rbbt-sources 3.1.51 → 3.1.52

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4cb92f1a2c300c6787870f06f5b4ea45967242a33e7c51cf94f002d1901000af
-  data.tar.gz: 793c56312f02532861451142988b24463de6a30d460f6f7a7238889b91c9c336
+  metadata.gz: 19722279e8b56d1b0cb9e72001e11333e8e436d3a59bf746be3ed744afd047e1
+  data.tar.gz: 7eca716d5a2ee5ab6ba35547230d01466417f373df775fda110cae79366b56d5
 SHA512:
-  metadata.gz: 95ab052b23bc28f919e9cf242e0a74e1e433335879d77fc38752925f74e5ec92b8b42b1c27f1fee170bb3125be1a04b83b24c7da524c46efce07241ff169bdff
-  data.tar.gz: bbe1970028942398b1b84e314d704acb0fac405045785e45516e8dbd65927852874bcceb0393748fd68a507c6d4292fda579d1df16d7831d47b5b9ea0de6bf7d
+  metadata.gz: 9acfd6f4718444fed3891d431e49396d42293105767915375bf86d485e84ed58c5dbb81eff6509ced77e6ae15c38bbbcf91d52590aca6d53e0b3feb4db8c90b6
+  data.tar.gz: 572c173002d2bad704df755542448fc294071c0849bfbac997ce254001591fb463ea8244c167066953224b1e758108ef91417a10b22c02bd0991d819efc68453

data/etc/allowed_biomart_archives

@@ -1,9 +1,4 @@
 may2009
-may2012
-dec2013
 feb2014
-dec2014
-dec2015
-oct2016
 may2017
 apr2019

data/lib/rbbt/sources/clinvar.rb

@@ -1,54 +1,95 @@
 require 'rbbt-util'
 require 'rbbt/resource'
+require 'rbbt/sources/organism'
 
 module ClinVar
   extend Resource
   self.subdir = 'share/databases/ClinVar'
 
-  def self.organism(org="Hsa")
-    Organism.default_code(org)
+  def self.organism_hg19(org="Hsa")
+    Organism.organism_for_build("hg19")
   end
 
+  def self.organism_hg38(org="Hsa")
+    Organism.organism_for_build("hg38")
+  end
+
+
   ClinVar.claim ClinVar.variant_summary, :url, "ftp://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variant_summary.txt.gz"
 
-  ClinVar.claim ClinVar.snv_summary, :proc do
-    url = "ftp://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variant_summary.txt.gz"
-    io = TSV.traverse ClinVar.variant_summary, :type => :array, :into => :stream do |line|
-      line = Misc.fixutf8 line
+  ClinVar.claim ClinVar.hg19.snv_summary, :proc do
+    parser = TSV::Parser.new ClinVar.variant_summary, :type => :list
+    dumper = TSV::Dumper.new :fields => parser.fields, :key_field => "Genomic Mutation", :organism => ClinVar.organism_hg19
+    dumper.init
+    chr_pos, start_pos, ref_pos, alt_pos, assembly_pos = %w(Chromosome PositionVCF ReferenceAlleleVCF AlternateAlleleVCF Assembly).collect{|f| parser.fields.index f}
+    TSV.traverse parser, :into => dumper, :bar => true do |allele,values,fields|
+      chr, start, ref, alt, assembly = values.values_at chr_pos, start_pos, ref_pos, alt_pos, assembly_pos
+      next if assembly != "GRCh37"
+      pos, muts = Misc.correct_vcf_mutation(start.to_i, ref, alt)
+      res = muts.collect{|m| [[chr, pos, m] * ":", values] }
+
+      res.extend MultipleResult
+
+      res
+    end
+    dumper.stream
+  end
+
+  ClinVar.claim ClinVar.hg38.snv_summary, :proc do
+    parser = TSV::Parser.new ClinVar.variant_summary, :type => :list
+    dumper = TSV::Dumper.new :fields => parser.fields, :key_field => "Genomic Mutation", :organism => ClinVar.organism_hg38
+    dumper.init
+    chr_pos, start_pos, ref_pos, alt_pos, assembly_pos = %w(Chromosome PositionVCF ReferenceAlleleVCF AlternateAlleleVCF Assembly).collect{|f| parser.fields.index f}
+    TSV.traverse parser, :into => dumper, :bar => true do |allele,values,fields|
+      chr, start, ref, alt, assembly = values.values_at chr_pos, start_pos, ref_pos, alt_pos, assembly_pos
+      next if assembly != "GRCh38"
+      pos, muts = Misc.correct_vcf_mutation(start.to_i, ref, alt)
+      res = muts.collect{|m| [[chr, pos, m] * ":", values] }
+
+      res.extend MultipleResult
+
+      res
+    end
+    dumper.stream
+  end
+
+
+  ClinVar.claim ClinVar.hg19.mi_summary, :proc do
+    require 'rbbt/workflow'
+    Workflow.require_workflow "Sequence"
+    variants = ClinVar.hg19.snv_summary.produce
+    muts = CMD.cmd('cut -f 1', :in => variants.open, :pipe => true)
+    consequence = Sequence.job(:mutated_isoforms_fast, "Clinvar", :mutations => muts, :non_synonymous => true, :organism => ClinVar.organism_hg19).clean.run(true)
+
+    options = TSV.parse_header(variants).options.merge({:key_field => "Mutated Isoform"})
+    fields = options[:fields].length
+    dumper = TSV::Dumper.new options
+    dumper.init
+    pasted = TSV.paste_streams([variants, TSV.get_stream(consequence)])
+    TSV.traverse pasted, :into => dumper, :bar => true do |mutation,values|
       begin
+        mis = values[fields..-1].flatten
+        next if mis.empty?
         res = []
-        if line =~ /^#/
-          parts = line.split("\t")
-          res << (["#Genomic Mutation"] + parts[1..12] + parts[15..23]) * "\t"
-        else
-          next unless line =~ /GRCh37/ 
-          next if line =~ /(copy number|NT expansion|duplication|indel)/
-          parts = line.split("\t")
-          chr,pos,ref,mut = parts.values_at 13, 14, 25, 26
-          next if ref == 'na' or mut == 'na'
-
-          pos, muts = Misc.correct_mutation(pos.to_i,ref,mut)
-          muts.each do |mut|
-            mutation = [chr,pos,mut] * ":"
-            res << ([mutation] + parts[1..12] + parts[15..23]) * "\t"
-          end
-        end
         res.extend MultipleResult
+        mis.each do |mi|
+          res << [mi, values[0..fields-1]]
+        end
         res
       rescue
         Log.exception $!
         raise $!
       end
     end
-    Misc.sort_stream(io)
+    dumper.stream
   end
 
-  ClinVar.claim ClinVar.mi_summary, :proc do
+  ClinVar.claim ClinVar.hg38.mi_summary, :proc do
     require 'rbbt/workflow'
     Workflow.require_workflow "Sequence"
-    variants = ClinVar.snv_summary.produce
+    variants = ClinVar.hg38.snv_summary.produce
     muts = CMD.cmd('cut -f 1', :in => variants.open, :pipe => true)
-    consequence = Sequence.job(:mutated_isoforms_fast, "Clinvar", :mutations => muts, :non_synonymous => true).clean.run(true)
+    consequence = Sequence.job(:mutated_isoforms_fast, "Clinvar", :mutations => muts, :non_synonymous => true, :organism => ClinVar.organism_hg38).clean.run(true)
 
     options = TSV.parse_header(variants).options.merge({:key_field => "Mutated Isoform"})
     fields = options[:fields].length
@@ -74,3 +115,10 @@ module ClinVar
   end
 end
 
+if __FILE__ == $0
+  Log.severity = 0
+  ClinVar.hg19.snv_summary.produce
+  ClinVar.hg19.mi_summary.produce(true)
+  ClinVar.hg38.snv_summary.produce
+  ClinVar.hg38.mi_summary.produce(true)
+end

data/lib/rbbt/sources/organism.rb

@@ -46,6 +46,12 @@ module Organism
     Rbbt.share.install.Organism.find.glob('???').collect{|f| File.basename(f)}
   end
 
+  def self.prepared_organisms
+    Rbbt.share.organisms.glob_all("???/???????/{identifiers,chromosome_1,scientific_name}").collect{|f| 
+      [File.basename(File.dirname(File.dirname(f))), File.basename(File.dirname(f))] * "/"
+    }.uniq
+  end
+
   def self.installable_organisms
     self.installed_organisms
   end

data/share/install/Organism/Rno/Rakefile

@@ -21,7 +21,7 @@ $biomart_lexicon = [
 $biomart_identifiers = [ 
   ['Entrez Gene ID', "entrezgene"],
   ['Ensembl Protein ID', "ensembl_peptide_id"  ],
-  ['Associated Gene Name' , "rgd_symbol"], 
+  ['Associated Gene Name' , "external_gene_name"], 
   ['Protein ID' , "protein_id"] , 
   ['UniProt/SwissProt ID' , "uniprot_swissprot"] , 
   ['UniProt/SwissProt Accession' , "uniprot_swissprot_accession"] , 
@@ -42,6 +42,14 @@ $biomart_identifiers = [
   #['Codelink ID ', "codelink"],
 ]
 
+$biomart_protein_identifiers = [
+  [ 'Protein ID', "protein_id"  ],
+  [ 'RefSeq Protein ID', "refseq_peptide"  ],
+  [ 'Unigene ID', "unigene"  ],
+  [ 'UniProt/SwissProt ID', "uniprot_swissprot"],
+  [ 'UniProt/SwissProt Accession', "uniprot_swissprot_accession"],
+]
+
 $namespace = File.basename(File.dirname(File.expand_path(__FILE__)))
 Thread.current["namespace"] = File.basename(File.dirname(File.expand_path(__FILE__)))
 load File.join(File.dirname(__FILE__), '../organism_helpers.rb')

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-sources
 version: !ruby/object:Gem::Version
-  version: 3.1.51
+  version: 3.1.52
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-01-28 00:00:00.000000000 Z
+date: 2021-04-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util
@@ -176,7 +176,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.6
+rubygems_version: 3.1.4
 signing_key: 
 specification_version: 4
 summary: Data sources for the Ruby Bioinformatics Toolkit (rbbt)