rbbt-dm 1.1.28 → 1.1.29

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 1d0744000025bfd9e9a1cefbcf61e54c2f81738d
-  data.tar.gz: b6169f95244d559e90a444b08b234a5f277c1ca7
+  metadata.gz: 27792718ef802b917873c9eec162237e4d90da2f
+  data.tar.gz: 8fd679e51737d0b95a72b09c2d3802614b9f856b
 SHA512:
-  metadata.gz: b3c06b87ba8b12e9d0451086fe888eaf5df84b5725916c61dd97ae6986c9c6c4150bc2a46004f5b07913f6c43e6bf6bce7119b093eac1f97731a70af8dd749ea
-  data.tar.gz: f7304dd025fff1f691f4be46bb4e932eb29145f32bae80366813e0d17516de8326cde6c9c24fa3ad58a08c15346dcc62d971df9a5d6ca0d85e0791590883363a
+  metadata.gz: c351b6b4ffdae4cdeeb135ed34690cd4c81eb5de8c4311c0c82e11d5e6100a0354047b9a5b73897c4fd32d77fef93dac31094008e99747a6bf5fb8dd0e0f751b
+  data.tar.gz: 915b6f886c732373161d1b6bc43b51691e692ca50dafa81307890803e435ba4a8ea8de9a6068d13b35028af173da87dbebbdca02aee5a5ce96eede50ebe7dc3f

data/lib/rbbt/matrix.rb

@@ -11,17 +11,17 @@ class Matrix
   end
 
   attr_accessor :data_file, :labels, :value_type, :format, :organism, :identifiers
-  def initialize(data_file, labels, value_type, format, organism=nil, identifiers=nil)
+  def initialize(data_file, labels = nil, value_type = nil, format = nil, organism=nil, identifiers=nil)
     @data_file = data_file
-    @labels = labels
-    @value_type = value_type
+    @labels = labels 
+    @value_type = value_type || 'count'
     @format = format
+    _header = nil
     @format ||=  begin
                    _header ||= TSV.parse_header(@data_file)
                    _header.key_field || "ID"
                  end
     @organism = organism 
-    _header = nil
     @organism ||=  begin
                      _header ||= TSV.parse_header(@data_file)
                      _header.namespace || Organism.default_code("Hsa")
@@ -150,12 +150,50 @@ class Matrix
     matrix
   end
 
+  def to_barcode_ruby(factor = 2)
+    name = data_file =~ /:>/ ? File.basename(data_file) : data_file
+
+    file = Persist.persist(data_file, :tsv, :prefix => "Barcode #{factor}", :dir => Matrix.matrix_dir.barcode, :no_load => true) do |filename|
+      barcode_ruby(filename, factor)
+    end
+    subsets = self.subsets
+    matrix = Matrix.new file, labels, value_type, "Ensembl Gene ID", organism
+    matrix.subsets = subsets
+    matrix
+  end
+
+  def to_barcode(factor = 2)
+    name = data_file =~ /:>/ ? File.basename(data_file) : data_file
+
+    file = Persist.persist(data_file, :tsv, :prefix => "Barcode R #{factor}", :dir => Matrix.matrix_dir.barcode, :no_load => true) do |filename|
+      barcode(filename, factor).to_list
+    end
+    subsets = self.subsets
+    matrix = Matrix.new file, labels, value_type, "Ensembl Gene ID", organism
+    matrix.subsets = subsets
+    matrix
+  end
+
+  def to_activity(factor = 2)
+    require 'rbbt/tsv/change_id'
+
+    name = data_file =~ /:>/ ? File.basename(data_file) : data_file
+
+    file = Persist.persist(data_file, :tsv, :prefix => "Activity #{factor}", :dir => Matrix.matrix_dir.barcode, :no_load => true) do |filename|
+      activity_cluster(filename, factor)
+    end
+    subsets = self.subsets
+    matrix = Matrix.new file, labels, value_type, "Ensembl Gene ID", organism
+    matrix.subsets = subsets
+    matrix
+  end
+
   def tsv(to_gene=true, identifiers = nil)
     if to_gene and TSV.parse_header(self.data_file).key_field != "Ensembl Gene ID"
       file =  self.to_gene(identifiers).data_file
-      file.tsv :persist => true, :persist_dir => Matrix.matrix_dir.persist, :type => :double, :merge => true, :cast => nil
+      file.tsv :persist => true, :persist_dir => Matrix.matrix_dir.persist, :type => :double, :merge => true
     else
-      self.data_file.tsv :persist => true, :persist_dir => Matrix.matrix_dir.persist, :type => :double, :merge => true, :cast => nil
+      self.data_file.tsv :persist => true, :persist_dir => Matrix.matrix_dir.persist, :merge => true
     end
   end
 

data/lib/rbbt/matrix/barcode.rb

@@ -6,12 +6,45 @@ class Matrix
     FileUtils.mkdir_p File.dirname(outfile) unless outfile.nil? or File.exists? File.dirname(outfile)
     cmd =<<-EOF
 source('#{Rbbt.share.R['barcode.R'].find}')
-rbbt.GE.barcode(#{ R.ruby2R self.data_file }, #{ R.ruby2R outfile }, #{ R.ruby2R factor })
+rbbt.GE.barcode.mode(#{ R.ruby2R self.data_file }, #{ R.ruby2R outfile }, #{ R.ruby2R factor })
     EOF
 
     R.run(cmd)
   end
 
+  def barcode_ruby(outfile, factor = 2)
+    parser = TSV::Parser.new self.data_file
+    dumper = TSV::Dumper.new parser.options.merge(:type => :list, :cast => :to_i)
+    dumper.init
+
+    TSV.traverse parser, :into => dumper, :bar => "Barcoding #{self.data_file}" do |key,values|
+      clean_values = values.flatten.compact.collect{|v| v.to_f}
+      modes = R.eval("rbbt.get.modes(#{R.ruby2R clean_values})$modes")
+      mode = Array === modes ? modes.first : modes
+      mode_values = clean_values.select{|v| v.to_f <= mode}
+      mode_values.concat mode_values.collect{|v| v+mode}
+      sd = Misc.sd mode_values 
+      if sd.nil?
+        [key, [nil] * values.length]
+      else
+        threshold = mode + sd
+        bars = if Array === values.compact.first 
+          values.collect do |v|
+            Misc.mean(v.compact.collect{|v| v.to_f}) > threshold ? 1 : 0
+          end
+        else
+          values.collect do |v|
+            v.to_f > threshold ? 1 : 0
+          end
+        end
+        key = key.first if Array === key
+      [key, bars]
+      end
+    end
+
+    Misc.sensiblewrite(outfile, dumper.stream)
+  end
+
   def activity_cluster(outfile, factor = 2)
 
     FileUtils.mkdir_p File.dirname(outfile) unless outfile.nil? or File.exists? File.dirname(outfile)

data/lib/rbbt/statistics/fisher.rb

@@ -0,0 +1,21 @@
+require 'rbbt/util/R'
+require 'rbbt/util/R/eval'
+
+module Fisher
+  def self.test_classification(classes1, classes2, alternative='greater')
+    matrix = [0,0,0,0]
+    classes1.each_with_index do |c1,i|
+      c2 = classes2[i]
+      if c1 == 1 and c2 == 1
+        matrix[0] += 1
+      elsif c1 == 0 and c2 == 1
+        matrix[1] += 1
+      elsif c1 == 1 and c2 == 0
+        matrix[2] += 1
+      else
+        matrix[3] += 1
+      end
+    end
+    R.eval("fisher.test(matrix(#{R.ruby2R matrix}, nrow=2), alternative = #{R.ruby2R alternative})$p.value")
+  end
+end

data/share/R/barcode.R

@@ -26,6 +26,42 @@ rbbt.GE.barcode <- function(matrix_file, output_file, sd.factor = 2, key.field =
   close(file.barcode)
 }
 
+rbbt.GE.barcode.mode <- function(matrix_file, output_file, sd.factor = 2, key.field = "Ensembl Gene ID"){
+  data = rbbt.tsv(matrix_file)
+  data.mean = rowMeans(data, na.rm=T)
+
+  data.mode = apply(data, 1, function(x){ mode = rbbt.get.modes(x)$modes[1]; lower = x[x <= mode]; return(c(lower, mode, lower+mode));})
+  data.empty = sapply(data.mode,function(x){ length(x) < 3})
+
+  data = data[rownames(data)[!data.empty],]
+  data.mode = data.mode[!data.empty]
+
+  data.sd = sapply(data.mode, sd, na.rm=T)
+  data.threshold = as.vector(sapply(data.mode, function(x){return(x[length(x)/2])})) + data.sd
+  names(data.threshold) = rownames(data)
+
+  file.barcode = file(output_file, 'w')
+
+  cat("#: :type=:list#:cast=:to_i\n", file = file.barcode)
+  cat("#", file = file.barcode)
+  cat(key.field, file = file.barcode)
+  cat("\t", file = file.barcode)
+  cat(colnames(data), file = file.barcode, sep="\t")
+  cat("\n", file = file.barcode)
+     
+  for (gene in rownames(data)){
+    barcode = (data[gene,] - data.threshold[gene]) > 0
+
+    barcode_value = rep(0, length(data[gene,]))
+    barcode_value[barcode] = 1
+
+    cat(gene, file = file.barcode)
+    cat("\t", file = file.barcode)
+    cat(barcode_value, file = file.barcode, sep = "\t")
+    cat("\n", file = file.barcode)
+  }
+  close(file.barcode)
+}
 
 rbbt.GE.activity_cluster <- function(matrix_file, output_file, key.field = "ID"){
 

data/test/rbbt/matrix/test_barcode.rb

@@ -0,0 +1,25 @@
+require File.join(File.expand_path(File.dirname(__FILE__)), '../..', 'test_helper.rb')
+require 'rbbt/matrix'
+require 'rbbt/matrix/barcode'
+
+class TestBarcode < Test::Unit::TestCase
+  def test_R_barcode
+    data = TSV.setup({}, :key_field => "Gene", :fields => %w(S1 S2 S3 S4 S5 S6) , :type => :list)
+    data["G1"] = [1,1,1,4,5,6]
+    data["G2"] = [1,6,1,6,1,6]
+    data["G3"] = [1,1,1,1,6,6]
+    data["G4"] = [6,6,1,1,1,1]
+
+    TmpFile.with_file(data.to_s) do |file|
+      m = Matrix.new file
+      m.barcode(file+'.barcode')
+      tsv =  TSV.open(file+'.barcode')
+      assert tsv["G2"] = [0,1,0,1,0,1]
+
+      m.barcode_ruby(file+'.barcode_ruby')
+      tsv =  TSV.open(file+'.barcode_ruby')
+      assert tsv["G2"] = [0,1,0,1,0,1]
+    end
+  end
+end
+

data/test/rbbt/statistics/test_fisher.rb

@@ -0,0 +1,11 @@
+require File.join(File.expand_path(File.dirname(__FILE__)), '../..', 'test_helper.rb')
+require 'rbbt/statistics/fisher'
+
+class TestFisher < Test::Unit::TestCase
+  def test_classification
+    class1 = [0,0,0,0,1,1,1,1,1]
+    class2 = [0,0,0,1,1,1,1,1,0]
+    iii Fisher.test_classification(class1, class2)
+  end
+end
+

data/test/test_helper.rb

@@ -1,9 +1,22 @@
-require 'test/unit'
+gem "test-unit", "~> 3.0"
+gem "minitest", "~> 5.5"
+
 $LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
 $LOAD_PATH.unshift(File.dirname(__FILE__))
 
+require 'test/unit'
+require 'fileutils'
+
+require 'rbbt'
+require 'rbbt/resource/path'
+
+
 class Test::Unit::TestCase
-  def get_test_datafile(file)
-    File.join(File.dirname(__FILE__), 'data', file)
+  def self.datafile_test(file)
+    Path.setup(File.join(File.dirname(__FILE__), 'data', file.to_s))
+  end
+
+  def datafile_test(file)
+    Test::Unit::TestCase.datafile_test(file)
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-dm
 version: !ruby/object:Gem::Version
-  version: 1.1.28
+  version: 1.1.29
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-10-03 00:00:00.000000000 Z
+date: 2016-10-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util
@@ -99,6 +99,7 @@ files:
 - lib/rbbt/plots/bar.rb
 - lib/rbbt/plots/heatmap.rb
 - lib/rbbt/statistics/fdr.rb
+- lib/rbbt/statistics/fisher.rb
 - lib/rbbt/statistics/hypergeometric.rb
 - lib/rbbt/statistics/random_walk.rb
 - lib/rbbt/statistics/rank_product.rb
@@ -107,8 +108,10 @@ files:
 - share/R/MA.R
 - share/R/barcode.R
 - share/R/heatmap.3.R
+- test/rbbt/matrix/test_barcode.rb
 - test/rbbt/network/test_paths.rb
 - test/rbbt/statistics/test_fdr.rb
+- test/rbbt/statistics/test_fisher.rb
 - test/rbbt/statistics/test_hypergeometric.rb
 - test/rbbt/statistics/test_random_walk.rb
 - test/rbbt/vector/model/test_svm.rb
@@ -142,6 +145,8 @@ test_files:
 - test/rbbt/vector/model/test_svm.rb
 - test/rbbt/vector/test_model.rb
 - test/rbbt/network/test_paths.rb
+- test/rbbt/matrix/test_barcode.rb
 - test/rbbt/statistics/test_random_walk.rb
 - test/rbbt/statistics/test_fdr.rb
 - test/rbbt/statistics/test_hypergeometric.rb
+- test/rbbt/statistics/test_fisher.rb