mutations_caller_pipeline_aws 0.0.12 → 0.0.13

data/bin/gatk_pipe_only_aws

@@ -44,7 +44,7 @@ options = { :bam_file_sorted => nil,
             :vcf => nil,
             :account => "",
             :project => "",
-            :debug => 1,
+            :debug => 1
           }
 
 optparse = OptionParser.new do |opts|
@@ -111,6 +111,7 @@ recal_bam = "#{random}_recal.bam"
 options[:account] = options[:project] if options[:account].empty?
 
 
+
 # Indexing
 SamtoolsIndexing.call(bam_file,
                       job_prefix,

data/bin/mutations_caller_pipeline_aws

@@ -48,6 +48,8 @@ options = { :mutant_r1 => nil,
             :account => "",
             :project => "",
             :debug => 1,
+            :cluster => false,
+            :coverage => false
           }
 
 optparse = OptionParser.new do |opts|
@@ -89,6 +91,14 @@ optparse = OptionParser.new do |opts|
     options[:debug] =  5 if i
   end
 
+  opts.on("-p","--cluster ", "On compute cluster (none AWS)" ) do |i|
+    options[:cluster] =  true
+  end
+
+  opts.on("-o","--coverage ", "GATK Coverage summary" ) do |i|
+    options[:coverage] =  true
+  end
+
   opts.on_tail("-h", "--help", "Show this message") do
     puts opts
     exit
@@ -125,6 +135,7 @@ sai_file_fwd = "#{random}_fwd.sai"
 sai_file_rev = "#{random}_rev.sai"
 
 options[:account] = options[:project] if options[:account].empty?
+options[:gatk] = "java -Xmx4g -jar #{options[:gatk]}" if options[:cluster]
 
 # BWA : ALN
 BwaCaller.call_aln(options[:mutant_r1],
@@ -219,3 +230,17 @@ GatkCaller.call(log_file,
                 job_prefix,
                 options[:account],
                 options[:debug])
+
+
+# GATK : coverage
+if options[:coverage]
+  outfile_prefix = options[:vcf].gsub(/.vcf/,"_coverage")
+  GatkCaller.coverage(
+    log_file,
+    options[:gatk],
+    options[:index_fa],
+    recal_bam,
+    job_prefix,
+    options[:account],
+    options[:debug])
+end

data/lib/mutations_caller_pipeline_aws/gatk_caller.rb

@@ -3,7 +3,7 @@ class GatkCaller
   # Genotyper
   def self.call(log_dir, gatk, index_fa, read_bam, read_vcf, job_prefix, account, debug)
     cmd = "echo 'starting GATK for mutant at ' `date` >> #{log_dir}
-      qsub -o #{log_dir} -V -cwd -b y -N genotyper_#{job_prefix} -l h_vmem=4G -hold_jid recalibration_#{job_prefix} #{account}\
+      qsub -o #{log_dir} -V -cwd -b y -N genotyper_#{job_prefix} -l h_vmem=4.5G -hold_jid recalibration_#{job_prefix} #{account}\
       #{gatk} -l INFO -R #{index_fa} -T UnifiedGenotyper \
       -I #{read_bam} \
       -o #{read_vcf} \
@@ -12,10 +12,22 @@ class GatkCaller
     system(cmd) if debug == 1
   end
 
+  # INDEX is normal genom.fa
+  # Coverage Summary
+  def self.coverage(log_dir, gatk, index_fa, read_bam, outfile_prefix, job_prefix, account, debug)
+    cmd = "echo 'starting coverage GATK for mutant at ' `date` >> #{log_dir}
+      qsub -o #{log_dir} -V -cwd -b y -N genotyper_#{job_prefix} -l h_vmem=4.5G -hold_jid recalibration_#{job_prefix} #{account}\
+      #{gatk} -R #{index_fa} -T DepthOfCoverage \
+      -I #{read_bam} \
+      -o #{outfile_prefix} "
+    puts cmd
+    system(cmd) if debug == 1
+  end
+
   # Making recalibration table
   def self.recalibrate_bam(log_dir ,gatk, index_fa, index_vcf, read_bam, recal_file, job_prefix, account, debug )
     cmd = "echo 'starting recalibration table ' `date` >> #{log_dir}
-      qsub -o #{log_dir} -V -cwd -b y -N recalibration_table_#{job_prefix} -l h_vmem=4G  -hold_jid realignment_#{job_prefix} #{account} \
+      qsub -o #{log_dir} -V -cwd -b y -N recalibration_table_#{job_prefix} -l h_vmem=4.5G  -hold_jid realignment_#{job_prefix} #{account} \
       #{gatk} -knownSites #{index_vcf} -I #{read_bam} \
       -R #{index_fa} -T CountCovariates \
       -cov ReadGroupCovariate -cov QualityScoreCovariate -cov DinucCovariate \
@@ -28,7 +40,7 @@ class GatkCaller
   # Using recalibration table
   def self.table_calibration(log_dir, gatk, index_fa, read_bam, recal_bam, recal_file, job_prefix, account, debug)
     cmd = "echo 'recalibrating bam_file at ' `date` >> #{log_dir}
-      qsub -V -o #{log_dir} -cwd -b y -N recalibration_#{job_prefix} -l h_vmem=4G -hold_jid recalibration_table_#{job_prefix} #{account} \
+      qsub -V -o #{log_dir} -cwd -b y -N recalibration_#{job_prefix} -l h_vmem=4.5G -hold_jid recalibration_table_#{job_prefix} #{account} \
       #{gatk} \
       -R #{index_fa} \
       -I #{read_bam} \
@@ -42,7 +54,7 @@ class GatkCaller
   # Preparation realignement
   def self.prepare_realigne(log_dir, gatk, read_bam, index_fa, target_intervals, job_prefix, account, debug)
     cmd = "echo 'preparing realignement at ' `date` >> #{log_dir}
-      qsub -o #{log_dir} -V -cwd -b y -N prep_realignment_#{job_prefix} -l h_vmem=4G -hold_jid indexing_#{job_prefix} #{account}\
+      qsub -o #{log_dir} -V -cwd -b y -N prep_realignment_#{job_prefix} -l h_vmem=4.5G -hold_jid indexing_#{job_prefix} #{account}\
       #{gatk} \
       -I #{read_bam} \
       -R #{index_fa} \
@@ -55,7 +67,7 @@ class GatkCaller
   # Realignment
   def self.realigne(log_dir, gatk, read_bam, index_fa, target_intervals, realigned_bam, job_prefix, account, debug)
     cmd = "echo 'preparing realignement at ' `date` >> #{log_dir}
-      qsub -o #{log_dir} -V -cwd -b y -N realignment_#{job_prefix} -l h_vmem=4G -hold_jid prep_realignment_#{job_prefix} #{account} \
+      qsub -o #{log_dir} -V -cwd -b y -N realignment_#{job_prefix} -l h_vmem=4.5G -hold_jid prep_realignment_#{job_prefix} #{account} \
       #{gatk} \
       -I #{read_bam} \
       -R #{index_fa} \

metadata

@@ -1,25 +1,35 @@
---- !ruby/object:Gem::Specification
+--- !ruby/object:Gem::Specification 
 name: mutations_caller_pipeline_aws
-version: !ruby/object:Gem::Version
-  version: 0.0.12
+version: !ruby/object:Gem::Version 
+  hash: 5
   prerelease: 
+  segments: 
+  - 0
+  - 0
+  - 13
+  version: 0.0.13
 platform: ruby
-authors:
+authors: 
 - Kaharina Hayer
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-01-20 00:00:00.000000000Z
+
+date: 2012-01-20 00:00:00 -05:00
+default_executable: 
 dependencies: []
+
 description: Using BWA to align and GATK to call the bases
-email:
+email: 
 - katharinaehayer@gmail.com
-executables:
+executables: 
 - mutations_caller_pipeline_aws
 - gatk_pipe_only_aws
 extensions: []
+
 extra_rdoc_files: []
-files:
+
+files: 
 - bin/gatk_pipe_only_aws
 - bin/mutations_caller_pipeline_aws
 - lib/mutations_caller_pipeline_aws.rb
@@ -28,28 +38,39 @@ files:
 - lib/mutations_caller_pipeline_aws/gatk_caller.rb
 - lib/mutations_caller_pipeline_aws/location_file.rb
 - lib/mutations_caller_pipeline_aws/samtools_indexing.rb
+has_rdoc: true
 homepage: https://github.com/khayer/mutations_caller_pipeline_aws
 licenses: []
+
 post_install_message: 
 rdoc_options: []
-require_paths:
+
+require_paths: 
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement
+required_ruby_version: !ruby/object:Gem::Requirement 
   none: false
-  requirements:
-  - - ! '>='
-    - !ruby/object:Gem::Version
-      version: '0'
-required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
-  requirements:
-  - - ! '>='
-    - !ruby/object:Gem::Version
-      version: '0'
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
 requirements: []
+
 rubyforge_project: mutations_caller_pipeline_aws
-rubygems_version: 1.8.10
+rubygems_version: 1.6.2
 signing_key: 
 specification_version: 3
 summary: Call Mutations for files.fq
 test_files: []
+