bio-rocker 1.1.12 → 1.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 77c08046b7d0b8f4528e38726c4b0681ce55a5c1
-  data.tar.gz: 8335d7484f0227f768a1cebdf0937bd8e7ef0513
+  metadata.gz: d4f32b4736d380d7e8b0ff7b6ef46b221b9a0651
+  data.tar.gz: cfbff18723f9b2d95d45350aba30a9b75cb32d5d
 SHA512:
-  metadata.gz: 1a359c796f73a44fb008032786df546e640079d186751429c100e06b09cbc9141ecbe19d7937dc15ae7a3e1a2a02885fa6f08c42c22c97a131d48c471275f96e
-  data.tar.gz: 38ca17ae7f4cd04f85ea799851e23cb6f095a17a3475dfb298b8863f1be8dcc3701e2d9dac08e2f3f8bc5b5f88726904488c411b99d03d277b4ef11f6b5a3a05
+  metadata.gz: 1a6d9fd0c94338415f334e50c928174ee595690c3c093b1478251d169b11265843eafccdec645d9997acd2ccff0cabcb7db67e77960e43a10d5bbc90298592ff
+  data.tar.gz: bd711c0f3a87f221b60014633130b73a581b5b5f6dd60eee5a948d5740f87bee68714f0b71b9b27412cf36ade84aac314ab20b267ae134e32a5d998e3855c056

data/bin/ROCker

@@ -103,6 +103,9 @@ opts = OptionParser.new do |opt|
       opt.on("--reuse-files",
 	 "Re-use existing result files. By default existing files are ignored."
 	 ){ |v| o[:reuse]=true }
+      opt.on("--keep-unlinked",
+         "Keep genomes with unlinked positive coordinates (missing references)."
+         ){ |v| o[:keep_unlinked] = v }
       
       opt.separator ""
       opt.separator "+ EXTERNAL SOFTWARE OPTIONS"
@@ -344,9 +347,6 @@ opts = OptionParser.new do |opt|
 end
 opts.parse!
 
-p ARGV
-
-
 #================================[ Main ]
 rocker = ROCker.new(o)
 begin

data/lib/rocker.rb

@@ -9,13 +9,13 @@ require "rocker/rocdata"
 
 class ROCker
    #================================[ Class ]
-   @@VERSION = "1.1.12"
+   @@VERSION = "1.2.0"
    @@CITATION = [
-      "Orellana, Rodriguez-R, & Konstantinidis. Under review.",
-      "Detecting and quantifying functional genes in short-read",
-      "metagenomic datasets: method development and application",
-      "to the nitrogen cycle genes."]
-   @@DATE = "2016-05-17"
+      "Orellana, Rodriguez-R & Konstantinidis, 2016. DOI:10.1093/nar/gkw900.",
+      "ROCker: accurate detection and quantification of target genes in",
+      "short-read metagenomic data sets by modeling sliding-window bitscores.",
+      "Nucleic Acids Research 45(3):e14."]
+   @@DATE = "2018-06-28"
    @@DEFAULTS = {
       # General
       q: false, r: "R", nucl: false, debug: false, thr: 2, search: :blast,

data/lib/rocker/genome-set.rb

@@ -18,7 +18,16 @@ class GenomeSet
       tmp_ids = Array.new(self.ids)
       ofh = File.open(file, "w")
       while tmp_ids.size>0
-	 ofh.print rocker.ebiFetch(:embl, tmp_ids.shift(200), :fasta)
+	 ofh.print rocker.
+                  ebiFetch(:embl, tmp_ids.shift(200), :fasta).
+                  each_line.to_a.select { |i|
+            if i =~ /^Entry: (\S+) (.*)/
+              warn "EBI returned an error fetching #{$1}: #{$2}"
+              false
+            else
+              true
+            end
+         }.join
       end
       ofh.close
    end
@@ -44,6 +53,9 @@ class GenomeSet
    end
    def size() self.ids.size end
    def empty?() self.ids.empty? end
+   def delete!(ids)
+      ids.map{ |i| @ids.delete(i) }.flatten.compact
+   end
 
    #================================[ Utilities ]
    def genome2taxon(genome_id, rank="species")

data/lib/rocker/protein-set.rb

@@ -63,6 +63,15 @@ class ProteinSet
       return [] if @genomes.empty?
       @genomes.values.reduce(:+).uniq
    end
+   def genome_by_prot_id(prot_id)
+     @genomes[prot_id]
+   end
+   ##
+   # Removes genomes linked to +prot_ids+ (an Array) and returns an Array
+   # of removed genomes.
+   def remove_genomes_by_prot_id!(prot_ids)
+     prot_ids.map{ |i| @genomes.delete(i) }.flatten.compact
+   end
    def tranids
       return [] if @tranids.empty?
       @tranids.values.reduce(:+).uniq

data/lib/rocker/step/build.rb

@@ -10,9 +10,9 @@ require "rocker/genome-set"
 
 class ROCker
    #================================[ Class ]
-   @@EBIREST = "http://www.ebi.ac.uk/Tools"
+   @@EBIREST = "https://www.ebi.ac.uk/Tools"
    @@DEFAULTS.merge!({positive:[], negative:[], seqdepth:0.03, readlen:100,
-      minovl:50,
+      minovl:50, keep_unlinked:false,
       # Ext. Software
       aligner: :clustalo, simulator: :grinder,
       simulatorbin:{grinder:"grinder"},
@@ -27,12 +27,12 @@ class ROCker
    def self.ebirest() @@EBIREST ; end
    def self.has_build_gems?
       return @@HAS_BUILD_GEMS unless @@HAS_BUILD_GEMS.nil?
-      @@HAS_BUILD_GEMS = TRUE
+      @@HAS_BUILD_GEMS = true
       begin
 	 require "rubygems"
 	 require "restclient"
       rescue LoadError
-	 @@HAS_BUILD_GEMS = FALSE
+	 @@HAS_BUILD_GEMS = false
       end
       @@HAS_BUILD_GEMS
    end
@@ -54,6 +54,7 @@ class ROCker
    def ebiFetch(db, ids, format, outfile=nil)
       url = "#{ROCker.ebirest}/dbfetch/dbfetch/" +
 	 "#{db.to_s}/#{ids.join(",")}/#{format.to_s}"
+      $stderr.puts url
       self.restcall url, outfile
    end
    def get_coords_from_gff3(genome_ids, pset, thread_id, json_file)
@@ -152,7 +153,7 @@ class ROCker
       ref_file = @o[:baseout] + ".ref.fasta"
       if not protein_set[:+].aln.nil?
 	 puts "  * reusing aligned sequences as positive set." unless @o[:q]
-	 protein_set[:+].get_from_aln(ref_file, aln)
+	 protein_set[:+].get_from_aln(ref_file, protein_set[:+].aln)
 	 @o[:noaln] = true
       elsif @o[:reuse] and File.size? ref_file
 	 puts "  * reusing positive set: #{ref_file}." unless @o[:q]
@@ -250,8 +251,16 @@ class ROCker
       raise "Cannot find the genomic location of any provided sequence." if
 	 found.nil?
       missing = protein_set[:+].ids - found
-      warn "\nWARNING: Cannot find genomic location of #{missing.size} " +
-	 "sequence(s) #{missing.join(",")}.\n\n" unless missing.empty?
+      unless missing.empty?
+         warn "\nWARNING: Cannot find genomic location of #{missing.size} " +
+	    "sequence(s) #{missing.join(",")}.\n\n"
+         unless @o[:keep_unlinked]
+            del_genomes = protein_set[:+].remove_genomes_by_prot_id!(missing)
+            warn "WARNING: Ignoring #{del_genomes.size} genomes with missing " +
+               "coordinates #{del_genomes.join(",")}.\n\n"
+            genome_set[ :+ ].delete!(del_genomes)
+         end
+      end
       
       # Download genomes
       genome_set[:all] = GenomeSet.new(self,
@@ -260,7 +269,9 @@ class ROCker
       if @o[:reuse] and File.size? genomes_file
 	 puts "  * reusing existing file: #{genomes_file}." unless @o[:q]
       else
-	 puts "  * downloading " + genome_set[:all].size.to_s +
+	 raise "Something went wrong: Nothing left to download." if
+            genome_set[:all].empty?
+         puts "  * downloading " + genome_set[:all].size.to_s +
 	    " genome(s) in FastA." unless @o[:q]
 	 $stderr.puts "   # #{genome_set[:all].ids}" if @o[:debug]
 	 genome_set[:all].download genomes_file
@@ -321,8 +332,8 @@ class ROCker
 		     if l =~ /^>/
 			rd = %r{
 			   ^>(?<id>\d+)\s
-			   reference=[A-Za-z]+\|
-			   (?<genome_id>[A-Za-z0-9_]+)\|.*\s
+			   reference=[A-Za-z_]+[\|:]
+			   (?<genome_id>[A-Za-z0-9_]+)(?:\|.*)?\s
 			   position=(?<comp>complement\()?(?<from>\d+)\.\.
 			   (?<to>\d+)\)?\s
 			}x.match(l)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bio-rocker
 version: !ruby/object:Gem::Version
-  version: 1.1.12
+  version: 1.2.0
 platform: ruby
 authors:
 - Luis (Coto) Orellana
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-05-17 00:00:00.000000000 Z
+date: 2018-06-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rest-client
@@ -64,7 +64,7 @@ files:
 - lib/rocker/step/search.rb
 homepage: http://enve-omics.ce.gatech.edu/rocker
 licenses:
-- artistic 2.0
+- Artistic-2.0
 metadata: {}
 post_install_message: 
 rdoc_options: []
@@ -82,7 +82,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.5.1
+rubygems_version: 2.6.13
 signing_key: 
 specification_version: 4
 summary: ROCker