rbbt-GE 0.1.0 → 0.2.0

data/lib/rbbt/GE/GEO.rb

@@ -58,20 +58,27 @@ module GEO
     GSE_URL="ftp://ftp.ncbi.nih.gov/pub/geo/DATA/SOFT/by_series/#SERIES#/#SERIES#_family.soft.gz"
 
     GSE_INFO = {
+      :DELIMITER        => "\\^PLATFORM",
       :title         => "!Series_title",
       :channel_count => "!Sample_channel_count",
       :value_type    => "!Series_value_type",
       :platform      => "!Series_platform_id",
       :description   => "!Series_summary*",      # Join with \n 
     }
-    
+ 
+    GSE_SAMPLE_INFO = {
+      :DELIMITER        => "\\^SAMPLE",
+      :title         => "!Sample_title",
+      :accession         => "!Sample_geo_accession",
+      :channel_count => "!Sample_channel_count",
+    }
+   
     GDS_INFO = {
-      :DELIMITER        => "\\^SUBSET",
-      :value_type       => "!dataset_value_type",
-      :channel_count    => "!dataset_channel_count",
-      :platform         => "!dataset_platform",
-      :reference_series => "!dataset_reference_series",
-      :description      => "!dataset_description",
+      :DELIMITER        => "\\^SUBSET|!sample_table_begin",
+      :title         => "!Sample_title",
+      :accession         => "!Sample_geo_accession",
+      :channel_count    => "!Sample_channel_count",
+      :platform         => "!Sample_platform_id",
     }
 
     GDS_SUBSET_INFO = {
@@ -253,6 +260,64 @@ module GEO
 
       info
     end
+
+
+
+    def self.series_samples(stream)
+      text = stream.read
+
+      values = nil
+
+      sample_info = {}
+      
+      samples = []
+      text.split(/\^SAMPLE/).each do |chunk|
+        info = get_info(chunk, GSE_SAMPLE_INFO)
+        sample = info[:accession]
+        next if sample.nil?
+
+        samples << sample
+
+        sample_values = TSV.open(StringIO.new(chunk.match(/!sample_table_begin(.*)!sample_table_end/msi)[0].strip), :type => :list, :header_hash => '')
+        sample_values.fields = [sample]
+
+        if values.nil?
+          values = sample_values
+        else
+          values.attach sample_values
+        end
+        sample_info[sample] = info
+      end
+
+      [values, sample_info]
+    end
+
+    def self.GSE(series, directory)
+      FileUtils.mkdir_p directory unless File.exists? directory
+
+      value_file = File.join(directory, 'values') 
+      info_file = File.join(directory, 'info.yaml') 
+
+      stream = Open.open(GSE_URL.gsub('#SERIES#', series), :nocache => true)
+
+      info = parse_header(stream, GSE_INFO)
+      info[:value_file]      = value_file
+      info[:data_directory] = directory
+
+      Log.medium "Producing values file for #{ series }"
+      values, sample_info = series_samples(stream)
+
+      key_field = TSV.parse_header(GEO[info[:platform]]['codes'].open).key_field
+      values.key_field = key_field
+
+      Open.write(value_file, values.to_s)
+      Open.write(info_file, info.to_yaml)
+
+      info[:channel_count] ||= sample_info.values.first[:channel_count]
+      info[:value_type] ||= sample_info.values.first[:value_type]
+
+      info
+    end
   end
 
   def self.compare(dataset, field, condition, control, path)

data/share/install/GEO/Rakefile

@@ -13,6 +13,14 @@ rule /^(GDS\d+)\/?(values|info\.yaml)?$/ do |t|
   GEO::SOFT.GDS(dataset, file.nil? ? t.name : File.dirname(t.name))
 end
 
+rule /^(GSE\d+)\/?(values|info\.yaml)?$/ do |t|
+  t.name =~ /^(GSE\d+)\/?(values|info\.yaml)?/
+  series = $1
+  file = $2
+  GEO::SOFT.GSE(series, file.nil? ? t.name : File.dirname(t.name))
+end
+
+
 rule /^(GDS\d+)\/comparison\/(.*)$/ do |t|
   t.name =~ /^(GDS\d+)\/comparison\/(.*)/
   dataset = $1

data/test/rbbt/GE/test_GEO.rb

@@ -71,9 +71,9 @@ class TestClass < Test::Unit::TestCase
   end
 
   def test_GSE
-    gse="GSE966"
-    info = GEO.GSE(gse)
-    assert_equal "GPL764", info[:platform]
+    gse = "GSE21029"
+    info = GEO::SOFT.GSE(gse, "/tmp/gse")
+    assert_equal "GPL570", info[:platform]
   end
 
 
@@ -98,7 +98,5 @@ class TestClass < Test::Unit::TestCase
 
     assert GEO[dataset].comparison[GEO.comparison_name field, condition, control].produce.exists?
   end
-
-
 end
 

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: rbbt-GE
 version: !ruby/object:Gem::Version 
-  hash: 27
+  hash: 23
   prerelease: 
   segments: 
   - 0
-  - 1
+  - 2
   - 0
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors: 
 - Miguel Vazquez
@@ -15,8 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2012-01-19 00:00:00 +01:00
-default_executable: 
+date: 2012-01-31 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: rbbt-util
@@ -48,7 +47,6 @@ files:
 - share/lib/R/MA.R
 - test/test_helper.rb
 - test/rbbt/GE/test_GEO.rb
-has_rdoc: true
 homepage: http://github.com/mikisvaz/rbbt-GE
 licenses: []
 
@@ -78,7 +76,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.6.2
+rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
 summary: Gene Expression in RBBT