mapp2g 0.1.4 → 0.1.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 80ed37b4687cdd64f5a292593e2b15addadbabe49223c79835b530e94ac23be6
-  data.tar.gz: 37a038d19321a88e10f44261a3027bda85c40c29d529b710bb93ed760ada585e
+  metadata.gz: ea23c58705cef813135bf96b23383cdeca589643ff08a25ba8b95fd473e26449
+  data.tar.gz: 7be36a76318408344b4402fafda2393ba51d6e2e36fe2c50e6ad393817ff6cbb
 SHA512:
-  metadata.gz: 79f3e7022e8532c1bb9fc0f333927042a4eead1a40d1ecbcb9408d9391a1f21e162830132c395411aac7065076fa14eed694f6ee5b6824a3edc1c193a99ff9ab
-  data.tar.gz: b6c93c34a6576ea1e57fdc2eecdb04e767a595b6b6a037885ab3a89ebf46f36272b6017f55441788ca1cb89dd80cbe273d6a7049f067445b48d8464a486402b2
+  metadata.gz: cfed44714ac742bd9cc8a17169f85c691292e2fe1d7c4d62960d3babe0cd61cf59b01c0b333af063975aef50f17c99aab9f06103b662da03e00813d456f45e02
+  data.tar.gz: c4f3711b154e3ad3d730dde9e5637061514563f49607a814c5bbf2a674de3b06a81f96926aac9d8c32bc5feedb06c7a3dab8a481a8f122b138d60530b923af7c

data/exe/mapp2g

@@ -62,6 +62,9 @@ rescue => e
 end
 
 begin
+  unless File.exist?(genome)
+    raise "genome file (#{genome}) not found"
+  end
   unless File.exist?("#{genome}.nsq") && File.exist?("#{genome}.nos")
     raise "genome is not indexed. Please run 'makeblastdb -in #{genome} -dbtype nucl -parse_seqids'"
   end
@@ -87,4 +90,7 @@ Bio::FlatFile.open(Bio::FastaFormat, query).each_with_index do |fas, i|
   res = mapper.run(query_file_path, genome)
   File.open(out_file_path, "w"){|o| o.puts res}
 
+  gff3 = Mapp2g::ExonerateOutput.new(res).to_gff3()
+  out_file_path = "#{outdir}/#{id}.exonerate.gff3"
+  File.open(out_file_path, "w"){|o| o.puts gff3}
 end

data/lib/mapp2g/report.rb

@@ -0,0 +1,75 @@
+module Mapp2g
+
+  class ExonerateOutput
+
+    def self.load(file)
+      self.new(File.read(file))
+    end
+
+    # @param exonerate_out [String] exonerate output text, not file path
+    def initialize(exonerate_out)
+      @exonerate_out = exonerate_out
+      @query_name = nil
+      @target = nil
+      @cigar = nil
+      @gff2_lines = []
+      #vulgar = nil
+      parse()
+    end
+
+    attr_reader :query_name, :target, :cigar, :gff2_lines
+
+    def parse(opt={})
+      @exonerate_out.each_line do |l|
+        if m = /\s+Query:\s/.match(l)
+          @query_name = m.post_match.chomp.split[0]
+        elsif m = /\s+Target:\s/.match(l)
+          @target = m.post_match.split[0]
+        elsif m = /^cigar:\s/.match(l)
+          @cigar = m.post_match.chomp
+        elsif /^#{@target}/ =~ l &&
+            (/\texonerate:est2genome\t/.match(l) || /\texonerate:protein2genome:local\t/.match(l)) &&
+            (/\texon\t/.match(l) || /\tgene\t/.match(l))
+          @gff2_lines << l.chomp
+        end
+      end
+    end
+
+    def to_gff3(opt={})
+      gff3_lines = []
+      @gff2_lines.each do |l|
+        a = l.chomp.split(/\t/)
+        b = Array.new(9)
+        a.each_with_index{|x, i| b[i] = x}
+        if b[2] == "gene"
+          b[2] = "match"
+          orig_attribute = b[8].split(";").map{|x| x.strip.split(/\s+/)}.to_h
+      #    p orig_attribute
+          c = @cigar.split(/\s+/)
+          cigar_pairs = c[9..-1].join.scan(/[MDI]\d+/)
+          attribute = {'ID' => @query_name,
+                       'Target' => [@query_name, c[1].to_i + 1, c[2]].join(" "),
+                       'Gap'=> cigar_pairs.join(" "),
+                       'identity' => orig_attribute['identity'],
+                       'similarity' => orig_attribute['similarity']}
+          b[8] = attribute.map{|k, v| "#{k}=#{v}"}.join(";")
+        elsif b[2] == "exon"
+          b[2] = "match_part"
+          orig_attribute = b[8].split(";").map{|x| x.strip.split(/\s+/)}.to_h
+          attribute = {'Parent' => @query_name,
+          'identity' => orig_attribute['identity'],
+          'similarity' => orig_attribute['similarity']}
+          b[8] = attribute.map{|k, v| "#{k}=#{v}"}.join(";")
+
+        else
+          raise
+        end
+        gff3_lines << b.join("\t")
+      end
+      return gff3_lines.join("\n")
+    end
+
+  end
+
+end
+

data/lib/mapp2g/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Mapp2g
-  VERSION = "0.1.4"
+  VERSION = "0.1.5"
 end

data/lib/mapp2g.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require_relative "mapp2g/mapper"
+require_relative "mapp2g/report"
 require_relative "mapp2g/version"
 
 module Mapp2g

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: mapp2g
 version: !ruby/object:Gem::Version
-  version: 0.1.4
+  version: 0.1.5
 platform: ruby
 authors:
 - Shuji Shigenobu
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2023-07-09 00:00:00.000000000 Z
+date: 2023-07-15 00:00:00.000000000 Z
 dependencies: []
 description: mapp2g is a bioinformatics software, which map and align protein sequences
   (amino acid sequences) to genome references in a splicing-aware way. mapp2g alignment
@@ -28,6 +28,7 @@ files:
 - exe/mapp2g
 - lib/mapp2g.rb
 - lib/mapp2g/mapper.rb
+- lib/mapp2g/report.rb
 - lib/mapp2g/version.rb
 - mapp2g.gemspec
 - scripts/add_annotation_from_uniprot_fasta_to_gff.rb
@@ -57,7 +58,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.4.10
+rubygems_version: 3.4.15
 signing_key: 
 specification_version: 4
 summary: mapp2g is the tool to map protein sequences to genome references in a splicing-aware