rbbt-sources 3.1.30 → 3.1.32

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7da6b2938f54788ddcd4ee03b78e8e8b237bf478
-  data.tar.gz: fc847e1ddf84cb477b4f07ac8b33a393009e72ac
+  metadata.gz: 0c4f4b556028ce477969e38d0b4a1d8a05b01305
+  data.tar.gz: 208e92cf212d96dc3333000d350e598ec45f6361
 SHA512:
-  metadata.gz: d2a2261e40edd858ca483b269d409a95e48f61acfeb07c9b31e2381e5bcffa619c06c6c5afaa4df05ba91405638bbaa94897722c66ef1610344b542746c2aa61
-  data.tar.gz: 7bd00cb785efd5db78ba14607436e42465055da71119d27741f20b7c7de4b791b90c321aaecc1de1d681c9d5ba75823401d03fd5da247d7a187dec56a27c4fd3
+  metadata.gz: 1f3e8855b0a233a6d97b80193a92fd90c45e1bfefb0f5b074a4aad2ec7151d9de1dd1c1eae842a84a9f15aeb436af4b97330845ec4ee8defe03735e65de1037f
+  data.tar.gz: a3f529915268806ac18920cfea3786ec2ab906b8492f3f8d6fd7786480a07bf1b17d6b4f5963a6175a1ff231a1697e53099fda66628e5d1fea9f7d525d7cf44e

data/etc/xena_hubs

@@ -0,0 +1,2 @@
+UCSC public hub (not in my data hubs): https://ucscpublic.xenahubs.net
+TCGA hub: https://tcga.xenahubs.net

data/lib/rbbt/sources/Xena.rb

@@ -0,0 +1,109 @@
+require 'rbbt-util'
+require 'rbbt/resource'
+require 'rest-client'
+
+module Xena
+  extend Resource
+  self.subdir = 'share/databases/Xena'
+
+  DEFAULT_XENA_HUB = "https://ucscpublic.xenahubs.net"
+  TCGA_HUB = "https://tcga.xenahubs.net/"
+  #def self.organism(org="Hsa")
+  #  Organism.default_code(org)
+  #end
+
+  #self.search_paths = {}
+  #self.search_paths[:default] = :lib
+
+  def self.query(query, hub = DEFAULT_XENA_HUB)
+    url = File.join(hub, 'data/')
+    
+    uri = URI.parse(url)
+    http = Net::HTTP.new(uri.host, uri.port)
+    http.use_ssl = true
+
+    header = {'Content-Type': 'text/plain'}
+    request = Net::HTTP::Post.new(uri.request_uri, header)
+    request.body = query
+    response = http.request(request)
+
+    JSON.parse(response.body)
+  end
+
+  def self.tsv(file, hub = DEFAULT_XENA_HUB)
+    url = "#{hub}/download/#{file}"
+    TSV.open(url, :monitor => true, :header_hash => '')
+  end
+
+  def self.cohorts(hub = DEFAULT_XENA_HUB)
+    query =<<-EOF
+(;allCohorts
+(fn []
+  (map :cohort
+    (query
+      {:select [[#sql/call [:distinct #sql/call [:ifnull :cohort "(unassigned)"]] :cohort]]
+      :from [:dataset]})))
+)
+    EOF
+
+    self.query query, hub
+  end
+
+  def self.cohort_dataset(cohort, hub = DEFAULT_XENA_HUB)
+    query =<<-EOF
+(; datasetList
+(fn [cohorts]
+    (let [count-table {:select [[:dataset.name :dname] [:%count.value :count]]
+                       :from [:dataset]
+                       :join [:field [:= :dataset.id :dataset_id]
+                       :code [:= :field.id :field_id]]
+                       :group-by [:dataset.name]
+                       :where [:= :field.name "sampleID"]}]
+        (query {:select [:d.name :d.longtitle :count :d.type :d.datasubtype :d.probemap :d.text :d.status [:pm-dataset.text :pmtext]]
+                   :from [[:dataset :d]]
+                   :left-join [[:dataset :pm-dataset] [:= :pm-dataset.name :d.probemap]
+                                count-table [:= :dname :d.name]]
+                   :where [:in :d.cohort cohorts]})
+))
+ ["#{ cohort }"])
+    EOF
+
+    self.query(query, hub)
+  end
+
+  
+  Xena.claim Xena.data, :proc do 
+  end
+end
+
+if __FILE__ == $0
+
+  Xena.cohorts(Xena::TCGA_HUB).each do |cohort|
+    puts Log.color :magenta, cohort
+    Xena.cohort_dataset(cohort, Xena::TCGA_HUB).each do |info|
+      puts Log.color(:blue, "Name: ") + Log.color(:yellow, info["name"].to_s)
+      puts Log.color(:blue, "Title: ") + info["longtitle"].to_s
+      puts Log.color(:blue, "Type: ") + info["type"].to_s
+      puts
+    end
+    puts
+  end
+
+  raise "STOP"
+
+  query_str =<<-EOF
+  (;allCohorts
+  (fn []
+    (map :cohort
+        (query
+             {:select [[#sql/call [:distinct #sql/call [:ifnull :cohort "(unassigned)"]] :cohort]]
+             :from [:dataset]}))))
+EOF
+
+  Log.severity = 0
+  iii Xena.query(query_str, 'https://ucscpublic.xenahubs.net/')
+  raise "STOP"
+  file = "GTEx_Analysis_v6_RNA-seq_RNA-SeQCv1.1.8_gene_rpkm_log.gz"
+  Log.tsv Xena.tsv(file)
+end
+

data/lib/rbbt/sources/organism.rb

@@ -259,4 +259,20 @@ module Organism
     
     Misc.total_length(exon_ranges)
   end
+
+  def self.chromosome_sizes(organism = Organism.default_code("Hsa"))
+    chromosome_sizes = {}
+
+    Organism[organism].glob_all("chromosome_*").each do |file|
+      chromosome = file.split("_").last.split(".").first
+      size = if Open.gzip?(file) || Open.bgzip?(file)
+               CMD.cmd("zcat '#{ file }' | wc -c ").read
+             else
+               File.size(file)
+             end
+      chromosome_sizes[chromosome] = size.to_i
+    end
+
+    chromosome_sizes
+  end
 end

data/share/install/Organism/organism_helpers.rb

@@ -573,12 +573,13 @@ end
 #{{{ Special files
 require 'bio'
 
-file 'transcript_sequence' => ["exons", "transcript_exons"] do |t|
+file 'transcript_sequence' => ["exons", "transcript_exons", "blacklist_chromosomes"] do |t|
   exon_info = TSV.open('exons', :type => :list, :fields => ["Exon Strand", "Exon Chr Start", "Exon Chr End", "Chromosome Name"], :unnamed => true)
 
   chr_transcript_ranges ||= {}
   transcript_strand = {}
 
+  blacklist_chromosomes = Path.setup(File.expand_path('blacklist_chromosomes')).list
   transcript_exons = Path.setup(File.expand_path('transcript_exons'))
   TSV.traverse transcript_exons do |transcript,values|
     transcript = transcript.first if Array === transcript
@@ -604,7 +605,7 @@ file 'transcript_sequence' => ["exons", "transcript_exons"] do |t|
   transcript_sequence = {}
   chr_transcript_ranges.each do |chr, transcript_ranges|
     begin
-      raise "LRG, GL, HG, NT, KI, and HSCHR chromosomes not supported: #{chr}" if chr =~ /(?:LRG_|GL0|HG|HSCHR|NT|KI)/
+      raise "LRG, GL, HG, NT, KI, and HSCHR chromosomes not supported: #{chr}" if blacklist_chromosomes.include? chr
       p = File.expand_path("./chromosome_#{chr}")
       Organism.root.annotate p
       p.sub!(%r{.*/organisms/},'share/organisms/')

data/test/rbbt/sources/test_organism.rb

@@ -66,6 +66,10 @@ class TestOrganism < Test::Unit::TestCase
     assert_equal ["ENSG00000133703"], Gene.setup("Kras", "Associated Gene Name", "Mmu/jun2011").ensembl.ortholog(Organism.default_code("Hsa"))
   end
 
+  def test_chr_sizes
+    assert Organism.chromosome_sizes["2"].to_i > 10_000_000
+  end
+
   #def test_genes_at_chromosome
   #  pos = [12, 117799500]
   #  assert_equal "ENSG00000089250", Organism::Hsa.genes_at_chromosome_positions(pos.first, pos.last)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-sources
 version: !ruby/object:Gem::Version
-  version: 3.1.30
+  version: 3.1.32
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-05-14 00:00:00.000000000 Z
+date: 2018-08-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util
@@ -89,6 +89,7 @@ files:
 - etc/allowed_biomart_archives
 - etc/biomart/missing_in_archive
 - etc/organisms
+- etc/xena_hubs
 - lib/rbbt/sources/CASCADE.rb
 - lib/rbbt/sources/COREAD_phospho_proteome.rb
 - lib/rbbt/sources/COSTART.rb
@@ -101,6 +102,7 @@ files:
 - lib/rbbt/sources/PRO.rb
 - lib/rbbt/sources/PSI_MI.rb
 - lib/rbbt/sources/STITCH.rb
+- lib/rbbt/sources/Xena.rb
 - lib/rbbt/sources/array_express.rb
 - lib/rbbt/sources/barcode.rb
 - lib/rbbt/sources/bibtex.rb