germ 0.1

data/lib/sam.rb

@@ -0,0 +1,196 @@
+require 'germ/data_types'
+require 'germ/printer'
+class Sam
+  include Enumerable
+  include Printer
+
+  class Header
+    class Record
+      attr_reader :type
+      attr_reader :tags
+      attr_reader :comment
+      attr_reader :line
+
+      def initialize line
+        @line = line
+        if line =~ /^@CO\t(.*)/
+          @type = :CO
+          @comment = $1
+        end
+        line.match /^@([A-Za-z][A-Za-z])\t((?:[A-Za-z][A-Za-z0-9]:[ -~]+(?:\t|$))+)/ do |m|
+          @type = m[1].to_sym
+          @tags = Hash[m[2].split(/\t/).map{|s| s.split(/:/)}]
+        end
+      end
+
+      def to_s
+        if @type == :CO
+          "@#{@type}\t#{@comment}"
+        else
+          "@#{@type}\t#{tags.map{ |t,v| "#{t}:#{v}" }.join("\t")}"
+        end
+      end
+    end
+
+    attr_reader :records
+    def initialize sam, lines
+      @sam = sam
+      @records = []
+      lines.each do |l|
+        @records.push Sam::Header::Record.new(l)
+      end
+    end
+
+    def output f
+      @records.each do |r|
+        f.puts r
+      end
+    end
+  end
+
+  class Read
+    extend DataTypes
+    attr_sym :qname
+    attr_integer :flag, :pos, :mapq, :pnext, :tlen
+    attr_string :rname, :rnext, :other, :sam, :seq, :qual, :cigar_string
+    alias :chrom :rname
+    alias :start_pos :pos
+    alias :mate_pos :pnext
+
+    def cigar
+      @cigar ||= @cigar_string.scan(/(\d+)(\w)/)
+    end
+
+    def paired?;  flag & 1; end
+    def mapped?;  flag & 2; end
+    def unmapped?;  flag & 4; end
+    def mate_unmapped?; flag & 8; end
+    def reversed?; flag & 16; end
+    def mate_reversed?; flag & 32; end
+    def first?; flag & 64; end
+    def second?; flag & 128; end
+    def secondary?; flag & 258; end
+    def unchaste?; flag & 512; end
+    def supplementary?; flag & 1024; end
+
+    def end_pos
+      epos = pos
+      cigar.each do |w,c|
+        case c
+        when 'M', 'D', 'N', '=', 'X'
+          epos += w.to_i
+        end
+      end
+      epos
+    end
+
+    def seq_at p
+      cpos = pos
+      cind = 0
+      cigar.each do |w,c|
+        case c
+        when 'M', '=', 'X'
+          epos = cpos + w.to_i
+          if (cpos...epos).include? p
+            return seq[cind + p - cpos]
+          end
+          cind = cind + w.to_i
+          cpos = epos
+        when 'D', 'N'
+          epos = cpos + w.to_i
+          if (cpos...epos).include? p
+            return nil
+          end
+          cind = cind + w.to_i
+          cpos = epos
+        when 'I'
+          cind = cind + w.to_i
+        when 'S', 'H', 'P'
+          return nil
+        end
+      end
+      nil
+    end
+
+    def junctions
+      epos = pos
+      cigar.map do |w,c|
+        case c
+        when 'M', 'D', '=', 'X'
+          epos += w.to_i
+          nil
+        when 'N'
+          [ epos-1, epos += w.to_i ]
+        end
+      end.compact
+    end
+
+    def mate_chrom
+      if rnext == "="
+        chrom
+      else
+        rnext
+      end
+    end
+
+    def mate
+      sam.mates[qname].find{|l| l.pos == mate_pos}
+    end
+
+    def initialize s, fields
+      @sam = s
+      @header = [:qname, :flag, :rname, :pos, :mapq, :cigar_string, :rnext, :pnext, :tlen, :seq, :qual, :other]
+      @header.each_with_index do |s,i|
+        send "#{s}=".to_sym, fields[i]
+      end
+      sam.add_mate qname, self
+    end
+
+    def output f
+      f.puts [ @qname, @flag, @rname, @pos, @mapq, @cigar_string, @rnext, @pnext, @tlen, @seq, @qual, @other].join("\t")
+    end
+  end
+
+  attr_reader :reads, :mates
+  attr_accessor :header
+  def initialize
+    @reads = []
+    @mates = {}
+  end
+
+  def add_mate mate, record
+    @mates[mate] ||= []
+    @mates[mate].push record
+  end
+
+  def each
+    @reads.each do |r|
+      yield r
+    end
+  end
+
+  def self.read file
+    sam = self.new
+    header = []
+    File.foreach(file) do |l|
+      if l =~ /^@/
+        header.push l
+        next
+      end
+      sam.reads.push Sam::Read.new(sam, l.chomp.split(/\t/,12))
+    end
+    sam.header = Sam::Header.new(sam, header)
+    sam
+  end
+
+  def output f
+    @header.output f
+    @reads.each do |r|
+      r.output f
+    end
+  end
+
+  def inspect
+    "#<#{self.class.name}:#{object_id} @reads=#{reads.size}>"
+  end
+end

data/lib/vcf.rb

@@ -0,0 +1,162 @@
+require 'mutation_set'
+require 'oncotator'
+require 'yaml'
+
+class VCF < MutationSet::Sample
+  requires "#CHROM", "POS", "ID", "REF", "ALT", "QUAL", "FILTER", "INFO", "FORMAT"
+  comments "##"
+
+  class Preamble
+    def initialize lines
+      @items = {}
+      lines.each do |line|
+        add_key line.chomp
+      end
+    end
+
+    protected
+    def add_key line
+      line.match(/^##(?<key>\w+)=(?<value>.*)$/) do |m|
+        add_key_item m[:key].to_sym, m[:value]
+      end
+    end
+
+    def add_key_item key, value
+      @items[key] ||= []
+      @items[key].push new_item(value)
+    end
+
+    def new_item value
+      case value
+      when /^<(.*)>$/
+        Hash[$1.split(/,/).map{|s| s.split(/=/)}]
+      else
+        value
+      end
+    end
+  end
+
+  def enforce_headers(array)
+    # kludge for empty vcf with no format line
+    missing = required.map(&:downcase) - array.map(&:downcase)
+    raise "VCF lacks required headers" if !missing.empty? && !(missing.first == "format" && missing.size == 1)
+
+    if array.size > required.size
+      @samples = array - required
+    end
+
+    @headers = array.map &:to_sym
+  end
+
+  class Line < MutationSet::Line
+    attr_reader :format, :mutation
+    alias_key :start, :pos
+    alias_key :ref_allele, :ref
+    def alt_allele; pick_alt; end
+    def stop; @stop || end_pos; end
+    def stop= nc; @stop = nc; end
+
+    def required
+      sample.required
+    end
+
+    def initialize(fields, s)
+      @sample = s
+      @mutation = Hash[clean_required.zip(fields[0...required.size])]
+      @mutation[:info] = Hash[@mutation[:info].split(/;/).map do |s| 
+        key, value = s.split(/=/)
+        value ||= true
+        [ key.to_sym, value ]
+      end]
+      @format = @mutation[:format] = @mutation[:format].split(/:/).map(&:to_sym)
+
+      if @sample.samples
+        sample_fields = fields[required.size..-1]
+        @genotypes = {}
+        @sample.samples.each_with_index do |s,i|
+          next if !sample_fields[i]
+          @genotypes[s] = VCF::Genotype.new self, sample_fields[i].split(/:/)
+        end
+      end
+    end
+
+    def skip_genotype? g
+      name, geno = g.first
+      geno = genotype(geno)
+
+      !geno || geno.empty? || criteria_failed?(geno, name)
+    end
+
+    def pick_alt
+      alt.split(/,/).first
+    end
+
+    def end_pos
+      pos.to_i + ref.length-1
+    end
+
+    def to_s
+      clean_required.map{ |h|
+        case h
+        when :info
+        mutation[h].map{|k,v|  "#{k}=#{v}" }.join(";")
+        when :format
+        mutation[h].join(":")
+        else
+        mutation[h]
+        end
+      }.join("\t") + "\t" + sample.samples.map{|s| genotype(s).to_s }.join("\t")
+    end
+
+    def genotype(s)
+      @genotypes[s] if @genotypes
+    end
+
+    def clean_required
+      sample.clean_headers[0...required.size]
+    end
+  end
+
+  class Genotype
+    attr_reader :info
+    def initialize(line,field)
+      @line = line
+      @info = Hash[line.format.zip(field)]
+    end
+
+    def homozygous?
+      @info[:GT] =~ /0.0/ || @info[:GT] =~ /1.1/
+    end
+
+    def heterozygous?
+      @info[:GT] =~ /0.1/ || @info[:GT] =~ /1.0/
+    end
+
+    def empty?
+      @info[:GT] =~ /\..\./
+    end
+
+    def callable?
+      @info[:GT] !~ /\..\./
+    end
+
+    def gt; @info[:GT]; end
+    def approx_depth; @info[:DP].to_i; end
+    def depth; alt_count + ref_count; end
+    def alt_count; @info[:AD] ? @info[:AD].split(/,/)[1].to_i : nil; end
+    def ref_count; @info[:AD] ? @info[:AD].split(/,/)[0].to_i : nil; end
+    def alt_freq; alt_count / depth.to_f; end
+    def ref_freq; ref_count / depth.to_f; end
+    def ref_length; @line.ref.length; end
+    def alt_length; @line.alt.length; end
+    def alt_base_quality; @info[:NQSBQ] ? @info[:NQSBQ].split(/,/)[0].to_f : nil; end
+    def alt_map_quality; @info[:MQS] ? @info[:MQS].split(/,/)[0].to_f : nil; end
+    def alt_mismatch_rate; @info[:NQSMM] ? @info[:NQSMM].split(/,/)[0].to_f : nil; end
+    def alt_mismatch_count; @info[:MM] ? @info[:MM].split(/,/)[0].to_f : nil; end
+    def quality; @info[:GQ].to_i; end
+
+    def to_s
+      @line.format.map{|f| @info[f]}.join(":")
+    end
+  end
+end

metadata

@@ -0,0 +1,115 @@
+--- !ruby/object:Gem::Specification
+name: germ
+version: !ruby/object:Gem::Version
+  version: '0.1'
+  prerelease: 
+platform: ruby
+authors:
+- Saurabh Asthana
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2014-07-16 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: extlib
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: net-http-persistent
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: sequel
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+description: See summary
+email: Saurabh.Asthana@ucsf.edu
+executables: []
+extensions:
+- ext/fasta_aux/extconf.rb
+- ext/hash_table_aux/extconf.rb
+extra_rdoc_files: []
+files:
+- lib/gtf.rb
+- lib/mutation_set.rb
+- lib/indelocator.rb
+- lib/maf.rb
+- lib/hash_table.rb
+- lib/sam.rb
+- lib/germ/config.rb
+- lib/germ/printer.rb
+- lib/germ/data_types.rb
+- lib/germ/flagstat.rb
+- lib/oncotator.rb
+- lib/mutect.rb
+- lib/vcf.rb
+- lib/germ.rb
+- lib/intervals.rb
+- lib/fasta.rb
+- ext/fasta_aux/FastaAux.c
+- ext/hash_table_aux/HashTableAux.c
+- ext/fasta_aux/extconf.rb
+- ext/hash_table_aux/extconf.rb
+homepage: http://github.com/mountetna/germ
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+- ext
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.23
+signing_key: 
+specification_version: 3
+summary: Collection of utilities for use in computational genomics
+test_files: []