dirseq 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7b2362c43bff2a2cd43f0dc727bf8d713e4ab3c4
-  data.tar.gz: 7a56c16b7bfe38543c8d217d9307dc4f18ea5d21
+  metadata.gz: 57eb18eb4e971726d26849e01d3a4dd32597584d
+  data.tar.gz: f38d044727e8163873617fad111d3c244507fe71
 SHA512:
-  metadata.gz: 526abeb2d9b53cc0b0d26b3688d9b01ed44dea5d9b010d575b64922c8b3ae3d7097134db5fe23cff29855aa05f46449563ce7fcc575601a30e901782d90aaf7d
-  data.tar.gz: 8702f8ba681c9e161f97f76cfd763593ad266a77acf813a194ecbc7390ef08d46d6c0a68d6b60bebe46f2b9d86bc6bb0a1dbf51c0f1ae3861d33b27e775b1e59
+  metadata.gz: 8842180ae34fa546362dbcd63a0d07e9269b9903cdec1b078c10f5031b848d59b8f4db6d1e9708942beb84b92c639365049e8fd563c59723e861402b92771797
+  data.tar.gz: c2f035c3dba77ee68fd2861d30c4764eb61095dac654e402a47a7213ddfec651bfb401caf2b5fa015c60efa783c7a50f09a7f4573bcfb544c621bbd3e8d6b0e3

data/Gemfile

@@ -9,11 +9,10 @@ gem "bio", "~>1.4", ">=1.4.2"
 # Add dependencies to develop your gem here.
 # Include everything needed to run rake, tests, features, etc.
 group :development do
-  gem "shoulda", "~> 3.5"
-  gem "rdoc", "~> 3.12"
-  gem "simplecov", "~> 0.8"
+  #gem "shoulda", "~> 3.5"
+  #gem "simplecov", "~> 0.8"
   gem "jeweler", "~> 2.0"
   gem "bundler", "~> 1.6"
-  gem "rspec", "~> 2.99"
+  gem "rspec", "~> 3.0"
   gem 'pry', '~>0.10'
 end

data/README.md

@@ -13,8 +13,8 @@ Won't work just yet:
 gem install dirseq
 ```
 Requires:
-* samtools (tested with 0.1.19)
-* bedtools (tested with 2.20.1)
+* samtools (tested with 0.1.19 and 1.0+)
+* bedtools (tested with 2.24.0) - old versions won't work.
 * Ruby (tested with 2.1.1)
 
 ## Usage

data/VERSION

@@ -1 +1 @@
-0.1.0
+0.2.0

data/bin/dirseq

@@ -9,11 +9,21 @@ require 'tempfile'
 
 SCRIPT_NAME = File.basename(__FILE__); LOG_NAME = SCRIPT_NAME.gsub('.rb','')
 
+COVERAGE_COUNT_TYPE = 'coverage'
+COUNT_COUNT_TYPE = 'count'
+COUNT_TYPES = [
+  COUNT_COUNT_TYPE,
+  COVERAGE_COUNT_TYPE
+]
+
 # Parse command line options into the options hash
 options = {
   :ignore_directions => false,
   :logger => 'stderr',
   :log_level => 'info',
+  :count_type => COVERAGE_COUNT_TYPE,
+  :forward_read_only => false,
+  :accepted_feature_types => ['CDS'],
 }
 o = OptionParser.new do |opts|
   opts.banner = "
@@ -28,9 +38,20 @@ o = OptionParser.new do |opts|
     options[:gff] = arg
   end
   opts.separator "\nOptional parameters:\n\n"
+  opts.on("--forward-read-only", "consider only forward reads (i.e. read1) and ignore reverse reads. [default #{options[:forward_read_only]}]") do
+    options[:forward_ready_only] = true
+  end
   opts.on("--ignore-directions", "ignore directionality, give overall coverage [default: false i.e. differentiate between directions]") do |arg|
     options[:ignore_directions] = true
   end
+  opts.on("--measure-type TYPE", "what to count for each gene [options: #{COUNT_TYPES.join(', ')}][default: #{options[:count_type]}]") do |arg|
+    raise "Unexpected count type detected" if not COUNT_TYPES.include?(arg)
+    options[:count_type] = arg
+  end
+  opts.on("--accepted-feature-types TYPE", Array,
+          "Print only features of these type(s) [default #{options[:accepted_feature_types].join(',')}]") do |arg|
+    options[:accepted_feature_types] = set(arg)
+  end
 
   # logger options
   opts.separator "\nVerbosity:\n\n"
@@ -47,6 +68,11 @@ Bio::Log::CLI.logger(options[:logger]); Bio::Log::CLI.trace(options[:log_level])
 
 gff_file = options[:gff]
 bam_file = options[:bam]
+accepted_feature_types = options[:accepted_feature_types]
+
+if options[:count_type] != COVERAGE_COUNT_TYPE and options[:ignore_directions]
+  raise "ignore_directions + count_type != coverage is currently unsupported"
+end
 
 
 calculate_cov = lambda do |covs, num_covs|
@@ -76,14 +102,26 @@ get_covs = lambda do |cov_lines|
     #96 #coverage
     #0.0208333
     feat = splits[0..8]
+    feature_type = feat[2]
+    if not accepted_feature_types.include?(feature_type)
+      log.debug "Skipping feature as it is of type #{feature_type}"
+      next
+    end
     if feat != previous_feature
       feature_to_covs[previous_feature] = calculate_cov.call(covs, num_covs) unless previous_feature.nil?
       covs = []
       num_covs = 0
     end
-    num = splits[10].to_i
-    covs.push num*splits[9].to_i
-    num_covs += num
+    if splits.length == 13 # -hist
+      num = splits[10].to_i
+      covs.push num*splits[9].to_i
+      num_covs += num
+    elsif splits.length == 10 # -count
+      covs.push splits[9].to_i
+      num_covs += 1
+    else
+      raise "Unexpected bedtools output line: #{line}"
+    end
     previous_feature = feat
   end
   feature_to_covs[previous_feature] = calculate_cov.call(covs, num_covs)
@@ -93,10 +131,54 @@ end
 
 # Remove the ##FASTA and afterwards from the GFF file as this makes bedtools <2.25 fail
 # https://github.com/arq5x/bedtools2/issues/235#issuecomment-103776618
-no_fasta_gff = Tempfile.new('dirseq')
+no_fasta_gff = Tempfile.new(['dirseq','.gff3'])
 Bio::Commandeer.run "sed '/^##FASTA$/,$d' #{gff_file.inspect} > #{no_fasta_gff.path}", :log => log
 gff_file = no_fasta_gff.path
 
+
+
+
+
+# Find featureless contigs. Need to so that bedtools coverage -sorted does not complain
+if not File.exists?("#{bam_file}.bai")
+  raise "Input bam file must be indexed, but the index file does not exist"
+end
+
+chromosome_file = Tempfile.new('bam_contigs')
+log.info "Listing contigs in sorted order .."
+cmd = "samtools idxstats #{bam_file.inspect} |cut -f1,2 >#{chromosome_file.path.inspect}"
+Bio::Commandeer.run(cmd, :log => log)
+
+log.info "Finding featureless contigs"
+cmd = "grep -v '^#' #{gff_file.inspect} |cut -f1 |sort |uniq |grep -vFw -f /dev/stdin #{chromosome_file.path.inspect} |cut -f1"
+featureless_contigs = Bio::Commandeer.run(cmd, :log => log).lines.map(&:chomp).reject{ |ref| ref=='*' }
+log.info "Found #{featureless_contigs.length} featureless contigs"
+
+# Sort the GFF
+dummy_features = featureless_contigs.collect do |ref|
+  [ref,
+   'dirseq',
+   'misc_RNA',
+   '1',
+   '2',
+   '.',
+   '+',
+   '0',
+   "ID=#{ref}_dummy_feature"].join("\t")
+end
+sorted_gff_file_f = Tempfile.new(['sorted_gff','.gff3'])
+sorted_gff_file = sorted_gff_file_f.path
+Tempfile.open(["extra_features",'.gff']) do |ef|
+  ef.puts dummy_features.join("\n")
+  ef.close
+
+  cmd = "cat #{ef.path} #{gff_file.inspect} |bedtools sort -i /dev/stdin -faidx #{chromosome_file.path.inspect} >#{sorted_gff_file.inspect}"
+  log.info "Running bedtools sort"
+  Bio::Commandeer.run(cmd, :log => log)
+end
+
+
+
 covs_fwd = nil
 if options[:ignore_directions]
   cmd1 = "bedtools coverage -b #{bam_file.inspect} -a #{gff_file.inspect} -hist"
@@ -108,10 +190,14 @@ else
   # fwd read 1
   read1_flag = '-F128' #account for read1 in pair, as well as single reads mapping
   read2_flag = '-f128'
-  cmdf1 = "samtools view -u #{read1_flag} #{bam_file.inspect} |bedtools coverage -b /dev/stdin -a #{gff_file.inspect} -hist -s"
-  cmdf2 = "samtools view -u #{read2_flag} #{bam_file.inspect} |bedtools coverage -b /dev/stdin -a #{gff_file.inspect} -hist -s"
-  cmdr1 = "samtools view -u #{read1_flag} #{bam_file.inspect} |bedtools coverage -b /dev/stdin -a #{gff_file.inspect} -hist -S"
-  cmdr2 = "samtools view -u #{read2_flag} #{bam_file.inspect} |bedtools coverage -b /dev/stdin -a #{gff_file.inspect} -hist -S"
+  bedtools_type_flag = '-hist'
+  if options[:count_type] == COUNT_COUNT_TYPE
+    bedtools_type_flag = '-counts'
+  end
+  cmdf1 = "samtools view -u #{read1_flag} #{bam_file.inspect} |bedtools coverage -sorted -g #{chromosome_file.path.inspect} -b /dev/stdin -a #{sorted_gff_file.inspect} -s #{bedtools_type_flag}"
+  cmdf2 = "samtools view -u #{read2_flag} #{bam_file.inspect} |bedtools coverage -sorted -g #{chromosome_file.path.inspect} -b /dev/stdin -a #{sorted_gff_file.inspect} -s #{bedtools_type_flag}"
+  cmdr1 = "samtools view -u #{read1_flag} #{bam_file.inspect} |bedtools coverage -sorted -g #{chromosome_file.path.inspect} -b /dev/stdin -a #{sorted_gff_file.inspect} -S #{bedtools_type_flag}"
+  cmdr2 = "samtools view -u #{read2_flag} #{bam_file.inspect} |bedtools coverage -sorted -g #{chromosome_file.path.inspect} -b /dev/stdin -a #{sorted_gff_file.inspect} -S #{bedtools_type_flag}"
 
   command_to_parsed = lambda do |cmds, name|
     covs_lines_initial = cmds.collect do |cmd|
@@ -122,15 +208,23 @@ else
       get_covs.call(lines)
     end
     covs = covs_initial[0]
-    covs_initial[1].each do |cov_key, cov|
-      covs[cov_key] += cov
+    if covs_initial.length > 1
+      covs_initial[1].each do |cov_key, cov|
+        covs[cov_key] += cov
+      end
     end
     covs #'return' from lambda
   end
 
   # Agreeing reads (those whose template are fwd along the reference sequence) are either first and fwd, or second and rev
-  covs_fwd = command_to_parsed.call([cmdf1,cmdr2], 'reads with same direction as their reference')
-  covs_rev = command_to_parsed.call([cmdf2,cmdr1], 'reads with opposing direction as their reference')
+  commands_fwd = [cmdf1,cmdr2]
+  commands_rev = [cmdf2,cmdr1]
+  if options[:forward_ready_only]
+    commands_fwd = [cmdf1]
+    commands_rev = [cmdr1]
+  end
+  covs_fwd = command_to_parsed.call(commands_fwd, 'reads with same direction as their reference')
+  covs_rev = command_to_parsed.call(commands_rev, 'reads with opposing direction as their reference')
 end
 
 headers = [
@@ -142,9 +236,14 @@ headers = [
 ]
 if options[:ignore_directions]
   headers.push 'average_coverage'
-else
+elsif options[:count_type] == COVERAGE_COUNT_TYPE
   headers.push 'forward_average_coverage'
   headers.push 'reverse_average_coverage'
+elsif options[:count_type] == COUNT_COUNT_TYPE
+  headers.push 'forward_read_count'
+  headers.push 'reverse_read_count'
+else
+  raise
 end
 headers.push 'annotation'
 puts headers.join("\t")

data/spec/script_spec.rb

@@ -51,4 +51,28 @@ describe 'script' do
 
     found.should == answer
   end
+
+  it 'should print counts correctly' do
+    answer = %w(
+    contig	type	start	end	strand	forward_read_count	reverse_read_count	annotation
+    ).join("\t")+"\n"+%w(
+    contig_100	CDS	2	127	+	0.0	2.0	putative
+    ).join("\t")+" methyltransferase YcgJ\n"
+
+    found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/realer.gff -q --measure-type count"
+
+    found.should == answer
+  end
+
+  it 'should count only the forward read when asked' do
+    answer = %w(
+    contig	type	start	end	strand	forward_read_count	reverse_read_count	annotation
+    ).join("\t")+"\n"+%w(
+    contig_100	CDS	2	127	+	0.0	1.0	putative
+    ).join("\t")+" methyltransferase YcgJ\n"
+
+    found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/realer.gff -q --measure-type count --forward-read-only"
+
+    found.should == answer
+  end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dirseq
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors:
 - Ben J. Woodcroft
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-01-08 00:00:00.000000000 Z
+date: 2018-02-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bio-commandeer
@@ -58,48 +58,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.4'
-- !ruby/object:Gem::Dependency
-  name: shoulda
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '3.5'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '3.5'
-- !ruby/object:Gem::Dependency
-  name: rdoc
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '3.12'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '3.12'
-- !ruby/object:Gem::Dependency
-  name: simplecov
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.8'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.8'
 - !ruby/object:Gem::Dependency
   name: jeweler
   requirement: !ruby/object:Gem::Requirement
@@ -134,14 +92,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.99'
+        version: '3.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.99'
+        version: '3.0'
 - !ruby/object:Gem::Dependency
   name: pry
   requirement: !ruby/object:Gem::Requirement
@@ -176,6 +134,7 @@ files:
 - lib/bio-rnaseq_transcription_directionality.rb
 - lib/bio-rnaseq_transcription_directionality/rnaseq_transcription_directionality.rb
 - spec/data/eg.bam
+- spec/data/eg.bam.bai
 - spec/data/eg.gff
 - spec/data/eg_with_fasta.gff
 - spec/data/realer.gff
@@ -201,7 +160,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.5.1
+rubygems_version: 2.6.13
 signing_key: 
 specification_version: 4
 summary: FPKG calculator for metatranscriptomics