neurohmmer 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d4482fb1f357fa6979328e0962b4d78fc3e976c2
-  data.tar.gz: f8261599fb51877933c4f9aa275dee98409f602b
+  metadata.gz: 64bfcd138dcaa9161c75b0f3432be1fb086f9b0e
+  data.tar.gz: 0c48aaea036d17b839b62d751313590224dce84c
 SHA512:
-  metadata.gz: 67613f69fbbf1118e6ced3ca3d7f08a3ee615926947dd82b98e49a793804900ca5dfca8858fb137793452e628101ab538e6298586de7ab9acfb73096f44aca88
-  data.tar.gz: c95c9fea86dd0ba0ce28d6f69fd74f703db570aec6eab7823c220af0d95bf760ada5dd65aaf0dcfcad1ec7c7ac29b1a1b078d70bd1687ddc3aafd24ecee78f3a
+  metadata.gz: ca2b38036ba66bb3856c24553dfe4785e3427858e8609fc00db26d4491f0fb70486c0a31508a0ef9462d02d5d56e84f31b3312e4911ce693ca71f53003782f6c
+  data.tar.gz: 6620ff16a2c77edad30b5bdd604178a33045ceb6d33ae9c7b4104919e3342d37ea0fcf0249c4e5a777308d828b4f86ce70d6a8d4a44c6cb319a2718e0265b654

data/bin/neurohmmer

@@ -30,6 +30,17 @@ BANNER
     opt[:temp_dir] = p
   end
 
+  opt[:signalp_path] = 'signalp'
+  opts.on('-s', '--signalp_path path',
+          'The full path to signalp executable') do |s|
+    opt[:signalp_path] = s
+  end
+
+  opts.on('-b', '--hmmer_bin_path path',
+          'The full path to hmmer binaries directory') do |h|
+    opt[:hmmer_path] = h
+  end
+
   opt[:num_threads] = 1
   opts.on('-n', '--num_threads num_of_threads', Integer,
           'The number of threads to use when analysing the input file') do |n|

data/lib/neurohmmer/arg_validators.rb

@@ -10,6 +10,7 @@ module Neurohmmer
         assert_input_file_not_empty(opt[:input_file])
         assert_input_file_probably_fasta(opt[:input_file])
         opt[:type] = assert_input_sequence(opt[:input_file])
+        export_bin_dirs(opt[:hmmer_bin]) if opt[:hmmer_bin]
         # TODO: Assert hmm & mafft binaries
         opt
       end
@@ -66,6 +67,23 @@ module Neurohmmer
         type = Bio::Sequence.new(cleaned_sequence).guess(0.9)
         (type == Bio::Sequence::NA) ? :genetic : :protein
       end
+
+      def export_bin_dirs(hmmer_bin)
+        bin = File.expand_path(bin)
+        if File.exist?(bin) && File.directory?(bin)
+          add_to_path(bin) 
+        else
+          $stderr.puts '*** The following bin directory does not exist:'
+          $stderr.puts "    #{bin}"
+        end
+      end
+
+      ## Checks if dir is in $PATH and if not, it adds the dir to the $PATH.
+      def add_to_path(bin_dir)
+        return unless bin_dir
+        return if ENV['PATH'].split(':').include?(bin_dir)
+        ENV['PATH'] = "#{bin_dir}:#{ENV['PATH']}"
+      end
     end
   end
 end

data/lib/neurohmmer/output.rb

@@ -1,5 +1,7 @@
-require 'slim'
 require 'forwardable'
+require 'slim'
+
+require 'neurohmmer/signalp'
 
 # Top level module / namespace.
 module Neurohmmer
@@ -7,12 +9,12 @@ module Neurohmmer
   class Output
     class <<self
       extend Forwardable
-      def_delegators Neurohmmer, :conf
+      def_delegators Neurohmmer, :opt, :conf
 
       def to_html(hmm_results)
         @html_results = format_seqs_for_html(hmm_results)
-        template_path = File.expand_path(File.join(__FILE__, '../../../',
-                                                   'template/contents.slim'))
+        template_path = File.expand_path('../../../template/contents.slim',
+                                         __FILE__)
         contents_temp = File.read(template_path)
         html_content = Slim::Template.new { contents_temp }.render(self)
         File.open(conf[:html_output], 'w') { |f| f.puts html_content }
@@ -35,9 +37,11 @@ module Neurohmmer
 
       def format_html_seq(seq, flatseq)
         seq.gsub!("\n", '')
+        signalp_output = Signalp.analyse_sequence(seq) if opt[:type] == :protein
         flatseq.each do |hsp|
           seq.gsub!(/#{hsp.gsub('-', '')}/i, '<span class=hsp>\0</span>')
         end
+        seq = format_signal_peptide(seq, signalp_output)
         seq.gsub(/KR|KK|RR/i, '<span class=clv>\0</span>')
           .gsub(/(K|R)<span class=hsp>(K|R)/i, '<span class=clv>\1</span>' \
                 '<span class=clv_i>\2</span><span class=hsp>')
@@ -45,6 +49,32 @@ module Neurohmmer
                 ' class=hsp>', 'R<span class=hsp>K')
           .gsub(/G<span class=clv>/, '<span class=gly>G</span><span class=clv>')
       end
+
+      def format_signal_peptide(seq, sp)
+        return seq if opt[:type] == :genetic || sp[:sp] == 'N'
+        add_signalp_formatting(seq, sp[:ymax_pos].to_i - 1)
+      end
+
+      def add_signalp_formatting(seq, sp_cut_off)
+        s1 = seq[0, sp_cut_off]
+        if s1 =~ /^</
+          s = seq.gsub(/^<span class=hsp>/, '<span class=sp_hsp>')
+          return s.insert(sp_cut_off + 19, '</span><span class=hsp>')
+        elsif s1 =~ /<span class=hsp>/
+          s = s1.gsub(/<span class=hsp>/, '</span><span class=sp_hsp>') +
+              seq[sp_cut_off..-1]
+          return s.insert(0, '<span class=sp>')
+                  .insert(sp_cut_off + 41, '</span><span class=hsp>')
+        elsif s1 =~ /</
+          s = s1.gsub(/<.+$/, '</span><span class=sp_hsp>') +
+              seq[sp_cut_off..-1].gsub(/^\w+?>/,'')
+          return s.insert(0, '<span class=sp>')
+                  .insert(sp_cut_off + 41, '</span><span class=hsp>')
+        else
+          return seq.insert(0, '<span class=sp>').insert(sp_cut_off + 15,
+                                                         '</span>')
+        end
+      end
     end
   end
 end

data/lib/neurohmmer/signalp.rb

@@ -0,0 +1,20 @@
+require 'forwardable'
+
+# Top level module / namespace.
+module Neurohmmer
+  # A class to hold sequence data
+  class Signalp
+    class << self
+      extend Forwardable
+      def_delegators Neurohmmer, :opt
+
+      def analyse_sequence(seq)
+        sp_headers = %w(name cmax cmax_pos ymax ymax_pos smax smax_pos smean d
+                        sp dmaxcut networks)
+        s = `echo ">seq\n#{seq}\n" | #{opt[:signalp_path]} -t euk -f short \
+             -U 0.3 -u 0.3`
+        Hash[sp_headers.map(&:to_sym).zip(s.gsub(/^#.*\n/, '').split)]
+      end
+    end
+  end
+end

data/lib/neurohmmer/version.rb

@@ -1,4 +1,4 @@
 # Top level module / namespace.
 module Neurohmmer
-  VERSION = '0.1.0'
+  VERSION = '0.1.1'
 end

data/template/contents.slim

@@ -15,6 +15,8 @@ html lang="en"
       a{color:#18bc9c; text-decoration:none}a:hover,a:focus{color:#18bc9c; text-decoration:underline}
       p {margin:0 0 10.5px}
       .id {font-weight:bold;}
+      .sp {color:#007AC0; font-weight:bold;}
+      .sp_hsp {color:#007AC0; font-weight:bold; background-color:#FFE4B5;}
       .sequence {word-break:break-all; font-family:Courier New, Courier, Mono;}
       .clv {color:#00B050; font-weight:bold;}
       .clv_i {color:#00B050; font-weight: bold; background-color:#FFE4B5;}
@@ -30,7 +32,7 @@ html lang="en"
           - hits.each do |hit|
             p.sequence
               span.id
-                = hit[:id]
+                = '>' + hit[:id]
               br
               span.seq
                 == hit[:seq]

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: neurohmmer
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Ismail Moghul
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-11-17 00:00:00.000000000 Z
+date: 2015-11-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -233,6 +233,7 @@ files:
 - lib/neurohmmer/arg_validators.rb
 - lib/neurohmmer/hmmer.rb
 - lib/neurohmmer/output.rb
+- lib/neurohmmer/signalp.rb
 - lib/neurohmmer/version.rb
 - neurohmmer.gemspec
 - spec/neurohmmer_spec.rb