bio-polyploid-tools 0.1.0

data/bin/snps_between_bams.rb

@@ -0,0 +1,106 @@
+#!/usr/bin/env ruby
+
+require 'bio'
+require 'rubygems'
+require 'pathname'
+require 'bio-samtools'
+
+require 'set'
+
+$: << File.expand_path(File.dirname(__FILE__) + '/../lib')
+$: << File.expand_path('.')
+path=File.expand_path(File.dirname(__FILE__) + '/../lib/bioruby-polyploid-tools.rb')
+$stderr.puts "Loading: #{path}"
+require path
+
+
+
+fasta_db = Bio::DB::Fasta::FastaFile.new( ARGV[0])
+fasta_db.load_fai_entries
+bam1 =  Bio::DB::Sam.new({:fasta=>ARGV[0], :bam=>ARGV[1]})
+bam2 =  Bio::DB::Sam.new({:fasta=>ARGV[0], :bam=>ARGV[2]})
+
+
+output_prefix = ARGV[3]
+
+block_size=300
+
+min_cov = ARGV[4].to_i ? ARGV[4].to_i : 10
+chunk = ARGV[5].to_i
+chunk_size = ARGV[6].to_i
+
+
+
+
+main_table="#{output_prefix}_#{block_size}_#{min_cov}_table.#{chunk}.csv"
+
+table_file = File.open(main_table, "w")
+table_file.puts "gene\tlength\tsnps_1\tcalled_1\tsnps_per_#{block_size}_1\tsnps_2\tcalled_2\tsnps_per_#{block_size}_2\tsnps_tot\tsnps_per_1k_tot"
+
+hist_1= Hash.new(0)
+hist_2= Hash.new(0)
+
+fasta_file = File.open("#{output_prefix}_#{min_cov}.#{chunk}.fa", "w")
+i = -1
+min = chunk * chunk_size
+max = min + chunk_size
+
+fasta_db.index.entries.each do | r |
+  i = i  + 1
+  next if i < min or i >= max
+  #Np r.get_full_region
+  #container.process_region( { :region => r.get_full_region.to_s, :output_file => output_file } )
+  region=r.get_full_region
+
+
+  begin
+
+    cons_1 = bam1.consensus_with_ambiguities({:region=>region, :case=>true, :min_cov=>min_cov})
+    cons_2 = bam2.consensus_with_ambiguities({:region=>region, :case=>true, :min_cov=>min_cov})
+    if cons_1 != cons_2
+
+      snps_1 = cons_1.count_ambiguities
+      snps_2 = cons_2.count_ambiguities
+
+      called_1 = cons_1.upper_case_count
+      called_2 = cons_2.upper_case_count
+
+      snps_tot = Bio::Sequence.snps_between(cons_1, cons_2)
+
+      snps_per_1k_1   = (block_size * snps_1.to_f   ) / called_1
+      snps_per_1k_2   = (block_size * snps_2.to_f   ) / called_2
+      snps_per_1k_tot = (block_size * snps_tot.to_f ) / region.size
+
+      hist_1[snps_per_1k_1.to_i] += 1
+      hist_2[snps_per_1k_2.to_i] += 1
+
+      table_file.print "#{r.id}\t#{region.size}\t"
+      table_file.print "#{snps_1}\t#{called_1}\t#{snps_per_1k_1}\t"
+      table_file.print "#{snps_2}\t#{called_2}\t#{snps_per_1k_2}\t"
+      table_file.print "#{snps_tot}\t#{snps_per_1k_tot}\n"
+      fasta_file.puts ">#{r.id}_1"
+      fasta_file.puts "#{cons_1}"
+      fasta_file.puts ">#{r.id}_2"
+      fasta_file.puts "#{cons_2}"
+    end
+  rescue Exception => e
+    $stderr.puts "Unable to process #{region}: #{e.to_s}"
+  end
+end
+fasta_file.close
+table_file.close
+
+hist_table="#{output_prefix}_#{block_size}_#{min_cov}_hist.#{chunk}.csv"
+hist_file = File.open(hist_table, "w")
+
+all_keys = SortedSet.new(hist_1.keys)
+all_keys.merge(hist_2.keys)
+hist_file.puts "SNPs/#{block_size}\thist_1\thist_2\n"
+all_keys.each do |k|
+  hist_file.puts "#{k}\t#{hist_1[k]}\t#{hist_2[k]}"
+end
+
+hist_file.close
+
+
+

data/bio-polyploid-tools.gemspec

@@ -0,0 +1,139 @@
+# Generated by jeweler
+# DO NOT EDIT THIS FILE DIRECTLY
+# Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
+# -*- encoding: utf-8 -*-
+# stub: bio-polyploid-tools 0.1.0 ruby lib
+
+Gem::Specification.new do |s|
+  s.name = "bio-polyploid-tools"
+  s.version = "0.1.0"
+
+  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 = "2014-03-31"
+  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", "hexaploid_primers.rb", "homokaryot_primers.rb", "map_markers_to_contigs.rb", "markers_in_region.rb", "polymarker.rb", "snps_between_bams.rb"]
+  s.extra_rdoc_files = [
+    "README"
+  ]
+  s.files = [
+    "Gemfile",
+    "Gemfile.lock",
+    "README",
+    "Rakefile",
+    "VERSION",
+    "bin/bfr.rb",
+    "bin/count_variations.rb",
+    "bin/filter_blat_by_target_coverage.rb",
+    "bin/find_best_blat_hit.rb",
+    "bin/hexaploid_primers.rb",
+    "bin/homokaryot_primers.rb",
+    "bin/map_markers_to_contigs.rb",
+    "bin/markers_in_region.rb",
+    "bin/polymarker.rb",
+    "bin/snps_between_bams.rb",
+    "bio-polyploid-tools.gemspec",
+    "conf/defaults.rb",
+    "conf/primer3_config/dangle.dh",
+    "conf/primer3_config/dangle.ds",
+    "conf/primer3_config/interpretations/dangle_i.dh",
+    "conf/primer3_config/interpretations/dangle_i.ds",
+    "conf/primer3_config/interpretations/loops_i.dh",
+    "conf/primer3_config/interpretations/loops_i.ds",
+    "conf/primer3_config/interpretations/stack_i.dh",
+    "conf/primer3_config/interpretations/stack_i.ds",
+    "conf/primer3_config/interpretations/stackmm_i_mm.dh",
+    "conf/primer3_config/interpretations/stackmm_i_mm.ds",
+    "conf/primer3_config/interpretations/tetraloop_i.dh",
+    "conf/primer3_config/interpretations/tetraloop_i.ds",
+    "conf/primer3_config/interpretations/triloop_i.dh",
+    "conf/primer3_config/interpretations/triloop_i.ds",
+    "conf/primer3_config/interpretations/tstack2_i.dh",
+    "conf/primer3_config/interpretations/tstack2_i.ds",
+    "conf/primer3_config/interpretations/tstack_i.dh",
+    "conf/primer3_config/interpretations/tstack_i.ds",
+    "conf/primer3_config/interpretations/tstack_tm_inf_i.dh",
+    "conf/primer3_config/interpretations/tstack_tm_inf_i.ds",
+    "conf/primer3_config/loops.dh",
+    "conf/primer3_config/loops.ds",
+    "conf/primer3_config/stack.dh",
+    "conf/primer3_config/stack.ds",
+    "conf/primer3_config/stackmm.dh",
+    "conf/primer3_config/stackmm.ds",
+    "conf/primer3_config/tetraloop.dh",
+    "conf/primer3_config/tetraloop.ds",
+    "conf/primer3_config/triloop.dh",
+    "conf/primer3_config/triloop.ds",
+    "conf/primer3_config/tstack.dh",
+    "conf/primer3_config/tstack2.dh",
+    "conf/primer3_config/tstack2.ds",
+    "conf/primer3_config/tstack_tm_inf.ds",
+    "lib/bio/BFRTools.rb",
+    "lib/bio/BIOExtensions.rb",
+    "lib/bio/PolyploidTools/ChromosomeArm.rb",
+    "lib/bio/PolyploidTools/ExonContainer.rb",
+    "lib/bio/PolyploidTools/Marker.rb",
+    "lib/bio/PolyploidTools/PrimerRegion.rb",
+    "lib/bio/PolyploidTools/SNP.rb",
+    "lib/bio/PolyploidTools/SNPSequence.rb",
+    "lib/bio/SAMToolsExtensions.rb",
+    "lib/bio/db/exonerate.rb",
+    "lib/bio/db/fastadb.rb",
+    "lib/bio/db/primer3.rb",
+    "lib/bioruby-polyploid-tools.rb",
+    "test/data/BS00068396_51.fa",
+    "test/data/BS00068396_51_contigs.aln",
+    "test/data/BS00068396_51_contigs.dnd",
+    "test/data/BS00068396_51_contigs.fa",
+    "test/data/BS00068396_51_exonerate.tab",
+    "test/data/BS00068396_51_genes.txt",
+    "test/data/LIB1716.bam",
+    "test/data/LIB1716.bam.bai",
+    "test/data/LIB1719.bam",
+    "test/data/LIB1719.bam.bai",
+    "test/data/LIB1721.bam",
+    "test/data/LIB1721.bam.bai",
+    "test/data/LIB1722.bam",
+    "test/data/LIB1722.bam.bai",
+    "test/data/S22380157.fa",
+    "test/data/S22380157.fa.fai",
+    "test/data/Test3Aspecific.csv",
+    "test/data/Test3Aspecific_contigs.fa",
+    "test/data/patological_cases5D.csv",
+    "test/data/short_primer_design_test.csv",
+    "test/data/test_primer3_error.csv",
+    "test/data/test_primer3_error_contigs.fa",
+    "test/test_bfr.rb",
+    "test/test_exon_container.rb",
+    "test/test_exonearate.rb",
+    "test/test_snp_parsing.rb"
+  ]
+  s.homepage = "http://github.com/tgac/bioruby-polyploid-tools"
+  s.licenses = ["MIT"]
+  s.rubygems_version = "2.2.1"
+  s.summary = "Tool to work with polyploids, NGS and molecular biology"
+
+  if s.respond_to? :specification_version then
+    s.specification_version = 4
+
+    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-samtools>, ["= 0.6.2"])
+      s.add_runtime_dependency(%q<rake>, [">= 0"])
+      s.add_runtime_dependency(%q<jeweler>, [">= 0"])
+    else
+      s.add_dependency(%q<bio>, ["= 1.4.2"])
+      s.add_dependency(%q<bio-samtools>, ["= 0.6.2"])
+      s.add_dependency(%q<rake>, [">= 0"])
+      s.add_dependency(%q<jeweler>, [">= 0"])
+    end
+  else
+    s.add_dependency(%q<bio>, ["= 1.4.2"])
+    s.add_dependency(%q<bio-samtools>, ["= 0.6.2"])
+    s.add_dependency(%q<rake>, [">= 0"])
+    s.add_dependency(%q<jeweler>, [">= 0"])
+  end
+end
+

data/conf/defaults.rb

@@ -0,0 +1 @@
+path_to_chromosomes="/Users/ramirezr/Documents/PHD/201305_Databases/iwgcs"

data/conf/primer3_config/dangle.dh

@@ -0,0 +1,128 @@
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+-500
+4700
+-4100
+-3800
+0
+0
+0
+0
+0
+0
+0
+0
+-5900
+-2600
+-3200
+-5200
+0
+0
+0
+0
+0
+0
+0
+0
+-2100
+-200
+-3900
+-4400
+0
+0
+0
+0
+0
+0
+0
+0
+-700
+4400
+-1600
+2900
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+-2900
+-4100
+-4200
+-200
+0
+0
+0
+0
+0
+0
+0
+0
+-3700
+-4000
+-3900
+-4900
+0
+0
+0
+0
+0
+0
+0
+0
+-6300
+-4400
+-5100
+-4000
+0
+0
+0
+0
+0
+0
+0
+0
+200
+600
+-1100
+-6900
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0
+0

data/conf/primer3_config/dangle.ds

@@ -0,0 +1,128 @@
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-1.1
+14.2
+-13.1
+-12.6
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-16.5
+-7.4
+-10.4
+-15
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-3.9
+-0.1
+-11.2
+-13.1
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-0.8
+14.9
+-3.6
+10.4
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-7.6
+-13
+-15
+-0.5
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-10
+-11.9
+-10.9
+-13.8
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+-17.1
+-12.6
+-14
+-10.9
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+2.3
+3.3
+-1.6
+-20
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf
+inf