rewriting uclust_seq

[biopieces.git] / bp_bin / uclust_seq

diff --git a/bp_bin/uclust_seq b/bp_bin/uclust_seq
index b20a29c672673ce690870e7f1e31d54df24e30b0..0eaee60bc66ba6fd54b15752d7c394d89fa28483 100755 (executable)
--- a/bp_bin/uclust_seq
+++ b/bp_bin/uclust_seq
@@ -65,23 +65,6 @@ class Uclust
                execute
   end
 
-  # Method to execute database search.
-  def usearch
-    # usearch --query query.fasta --db db.fasta --uc results.uc --id 0.90 [--evalue E]
-    @command << "usearch --query #{@infile} --db #{@options[:database]} --uc #{@outfile} --id #{@options[:identity]}"
-    @command << "--evalue #{@options[:e_val]}" if @options.has_key? :e_val
-
-               execute
-  end
-
-       # Method to execute clustering to database plus de novo if not matched.
-  def usearch_uclust
-    # usearch --cluster seqs_sorted.fasta --db db.fasta --uc results.uc --id 0.90
-    @command << "usearch --cluster #{@infile} --db #{@options[:database]} --uc #{@outfile} --id #{@options[:identity]}"
-
-               execute
-  end
-
        # Method to parse a Uclust .uc file and for each line of data
        # yield a Biopiece record.
   def each
@@ -92,18 +75,12 @@ class Uclust
         if line !~ /^#/
           fields = line.chomp.split("\t")
 
-          record[:REC_TYPE] = "UCLUST"
+          next if fields[0] == 'C'
+
           record[:TYPE]     = fields[0]
           record[:CLUSTER]  = fields[1].to_i
-          record[:SEQ_LEN]  = fields[2].to_i
           record[:IDENT]    = fields[3].to_f
-          record[:STRAND]   = fields[4]
-          record[:Q_BEG]    = fields[5].to_i
-          record[:S_BEG]    = fields[6].to_i
-          record[:S_END]    = fields[6].to_i + fields[2].to_i
-          record[:CIGAR]    = fields[7]
           record[:Q_ID]     = fields[8]
-          record[:S_ID]     = fields[9]
 
           yield record
         end
@@ -129,28 +106,20 @@ class Uclust
        end
 end
 
-ok_methods = "uclust,usearch,usearch_uclust"
-
 casts = []
 casts << {:long=>'no_sort',  :short=>'n', :type=>'flag',   :mandatory=>false, :default=>nil,      :allowed=>nil,        :disallowed=>nil}
-casts << {:long=>'method',   :short=>'m', :type=>'string', :mandatory=>true,  :default=>"uclust", :allowed=>ok_methods, :disallowed=>nil}
-casts << {:long=>'database', :short=>'d', :type=>'file!',  :mandatory=>false, :default=>nil,      :allowed=>nil,        :disallowed=>nil}
 casts << {:long=>'comp',     :short=>'c', :type=>'flag',   :mandatory=>false, :default=>nil,      :allowed=>nil,        :disallowed=>nil}
 casts << {:long=>'identity', :short=>'i', :type=>'float',  :mandatory=>true,  :default=>0.9,      :allowed=>nil,        :disallowed=>nil}
-casts << {:long=>'e_val',    :short=>'e', :type=>'float',  :mandatory=>false, :default=>nil,      :allowed=>nil,        :disallowed=>nil}
 
 options = Biopieces.options_parse(ARGV, casts)
 
-# At high identities, around 96% and above, compressed indexes are often more sensitive, faster 
-# and use less RAM. Compressed indexes are disabled by default, so I generally recommend that 
-# you specify the --slots and --w options when clustering at high identities.
-
-tmpdir  = Biopieces.mktmpdir
-infile  = File.join(tmpdir, "in.fna")
-outfile = File.join(tmpdir, "out.uc")
+tmpdir       = Biopieces.mktmpdir
+file_records = File.join(tmpdir, "data.stream")
+file_fasta   = File.join(tmpdir, "in.fna")
+file_uclust  = File.join(tmpdir, "out.uc")
 
-Biopieces.open(options[:stream_in], options[:stream_out]) do |input, output|
-  Fasta.open(infile, mode="w") do |fasta_io|
+Biopieces.open(options[:stream_in], file_records) do |input, output|
+  Fasta.open(file_fasta, mode="w") do |fasta_io|
     input.each_record do |record|
       output.puts record
 
@@ -159,17 +128,29 @@ Biopieces.open(options[:stream_in], options[:stream_out]) do |input, output|
       end
     end
   end
+end
 
-  uclust = Uclust.new(infile, outfile, options)
-  uclust.sort unless options[:no_sort] or options[:method] == "usearch"
+uc = Uclust.new(file_fasta, file_uclust, options)
+uc.sort unless options[:no_sort]
+uc.cluster
 
-  case options[:method].to_s
-  when "uclust"         then uclust.cluster
-  when "usearch"        then uclust.usearch
-  when "usearch_uclust" then uclust.usearch_uclust
-  end
+hash = {}
+
+uc.each do |record|
+  hash[record[:Q_ID].to_sym] = record.dup
+end
+
+Biopieces.open(file_records, options[:stream_out]) do |input, output|
+  input.each_record do |record|
+    if record.has_key? :SEQ_NAME and record.has_key? :SEQ
+      if hash.has_key? record[:SEQ_NAME].to_sym
+        uc = hash[record[:SEQ_NAME].to_sym]
+        record[:CLUSTER] = uc[:CLUSTER].to_i
+        record[:IDENT]   = uc[:IDENT].to_i
+        record[:IDENT] = '*' if uc[:TYPE] == 'S'
+      end
+    end
 
-  uclust.each do |record|
     output.puts record
   end
 end