bio-gadget 0.4.8 → 0.5.0

data/lib/bio/gadget/strt.rb

@@ -0,0 +1,605 @@
+require 'csv'
+require 'fileutils'
+require 'open3'
+require 'parallel'
+require 'thread'
+
+require 'bio/gadget/strt/prepare_transcriptome.rb'
+
+module Bio
+  class Gadget
+    class Strt < Bio::Gadget
+
+      OPT_GENOME = [ :genome, { :default => 'hg38',
+                                :desc => 'Genome assembly' } ]
+      
+      OPT_UMI_LENGTH = [ :umi_length, { :banner => 'NT',
+                                        :default => 6,
+                                        :desc => 'Length of UMI',
+                                        :type => :numeric } ]
+
+      # strt:alignment
+
+      desc 'alignment REFDIR SEQDIR MAPDIR', 'Align reads to reference'
+      long_desc <<-DESC
+Align STRT reads (*.fq.gz files at SEQDIR) to a reference (REFDIR/ref.*.ht2). The alignments will be at MAPDIR/*.bam, and per-base 5'-end counts will be at MAPDIR/*.bed.gz.
+DESC
+
+      method_option *OPT_BUFFER_SIZE
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_GREP_PREFIX
+      method_option *OPT_PARALLEL
+
+      def alignment(refdir, seqdir, mapdir)
+        
+        Dir.glob("#{File.expand_path(seqdir)}/*.fq.gz").each do |fqgz|
+          base = File.basename(fqgz, '.fq.gz')
+          STDERR.puts "#{`date`.strip}: Align #{base}..."
+          bam = "#{mapdir}/#{base}.bam"
+          pipeline(
+            "hisat2 --no-unal --rna-strandness F --dta-cufflinks -p #{options.parallel} -x #{refdir}/ref -U #{fqgz}",
+            "#{grep_command} -v -E 'NH:i:([2-9][0-9]*|1[0-9]+)'",
+            "samtools sort -@ #{options.parallel} -o #{bam}")
+          sh "samtools index #{bam}"
+        end
+
+        STDERR.puts "#{`date`.strip}: Count from all alignments."
+        Parallel.map(Dir.glob("#{File.expand_path(mapdir)}/*.bam"),
+                     in_threads: options.parallel) do |bam|
+          pipeline(
+            "strt count_per_base#{buffer_size_option}#{coreutils_prefix_option}#{parallel_option(options)} #{bam}",
+            "pigz -c > #{mapdir}/#{File.basename(bam, '.bam')}.bed.gz")
+        end
+        
+      end
+
+      # strt:build_index
+
+      desc 'build_index DIR', 'Build index for alignment'
+      long_desc <<-DESC
+Build index for alignment of STRT reads, from the speficied GENOME, TRANSCRIPTOME and VARIATION, at DIR.
+DESC
+
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_GENOME
+      method_option *OPT_GREP_PREFIX
+      method_option *OPT_PARALLEL
+      
+      def build_index(dir0)
+
+        dir = File.expand_path(dir0)
+        FileUtils.mkdir_p(dir)
+
+        STDERR.puts "#{`date`.strip}: Preparing data files..."
+        
+        Parallel.map(
+          ["strt prepare_genome#{coreutils_prefix_option}#{genome_option(options)} #{dir}",
+           "strt prepare_variation#{coreutils_prefix_option}#{genome_option(options)} --download=only #{dir}",
+           "strt prepare_transcriptome #{options.genome}#{coreutils_prefix_option}#{grep_prefix_option(options)} --download=only #{dir}",
+           "strt prepare_spikein#{coreutils_prefix_option} #{dir}",
+           "strt prepare_ribosome#{coreutils_prefix_option}#{genome_option(options)} #{dir}"], in_threads: options.parallel) do |cmd|
+          system cmd or exit $?.exitstatus
+        end
+
+        system "unpigz -c #{dir}/genome.fa.gz #{dir}/spikein.fa.gz #{dir}/ribosome.fa.gz > #{dir}/ref.fa"
+        system "samtools faidx #{dir}/ref.fa"
+
+        Parallel.map(["strt prepare_transcriptome #{options.genome}#{coreutils_prefix_option}#{grep_prefix_option(options)} --download=no #{dir}",
+                      "strt prepare_variation#{coreutils_prefix_option}#{genome_option(options)} --download=no #{dir}"], in_threads: options.parallel) do |cmd|
+          STDERR.puts cmd
+          system cmd or exit $?.exitstatus
+        end
+        
+        STDERR.puts "#{`date`.strip}: Building index..."
+        
+        system "hisat2-build -f -p #{options.parallel} --snp #{dir}/variation.snp --haplotype #{dir}/variation.haplotype --ss #{dir}/transcriptome.splice_sites --exon #{dir}/transcriptome.exons #{dir}/ref.fa #{dir}/ref"
+        
+      end
+
+      # strt:call_allele
+
+      desc 'call_allele CSV REFDIR MAPDIR', 'Call allele frequency'
+      long_desc <<-DESC
+Call allele frequencies of multiple samples specified in a design CSV, based on alignment files at BAMDIR, and reference sequence 'ref.fa' and the index at REFDIR.
+DESC
+
+      method_option *OPT_GENOME
+
+      def call_allele(csv, refdir, mapdir)
+
+        design = CSV.table(csv)
+        bams = get_temporary_path('strt.call_allele', 'bams')
+        fp = open(bams, 'w')
+        design[:base].each {|bam| fp.puts "#{mapdir}/#{bam}.bam" }
+        fp.close
+        csvdir = File.dirname(csv)
+        bcf = "#{csvdir}/strt-call_allele.bcf"
+
+        pipeline("samtools mpileup -u -t AD,ADF,ADR,DP -f #{refdir}/ref.fa -b #{bams}",
+                 "bcftools call --multiallelic-caller --variants-only --output-type u",
+                 "bcftools filter -s LowQual -e '%QUAL<20 || MIN(FORMAT/DP)<20' --output-type b > #{bcf}")
+        pipeline("bcftools view #{bcf}",
+                 "table_annovar.pl - #{refdir} --buildver #{options.genome} --outfile #{csvdir}/strt-call_allele --remove --protocol refGene --operation g --nastring . --vcfinput")
+
+      end
+
+      # strt:count_per_base
+
+      desc 'count_per_base BAM',  'Count reads per base'
+      long_desc <<-DESC
+Count reads per base , based on an alignment BAM.
+DESC
+
+      method_option *OPT_BUFFER_SIZE
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_PARALLEL
+
+      def count_per_base(bam)
+
+        pipeline(
+          "bedtools bamtobed -i #{bam}",
+          "ruby -F'\t' -anle 'puts [$F[0], $F[5]==\"+\" ? $F[1] : $F[2].to_i-1, $F[5]==\"+\" ? $F[1].to_i+1 : $F[2], $F[5]].join(\"\t\")'",
+          "#{sort_command} -t '\t' -k 1,1 -k 2,2n",
+          "#{uniq_command(options)} -c",
+          "ruby -anle 'puts ($F[1..3]+[\"#{File.basename(bam, '.bam')}\", $F[0], $F[4]]).join(\"\t\")'")
+        
+      end
+
+      # strt:count_per_region
+
+      desc 'count_per_region COUNT REG [REG ...]',
+           'Count reads per region'
+      long_desc <<-DESC
+Count reads per region for a sample. Read counts (a BED-format COUNT) within regions (BED-format REGions) are summed by the region names.
+DESC
+      
+      method_option *OPT_COREUTILS_PREFIX
+
+      def count_per_region(count, region0, *regions0)
+
+        pipeline(
+          "bedtools intersect -nonamecheck -s -wa -wb -a #{count} -b #{([region0]+regions0).join(' ')}",
+          "#{cut_command} -f 5,11",
+          "ruby -F'\t' -e 'n2c={}; while gets; c,n=$_.strip.split /\\t/; n2c[n]=(n2c.key?(n) ? n2c[n] : 0)+c.to_i; end; puts \"ID,COUNT\"; n2c.each {|n,c| puts \"\#{n},\#{c}\"}'")
+        
+      end
+      
+      # strt:prepare_genome
+
+      desc 'prepare_genome DIR', 'Prepare genome data'
+      long_desc <<-DESC
+Prepare data files of the specified GENOME at DIR.
+DESC
+
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_DOWNLOAD
+      method_option *OPT_GENOME
+
+      def prepare_genome(dir0)
+
+        dir = File.expand_path(dir0)
+        tgz = "#{dir}/#{options.genome}.chromFa.tar.gz"
+        ucsc = "rsync://hgdownload.cse.ucsc.edu/goldenPath/#{options.genome}/bigZips"
+
+        if options.download != 'no'
+          if options.genome == 'hg38'
+            rsync_file("#{ucsc}/#{options.genome}.chromFa.tar.gz", tgz)
+          else
+            rsync_file("#{ucsc}/chromFa.tar.gz", tgz)
+          end
+        end
+        pipeline("unpigz -c #{tgz}",
+                 "#{options.coreutils_prefix}tar -xOf - --exclude \"*_*\"",
+                 "gawk 'BEGIN{p=\"\"} /^>/{ print p $1 } !/^>/{ printf $1; p=\"\\n\" } END{ print }'",
+                 "#{fold_command(options)} -w 50",
+                 "pigz -c > #{dir}/genome.fa.gz"
+                ) if options.download != 'only'
+        
+      end
+      
+      # strt:prepare_reads
+
+      desc 'prepare_reads BASE MAP FQGZ ...', 'Prepare STRT reads'
+      long_desc <<-DESC
+Prepare STRT reads from raw sequence files before alignment. After demultiplexing, it performs
+(i) exclusion of redundant reads,
+(ii) exclusion of noncanonical reads, which does not begin with template switching primer,
+(iii) trimming from low-quality base,
+(iv) trimming from sequence similar to HiSeq universal primer,
+and (v) trimming of the template switching primer.
+
+Mandatory paramters are (1) BASE; basename for demulplexed and gzipped fastq files, (2) MAP; filename of comma-separated table between barcode and well, and (3) FQGZs; comma-separated filenames of raw sequences; each file is gzipped fastq. When MAP contains 'CAAAGT,A2' and BASE is '~/test', reads having CAAAGT-like barcode are in '~/test.A2.fq.gz' file after the preprocesses.
+DESC
+
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_GREP_PREFIX
+      method_option *OPT_PARALLEL
+      method_option *OPT_UMI_LENGTH
+
+      method_option :maximum_memory,
+                    default: 50,
+                    desc: 'Maximum memory usage in percent rate',
+                    type: :numeric
+      
+      method_option :minimum_length,
+                    banner: 'NT',
+                    default: 24,
+                    desc: 'Minimum length after the preprocess',
+                    type: :numeric
+      
+      method_option :reads,
+                    desc: 'Number of raw reads for the preprocess',
+                    type: :numeric
+
+      # method_option :maximum_distance,
+      #               default: 1,
+      #               desc: 'Maximum distance between barcode and sequence',
+      #               type: :numeric
+
+      def prepare_reads(base, map, fqgz0, *fqgzs0)
+
+        fqgzs = [fqgz0] + fqgzs0
+
+        bcs = Hash.new
+        open(map, 'r').each do |line|
+          bc, well = line.rstrip.split(',')
+          bcs[bc] = well
+        end
+        
+        bcl = bcs.keys.map!{|key| key.length}.sort.uniq[0]
+
+        tso_pattern = '.'*options.umi_length + '.'*bcl + 'GG'
+
+        #
+        
+        STDERR.puts "#{`date`.strip}: Demultiplexing each raw sequence files..."
+        
+        fqgz2csv0 = Hash.new
+        fqgz2csv1 = Hash.new
+        fqgz2base = Hash.new
+        fqgzs.each do |fqgz|
+          fqgz2csv0[fqgz] = get_temporary_path('strt.preprocess', 'csv', false)
+          fqgz2csv1[fqgz] = get_temporary_path('strt.preprocess', 'csv', false)
+          fqgz2base[fqgz] = get_temporary_path('strt.preprocess', 'base', false)
+        end
+
+        Parallel.map(fqgz2csv0.keys, in_processes: options.parallel) do |fqgz|
+          cmds = [
+            "unpigz -c #{fqgz}",
+            "#{fq1l_convert_command(options)}",
+            "#{fq1l_count_command(options)} #{fqgz2csv0[fqgz]}",
+            "fq1l match_5end#{grep_prefix_option(options)} #{tso_pattern}",
+            "#{fq1l_count_command(options)} #{fqgz2csv1[fqgz]}",
+            "fq1l annotate_index --first-cycle=#{options.umi_length+1} --last-cycle=#{options.umi_length+bcl}",
+            "fq1l annotate_umi --first-cycle=1 --last-cycle=#{options.umi_length}",
+            "fq1l sort_index#{coreutils_prefix_option}#{parallel_option(options)} --buffer-size=#{(options.maximum_memory/(fqgz2csv0.keys.size+1)).to_i}%",
+            "fq1l demultiplex #{fqgz2base[fqgz]} #{map}"
+          ]
+          cmds.insert(2, "#{head_command(options)} -n #{options.reads}") unless options.reads.nil?
+          stats = Open3.pipeline(*cmds)
+          stats.each_index do |i|
+            raise "Fail at process #{i}; #{stats[i]}; #{cmds[i]}" unless stats[i].success? || (stats[i].signaled? && stats[i].termsig == 13)
+          end
+        end
+
+        system "fq1l sum_counts #{fqgz2csv0.values.join(' ')} > #{base}.count.step1.csv"
+        unlink_files(fqgz2csv0.values)
+        
+        system "fq1l sum_counts #{fqgz2csv1.values.join(' ')} > #{base}.count.step2.csv"
+        unlink_files(fqgz2csv1.values)
+
+        #
+        
+        (bcs.values + ['NA']).each do |well|
+
+          STDERR.puts "#{`date`.strip}: Finishing well #{well}..."
+          
+          tmpfqgzs = fqgz2base.values.map {|base| "#{base}.#{well}.fq.gz"}
+          csvs = Array.new(6) {|i| "#{base}.#{well}.count.step#{i+3}.csv"}
+          
+          pipeline("unpigz -c #{tmpfqgzs.join(' ')}",
+                   "#{fq1l_convert_command(options)}",
+                   "#{fq1l_count_command(options)} #{csvs[0]}",
+                   "#{fq1l_sort_command} --buffer-size=#{(options.maximum_memory/2).to_i}%",
+                   "fq1l exclude_duplicate",
+                   "#{fq1l_count_command(options)} #{csvs[1]}",
+                   "fq1l trim_3end_quality",
+                   "#{fq1l_count_command(options)} #{csvs[2]}",
+                   "fq1l trim_3end_primer#{coreutils_prefix_option}#{grep_prefix_option(options)}#{parallel_option(options)}",
+                   "#{fq1l_count_command(options)} #{csvs[3]}",
+                   "#{fq1l_sort_command} --buffer-size=#{(options.maximum_memory/2).to_i}%",
+                   "fq1l exclude_degenerate",
+                   "#{fq1l_count_command(options)} #{csvs[4]}",
+                   "fq1l trim_5end --minimum-length=#{options.minimum_length} #{tso_pattern}+",
+                   "#{fq1l_count_command(options)} #{csvs[5]}",
+                   "fq1l restore#{coreutils_prefix_option}",
+                   "pigz -c > #{base}.#{well}.fq.gz")
+          
+          unlink_files(tmpfqgzs)
+          
+        end
+                             
+      end
+
+      # strt:prepare_ribosome
+
+      desc 'prepare_ribosome DIR', 'Prepare ribosome data'
+      long_desc <<-DESC
+Prepare ribosome data files for the specified GENOME at DIR.
+DESC
+
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_DOWNLOAD
+      method_option *OPT_GENOME
+
+      def prepare_ribosome(dir0)
+
+        dir = File.expand_path(dir0)
+
+        if options.genome[0..1] == 'hg'
+          if options.download != 'no'
+            download_file("https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=nuccore&id=555853&strand=1&rettype=fasta&retmode=text", "#{dir}/U13369.fa")
+            system "pigz #{dir}/U13369.fa" or exit $?.exitstatus
+          end
+          pipeline("unpigz -c #{dir}/U13369.fa.gz",
+                   "gawk '/^>/{ print \">RIBO_U13369.1\" } !/^>/{ printf $1 } END{ print }'",
+                   "#{fold_command(options)} -w 50",
+                   "pigz -c > #{dir}/ribosome.fa.gz"
+                  ) if options.download != 'only'
+        else
+          pipeline("echo", "pigz -c > #{dir}/ribosome.fa.gz")
+        end
+        
+      end
+
+      # strt:prepare_spikein
+
+      desc 'prepare_spikein DIR', 'Prepare spikein data'
+      long_desc <<-DESC
+Prepare spikein data files at DIR.
+DESC
+
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_DOWNLOAD
+
+      def prepare_spikein(dir0)
+        
+        dir = File.expand_path(dir0)
+        zip = "#{dir}/ERCC92.zip"
+        
+        download_file("https://tools.thermofisher.com/content/sfs/manuals/ERCC92.zip", zip) if options.download != 'no'
+        pipeline("unzip -cq #{zip} ERCC92.fa",
+                 "gawk 'BEGIN{p=\"\"} /^>/{ print p \">RNA_SPIKE_\" substr($1, 2); printf(\"AATTC\" ($1 == \">ERCC-00130\" ? \"GAGCTC\" : \"\") ) } /^[ACGT]/{ printf $1; p=\"\\n\" } END{ print }'",
+                 "#{fold_command(options)} -w 50",
+                 "pigz -c > #{dir}/spikein.fa.gz") if options.download != 'only'
+        
+      end
+      
+      # strt:prepare_transcriptome
+
+      register(Bio::Gadget::StrtPrepareTranscriptome,
+               'prepare_transcriptome',
+               'prepare_transcriptome GENOME',
+               'Prepare transcriptome data')
+      
+      # strt:prepare_variation
+
+      desc 'prepare_variation DIR', 'Prepare variation data'
+      long_desc <<-DESC
+Prepare genome variation data files for the specified GENOME dir based on common variations in dbSNP BUILD, at DIR.
+DESC
+      
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_DOWNLOAD
+      method_option *OPT_GENOME
+
+      method_option :dbsnp,
+                    banner: 'BUILD',
+                    default: 146,
+                    desc: 'Build number of dbSNP',
+                    type: :numeric
+      
+      def prepare_variation(dir0)
+        
+        dir = File.expand_path(dir0)
+        snp = "#{dir}/#{options.genome}.snp#{options.dbsnp}Common.txt.gz"
+
+        rsync_file("rsync://hgdownload.soe.ucsc.edu/goldenPath/#{options.genome}/database/snp#{options.dbsnp}Common.txt.gz", snp) if options.download != 'no'
+        pipeline("unpigz -c #{dir}/genome.fa.gz",
+                 "hisat2_extract_snps_haplotypes_UCSC.py - #{snp} #{dir}/variation"
+                ) if options.download != 'only'
+        
+      end
+
+      # strt:qualify
+
+      desc 'qualify CSV REFDIR SEQDIR MAPDIR', 'Qualify samples'
+long_desc <<-DESC
+Qualify samples in a design CSV.'
+DESC
+
+      method_option *OPT_BUFFER_SIZE
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_PARALLEL
+
+      def qualify(csv, refdir, seqdir, mapdir)
+
+        count_commands = ["#{cut_command} -f 5",
+                          "ruby -e 'n=0; while gets; n+=$_.to_i; end; puts n'"]
+
+        samples = CSV.read(csv, {
+                             headers: true,
+                             converters: :numeric
+                           })
+        bases = samples["BASE"]
+
+        samples["TOTAL_READS"] =
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          stat = CSV.table("#{seqdir}/#{base}.count.step8.csv")
+          n = 0
+          stat[:reads].each {|i| n += i }
+          n
+        end
+
+        samples["MAPPED_READS"] =
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          pipeline_readline("unpigz -c #{mapdir}/#{base}.bed.gz",
+                            *count_commands).to_i
+        end
+
+        tmp = Array.new
+        samples.each do |row|
+          tmp << row["MAPPED_READS"].to_f / row["TOTAL_READS"]
+        end
+        samples["MAPPED_RATE"] = tmp
+
+        samples["RIBOSOME_READS"] = 
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          pipeline_readline("bedtools intersect -nonamecheck -u -s -a #{mapdir}/#{base}.bed.gz -b #{refdir}/ribosome.bed.gz",
+                            *count_commands).to_i
+        end
+
+        samples["SPIKEIN_READS"] =
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          pipeline_readline("bedtools intersect -nonamecheck -u -s -a #{mapdir}/#{base}.bed.gz -b #{refdir}/spikein_whole.bed.gz",
+                            *count_commands).to_i
+        end
+
+        samples["SPIKEIN_5END_READS"] =
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          pipeline_readline("bedtools intersect -nonamecheck -u -s -a #{mapdir}/#{base}.bed.gz -b #{refdir}/spikein_5end.bed.gz",
+                            *count_commands).to_i
+        end
+
+        tmp = Array.new
+        samples.each do |row|
+          tmp << row["SPIKEIN_5END_READS"].to_f / row["SPIKEIN_READS"]
+        end
+        samples["SPIKEIN_5END_RATE"] = tmp
+
+        tmp = Array.new
+        samples.each do |row|
+          tmp << (row["MAPPED_READS"] - row["RIBOSOME_READS"] - row["SPIKEIN_READS"]) / row["SPIKEIN_5END_READS"].to_f
+        end
+        samples["RELATIVE_POLYA_RNAS"] = tmp
+
+        samples["CODING_READS"] =
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          pipeline_readline("bedtools intersect -nonamecheck -u -s -a #{mapdir}/#{base}.bed.gz -b #{refdir}/transcriptome.coding_whole.bed.gz",
+                            *count_commands).to_i
+        end
+
+        samples["CODING_5END_READS"] =
+          Parallel.map(bases, in_threads: options.parallel) do |base|
+          pipeline_readline("bedtools intersect -nonamecheck -u -s -a #{mapdir}/#{base}.bed.gz -b #{refdir}/transcriptome.coding_5end.bed.gz",
+                            *count_commands).to_i
+        end
+
+        tmp = Array.new
+        samples.each do |row|
+          tmp << row["CODING_5END_READS"].to_f / row["CODING_READS"]
+        end
+        samples["CODING_5END_RATE"] = tmp
+
+        tmp = Array.new
+        samples.each do |row|
+          tmp << row["CODING_5END_READS"].to_f / row["SPIKEIN_5END_READS"]
+        end
+        samples["RELATIVE_MRNAS"] = tmp
+
+        puts samples
+
+      end
+      
+      # strt:quantitate
+
+      desc 'quantify CSV REFDIR MAPDIR REGBASE',
+           'Quantify samples'
+      long_desc <<-DESC
+Count reads per region for multiple samples in CSV. Read counts (a BED-format COUNT) within regions (REFDIR/ribosome.bed.gz, REFDIR/spikein_5end.bed.gz & REGBASE.bed.gz) are summed by the region names. Addional columns in REGBASE.csv is attached as annotations.
+DESC
+
+      method_option *OPT_COREUTILS_PREFIX
+      method_option *OPT_PARALLEL
+      
+      def quantify(csv, refdir, mapdir, regbase)
+
+        samples = CSV.read(csv, { headers: true, converters: :numeric })
+
+        tmp = CSV.read("#{regbase}.csv", { headers:true, converters: :numeric })
+        anns = tmp.headers; anns.delete('ID')
+        name2anns = Hash.new
+        tmp.each {|row| name2anns[row['ID']] = row.values_at(*anns)}
+
+        bases = samples["BASE"]
+        puts ( ['ID'] + anns +
+               bases.map {|base| "N|#{base}"} +
+               bases.map {|base| "R|#{base}"} ).join(',')
+
+        name2base2cnt = Hash.new
+        base2spike = Hash.new
+        @locker = Mutex.new
+        Parallel.map(samples["BASE"], in_threads: options.parallel) do |base|
+          base2spike[base] = 0.0
+          fp = open("| strt count_per_region#{coreutils_prefix_option} #{mapdir}/#{base}.bed.gz #{refdir}/ribosome.bed.gz #{refdir}/spikein_5end.bed.gz #{regbase}.bed.gz")
+          fp.gets
+          fp.each do |line|
+            name, cnt = line.strip.split /,/
+            @locker.synchronize do
+              if !name2base2cnt.key?(name)
+                name2base2cnt[name] = Hash.new
+                name2base2cnt[name][base] = Hash.new
+              elsif !name2base2cnt[name].key?(base)
+                name2base2cnt[name][base] = Hash.new
+              end
+              name2base2cnt[name][base] = cnt
+              base2spike[base] += cnt.to_f if name =~ /^RNA_SPIKE_/
+            end
+          end
+          fp.close
+        end
+        
+        name2base2cnt.each do |name, base2cnt|
+          puts ( [name] +
+                 (name2anns.key?(name) ?
+                    name2anns[name] : Array.new(anns.length, 'NA')) +
+                 bases.map {|base| base2cnt.key?(base) ?
+                              base2cnt[base].to_f/base2spike[base]*1000 : 0} +
+                 bases.map {|base| base2cnt.key?(base) ? base2cnt[base] : 0}
+               ).join(',')
+        end
+        
+      end
+      
+      #
+
+      no_commands do
+
+        def download_option(options)
+          " --download=#{options.download}"
+        end
+
+        def genome_option(options)
+          " --genome=#{options.genome}"
+        end
+
+        def pipeline_readline(*cmds)
+          fp, ths = Open3.pipeline_r(*cmds)
+          line = fp.gets.strip
+          fp.close
+          ths[-1].join
+          line
+        end
+
+        def rsync_file(remote, local)
+          system "rsync -a #{remote} #{local}" or exit $?.exitstatus
+        end
+        
+      end
+      
+    end
+  end
+end
+
+require 'bio/gadget/strt/count.rb'
+require 'bio/gadget/strt/depth.rb'