transfuse 0.4.3 → 0.4.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: ec08ab7f1e09b7ab11ed3aa80f48650c4cebd13c
-  data.tar.gz: f56e079bb4c91c8d1636472409d5f2dccec54a0c
+  metadata.gz: 94b55c79a713ddce1aaba5c302e3ce0d7a3dbd4a
+  data.tar.gz: 4472cc10a3a5a5e29e11bc22ad3d17b07fb7c332
 SHA512:
-  metadata.gz: 0bc631582f019658ece91918c0b2278c24217cc8bdfe0ccd721000c0d75892c59a3d70c78d13d3f32666ea696aa2886ff3a05b8fae4ce158c7debb16daacac06
-  data.tar.gz: dcfd447ef6bedba5d5a0ede16421c01c6fc230ad81c9cd4e1215593ec47f59cad25cd10e364beaeacce0b6d33ce235af22523a3644b9b20f86bca366512c1e82
+  metadata.gz: 8b61add12cd4c12db7ae2a65739c7895ea31c82b188a717f17e860a2009036a28962d53e275e67e244ef3f85e21a291ca15384df14ec5c1c51606a6893c2e25c
+  data.tar.gz: 1443f82858f58c9371f2223fb91fefa802cf809aa6802bbb84c014781641304e8c1a02d9dc6e289291e8b52569bce71438931472b18dfc009d9e9472fcc130c4

data/README.md

@@ -1,40 +1,55 @@
 ## Transfuse
 
-**Transfuse is currently in development and is not yet ready for use**
-
 Transfuse intelligently merges your multiple de novo transcriptome assemblies. Run multiple assemblies with different de novo assemblers, or different settings in the same assembler and have them combined into a single high quality transcriptome.
 
 Transfuse takes in the reads you used to do the assembly and a list of fasta files and produces a single output fasta file.
 
 ### Installation and Running
 
-To install Transfuse, clone this repo:
+To install Transfuse you can get it from rubygems.com
+
+`gem install transfuse`
+
+or you can clone this repo:
 
 `git clone https://github.com/cboursnell/transfuse.git`
 
-Then build and install the ruby gem
+then build and install the ruby gem
 
 `gem build *spec; gem install *gem`
 
+Transfuse also requires `vsearch` to be installed which can be downloaded from:
+
+`https://github.com/torognes/vsearch`
+
 ### Usage
 
 Transfuse is run on the command line. The options are:
 
 ```
-  -a, --assembly=<s>    assembly files in FASTA format, comma-separated
-  -l, --left=<s>        left reads file in FASTQ format
-  -r, --right=<s>       right reads file in FASTQ format
-  -o, --output=<s>      write merged assembly to file
-  -t, --threads=<i>     number of threads (default: 1)
-  -v, --verbose         be verbose
-  -e, --version         Print version and exit
-  -h, --help            Show this message
+  -a, --assemblies=<s>    assembly files in FASTA format, comma-separated
+  -l, --left=<s>          left reads file in FASTQ format
+  -r, --right=<s>         right reads file in FASTQ format
+  -o, --output=<s>        write merged assembly to file
+  -t, --threads=<i>       number of threads (default: 1)
+  -i, --id=<f>            sequence identity to cluster at (default: 1.0)
+  -v, --verbose           be verbose
+  -e, --version           Print version and exit
+  -h, --help              Show this message
 ```
 
 An example command:
 
 `transfuse --assembly soap-k31.fa,soap-k41.fa,soap-k51.fa --left reads_1.fq --right reads_2.fq --output soap-merged.fa --threads 12`
 
+### Contributing
+
+[![Join the chat at https://gitter.im/cboursnell/transfuse](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/cboursnell/transfuse?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
+
+Tranfuse is currently in development.
+
+If you want to suggest, and maybe implement, a new feature, please suggest it on the tracker first.
+
 ### License
 
 This is adademic software - please cite us if you use it in your work.

data/bin/transfuse

@@ -23,26 +23,41 @@ opts = Trollop::options do
 
   EOS
   opt :assemblies, "assembly files in FASTA format, comma-separated",
-      :type => String, :required => true
+      :type => String
   opt :left, "left reads file in FASTQ format",
       :type => String
   opt :right, "right reads file in FASTQ format",
       :type => String
-  opt :scores, "transrate contig score output files, comma-separated. Ignored if reads are provided",
-      :type => String
-  opt :output, "write merged assembly to file",
-      :type => String, :required => :true
+  opt :output, "write merged assembly to file", :type => String
   opt :threads, "number of threads", :type => :int, :default => 1
   opt :id, "sequence identity to cluster at", :type => :float, :default => 1.0
+  # opt :install, "install dependencies"
   opt :verbose, "be verbose"
 end
 
 transfuse = Transfuse::Transfuse.new opts.threads, opts.verbose
 
-assembly_files = transfuse.check_files opts.assemblies
-score_files = transfuse.check_files opts.score if opts.score
-left = transfuse.check_files opts.left if opts.left
-right = transfuse.check_files opts.right if opts.right
+# if opts.install
+#   transfuse.install_dependencies
+# else
+#   missing = transfuse.check_dependencies
+#   unless missing.empty?
+#     list = missing.collect {|i| "#{i.name}:#{i.version}"}.join("\n - ")
+#     msg = "Not installed: \n - #{list}"
+#     abort msg
+#   end
+# end
+
+assembly_files = transfuse.check_files(opts.assemblies, "assemblies")
+left = transfuse.check_files(opts.left, "left")
+right = transfuse.check_files(opts.right, "right")
+if opts.output
+  if File.exist?(opts.output)
+    abort "Output #{opts.output} already exists"
+  end
+else
+  abort "Please specify an output with the --output option"
+end
 
 if opts.scores
 # load the scores from the comma separated list of files

data/lib/transfuse/consensus.rb

@@ -12,87 +12,81 @@ module Transfuse
       @verbose = verbose
     end
 
+    def count_exons list
+      exon_counts = {}
+      list.each_with_index do |hash, index|
+        seq = hash[:seq]
+        exon_count=0
+        gap_count=0
+        prev=""
+        seq.each_char do |c|
+          if c=="-"
+            base = "-"
+          else
+            base = "*"
+          end
+
+          if base!=prev
+            if c=="-"
+              gap_count+=1
+            else
+              exon_count+=1
+            end
+          end
+
+          if c=="-"
+            prev = "-"
+          else
+            prev = "*"
+          end
+        end
+        exon_counts[exon_count] ||= []
+        exon_counts[exon_count] << index
+      end
+      return exon_counts
+    end
+
     def run msa, scores, output
-      return 1 if File.exist?(output)
+      preoutput = "#{File.basename(output, File.extname(output))}_cons.fa"
+      return preoutput if File.exist?(preoutput)
       print "writing consensus " if @verbose
       # msa is a hash
       #   key = cluster id
       #   value = list
       #     list of sequences in cluster aligned with gaps
-      preoutput = "#{File.basename(output, File.extname(output))}_cons.fa"
       count = 0
       File.open("#{output}.data", "w") do |out2|
         File.open(preoutput, "w") do |out|
-          msa.each do |id, list|
+          msa.each do |id, seq_list|
             count+=1
             print "." if count%5_000==0 and @verbose
             exons={}
-            cons = []
-            length = list[0][:seq].length
-              list.each_with_index do |hash, index|
-                seq = hash[:seq]
-                name = hash[:name]
-                out2.write "#{id}\t#{scores[name][:score]}\t#{name}\n"
-                prev = ""
-                gap = 0
-                exon = 0
-                seq.each_char do |c|
-                  if c=="-"
-                    base="-"
-                  else
-                    base="*"
-                  end
-                  if base!=prev
-                    if c=="-"
-                      gap+=1
-                    else
-                      exon+=1
-                    end
-                  end
-                  if c=="-"
-                    prev = "-"
-                  else
-                    prev = "*"
-                  end
-                end
-                exons[index] = exon
-              end
+            cons = {}
+            length = seq_list[0][:seq].length
+            exons = count_exons(seq_list)
 
-            consensus = ""
-            0.upto(length-1) do |i|
-              base="N"
-              counts = {}
-              list.each_with_index do |hash, index|
-                seq = hash[:seq]
-                if seq[i] != "-" and seq[i] != "N"
-                  counts[seq[i]]||=0
-                  counts[seq[i]] += 1
-                  if exons[index]==1
-                    base = seq[i]
+            exons.each do |count, list|
+              0.upto(length-1) do |pos|
+                base = "N"
+                list.each do |index|
+                  b = seq_list[index][:seq][pos]
+                  if b != "-" and b != "N"
+                    base = b
                   end
                 end
-              end
-              if counts.size>0
-                base = counts.sort.last.first
-              end
-              consensus << base
-            end
-
-            if consensus.count("N") < consensus.length.to_f*0.5
-              cons << consensus
-            end
-
-            list.each_with_index do |hash, index|
-              if exons[index] > 1
-                cons << hash[:seq].delete("-")
+                if base != "N"
+                  cons[count]||=""
+                  cons[count]<<base
+                end
               end
             end
 
-            cons.each_with_index do |s,index|
+            cons.each_with_index do |s, index|
               out.write ">contig#{id}.#{index+1}\n"
-              out.write "#{s}\n"
+              out.write "#{s[1]}\n"
             end
 
+
           end # msa.each
         end # file
       end # file open

data/lib/transfuse/transfuse.rb

@@ -18,8 +18,29 @@ module Transfuse
       @verbose = verbose
     end
 
-    def check_files string
+    def check_dependencies
+      # Check dependencies if they are relevant to the command issued,
+      # and handle any commands to install missing ones
+      gem_dir = Gem.loaded_specs['transfuse'].full_gem_path
+      gem_deps = File.join(gem_dir, 'deps', 'deps.yaml')
+
+      return Bindeps.missing gem_deps
+
+    end # check_dependencies
+
+    def install_dependencies
+      # Check dependencies if they are relevant to the command issued,
+      # and handle any commands to install missing ones
+      gem_dir = Gem.loaded_specs['transfuse'].full_gem_path
+      gem_deps = File.join(gem_dir, 'deps', 'deps.yaml')
+
+      Bindeps.require gem_deps
+
+    end # check_dependencies
+
+    def check_files string, option
       # puts "check file string: #{string}" if @verbose
+      abort "Please specify --#{option} option" if string.nil?
       list = []
       string.split(",").each do |file|
         file = File.expand_path(file)
@@ -205,8 +226,8 @@ module Transfuse
 
             end
             File.open("summary.txt","w") do |out|
-              out.write "fasta\tscore\toptimal\n"
-              out.write "#{fasta}\t#{transrater.assembly_score}\t#{transrater.assembly_optimal_score("prefix")}\n"
+              out.write "fasta\tscore\toptimal\tcutoff\n"
+              out.write "#{fasta}\t#{transrater.assembly_score}\t#{transrater.assembly_optimal_score("prefix").join("\t")}\n"
             end
           end
         end

data/lib/transfuse/version.rb

@@ -8,7 +8,7 @@ module Transfuse
   module VERSION
     MAJOR = 0
     MINOR = 4
-    PATCH = 3
+    PATCH = 5
     BUILD = nil
 
     STRING = [MAJOR, MINOR, PATCH, BUILD].compact.join('.')

data/test/test_cluster.rb

@@ -8,7 +8,7 @@ class TestCluster < Test::Unit::TestCase
   context 'cluster' do
 
     setup do
-      @cluster = Transfuse::Cluster.new 4
+      @cluster = Transfuse::Cluster.new 4, true, 1.0
     end
 
     teardown do
@@ -18,14 +18,11 @@ class TestCluster < Test::Unit::TestCase
       assert @cluster
     end
 
-    should 'generate cd-hit command' do
-      cmd = @cluster.generate_cdhit_command "assembly1.fasta", "output.fa"
-    end
-
     should 'generate vsearch command' do
-      output = @cluster.generate_vsearch_command "assembly1.fasta", "output.txt"
-      a = "vsearch --cluster_fast assembly1.fasta --id 1.00 "
-      a << "--strand both --uc output.txt --threads 4"
+      output = @cluster.generate_vsearch_command "assembly1.fasta", "output.txt", "output.msa"
+      a = "vsearch --cluster_fast assembly1.fasta --id 1.0 "
+      a << "--iddef 0 --qmask none --strand both --uc output.txt "
+      a << "--msaout output.msa --threads 4"
       b = output.split(" ")
       b[0] = File.basename(b[0])
       output = b.join(" ")

data/test/test_transfuse.rb

@@ -18,7 +18,7 @@ class TestTransfuse < Test::Unit::TestCase
       list = []
       list << File.join(File.dirname(__FILE__), 'data', 'assembly1.fasta')
       list << File.join(File.dirname(__FILE__), 'data', 'assembly2.fasta')
-      files = @fuser.check_files list.join(",")
+      files = @fuser.check_files(list.join(","), "option")
       assert_equal 2, files.length, "length"
     end
 
@@ -41,7 +41,7 @@ class TestTransfuse < Test::Unit::TestCase
       tmpdir = Dir.mktmpdir
         Dir.chdir(tmpdir) do
           file = File.join(File.dirname(__FILE__), 'data', 'assembly1.fasta')
-          hash = @fuser.cluster file
+          hash = @fuser.cluster(file, 1.0)
           assert_equal 250, hash.size, "output size"
         end
       # end

data/transfuse.gemspec

@@ -15,15 +15,16 @@ Gem::Specification.new do |gem|
   gem.homepage    = 'https://github.com/cboursnell/transfuse'
   gem.license     = 'MIT'
 
-  gem.add_dependency 'trollop', '~> 2.0'
-  gem.add_dependency 'bio', '~> 1.4', '>= 1.4.3'
+  gem.add_dependency 'trollop', '~> 2.1', '>= 2.1.2'
+  gem.add_dependency 'bio', '~> 1.5', '>= 1.5.0'
   gem.add_dependency 'fixwhich', '~> 1.0', '>= 1.0.2'
-  gem.add_dependency 'bindeps', '~> 1.0', '>= 1.0.1'
+  gem.add_dependency 'bindeps', '~> 1.2', '>= 1.2.0'
   gem.add_dependency 'transrate', '~> 1.0', '>= 1.0.1'
+  gem.add_dependency 'bundler', '~> 1.10', '>= 1.10.6'
 
-  gem.add_development_dependency 'rake', '~> 10.3', '>= 10.3.2'
+  gem.add_development_dependency 'rake', '~> 10.4', '>= 10.4.2'
   gem.add_development_dependency 'turn', '~> 0.9', '>= 0.9.7'
-  gem.add_development_dependency 'simplecov', '~> 0.8', '>= 0.8.2'
+  gem.add_development_dependency 'simplecov', '~> 0.10', '>= 0.10.0'
   gem.add_development_dependency 'shoulda-context', '~> 1.2', '>= 1.2.1'
-  gem.add_development_dependency 'coveralls', '~> 0.7'
+  gem.add_development_dependency 'coveralls', '~> 0.8', '>= 0.8.2'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: transfuse
 version: !ruby/object:Gem::Version
-  version: 0.4.3
+  version: 0.4.5
 platform: ruby
 authors:
 - Richard Smith-Unna
@@ -17,34 +17,40 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.0'
+        version: '2.1'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 2.1.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.0'
+        version: '2.1'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 2.1.2
 - !ruby/object:Gem::Dependency
   name: bio
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.4'
+        version: '1.5'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.4.3
+        version: 1.5.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.4'
+        version: '1.5'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.4.3
+        version: 1.5.0
 - !ruby/object:Gem::Dependency
   name: fixwhich
   requirement: !ruby/object:Gem::Requirement
@@ -71,20 +77,20 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: '1.2'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.0.1
+        version: 1.2.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: '1.2'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.0.1
+        version: 1.2.0
 - !ruby/object:Gem::Dependency
   name: transrate
   requirement: !ruby/object:Gem::Requirement
@@ -105,26 +111,46 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 1.0.1
+- !ruby/object:Gem::Dependency
+  name: bundler
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.10'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 1.10.6
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.10'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 1.10.6
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '10.3'
+        version: '10.4'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 10.3.2
+        version: 10.4.2
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '10.3'
+        version: '10.4'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 10.3.2
+        version: 10.4.2
 - !ruby/object:Gem::Dependency
   name: turn
   requirement: !ruby/object:Gem::Requirement
@@ -151,20 +177,20 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '0.10'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.8.2
+        version: 0.10.0
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '0.10'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.8.2
+        version: 0.10.0
 - !ruby/object:Gem::Dependency
   name: shoulda-context
   requirement: !ruby/object:Gem::Requirement
@@ -191,14 +217,20 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.7'
+        version: '0.8'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.8.2
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.7'
+        version: '0.8'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.8.2
 description: See summary
 email:
 - rds45@cam.ac.uk