyploid-tools
 version: !ruby/object:Gem::Version
-  version: 0.2.3
+  version: 0.3.1
 platform: ruby
 authors:
 - Ricardo H.  Ramirez-Gonzalez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-04-27 00:00:00.000000000 Z
+date: 2014-05-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bio
@@ -30,14 +30,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 2.0.3
+        version: 2.0.4
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 2.0.3
+        version: 2.0.4
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -182,6 +182,7 @@ files:
 - test/data/S22380157.fa.fai
 - test/data/Test3Aspecific.csv
 - test/data/Test3Aspecific_contigs.fa
+- test/data/bfr_out_test.csv
 - test/data/patological_cases5D.csv
 - test/data/short_primer_design_test.csv
 - test/data/test_primer3_error.csv