Class: Exodb::Generef

Inherits:
Reference show all
Includes:
GenomeLocationField
Defined in:
lib/exodb/datamodel/reference.rb

Instance Method Summary collapse

Methods included from GenomeLocationField

#begin, #chromosome, #end, #location_str, #parse_location, #start, #stop

Methods included from XrefsField

#get_xref

Instance Method Details

#can_translated? ⇒ Boolean

Check if Generef can translate

Returns:

  • (Boolean) —

    Return true if this can be translate



212
213
214
 
# File 'lib/exodb/datamodel/reference.rb', line 212

def can_translated?()
  return self.has_sequence? && self.has_splices? && self.longest_splice != nil ? true : false
end

#dl_occurrent! ⇒ Object

Download incident data from TCGA



116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
 
# File 'lib/exodb/datamodel/reference.rb', line 116

def dl_occurrent!
  
  if self.get_xref('urn:miriam:hgnc.symbol')
    
    cancerstudies = []
    
    open("http://www.cbioportal.org/public-portal/webservice.do?cmd=getCancerStudies") {|f|
      f.each_line {|line| cancerstudies.push(line.chomp.split("\t")[0])}
    }
    
    occurrents = {}
    totalcase = {}
    
    cancerstudies.each do |study|
      
      totalcase[study] = 0 if !totalcase.has_key?(study)
      
      open("http://www.cbioportal.org/public-portal/webservice.do?cmd=getCaseLists&cancer_study_id=#{study}") do |f|
        f.each_line do |line|
          totalcase[study] += line.chomp.split(/\t/)[4].split(' ').length if line =~ /\tSequenced Tumors\t/
        end
      end
      
      occurrents[study] = {} if !occurrents.has_key?(study)
      
      open("http://www.cbioportal.org/public-portal/webservice.do?cmd=getMutationData&genetic_profile_id=#{study}_mutations&gene_list=#{self.get_xref('urn:miriam:hgnc.symbol').id}") do |f|
        f.each_line do |line|
          dat = line.chomp.split(/\t/)
          
          if dat[5] == 'Missense_Mutation'
            
            occurrents[study][dat[7].split(/(\d+)/)[1]] = [] if !occurrents[study].has_key?(dat[7].split(/(\d+)/)[1])
            occurrents[study][dat[7].split(/(\d+)/)[1]].push(dat[2])
            
          end
          
        end
        
        
      end
    end
    
    self.occurrents.clear if self.occurrents
    
    occurrents.each_pair do |cancertype, v|
      v.each_pair do |position, occur|
        self.occurrents << Occurrent.new({cancertype: cancertype, position: position, occur: occur.uniq.sort, casenumber: totalcase[cancertype]})
      end
    end
    
    self.save!
    
  end
  
end

#dl_seq! ⇒ Object

Download sequence from web service please use by caution. NCBI will block scamming sequest



71
72
73
74
75
76
77
78
79
80
 
# File 'lib/exodb/datamodel/reference.rb', line 71

def dl_seq!
  
  case self.chrrefseq
  when /\Aurn:miriam:refseq:/
    self.sequence = Bio::FastaFormat.new(Bio::NCBI::REST.efetch(self.chrrefseq.split(':', 4), {"db"=>"nucleotide", "rettype"=>"fasta", "retmode"=>"text", "seq_start"=>self.start, "seq_stop"=>self.end})).seq
  end
  
  self.save!
  
end

#dl_symbol! ⇒ Object

Download gene symbol from HGNC service



84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/exodb/datamodel/reference.rb', line 84

def dl_symbol!
  
  baseuri = "http://rest.genenames.org/search"
  
  
  query = ""
  
  if self.get_xref('urn:miriam:refseq')
    query = "#{baseuri}/refseq_accession/#{self.chrrefseq.id.split('.')[0]}"
  elsif self.get_xref('urn:miriam:ncbigene')
    query = ""
  end
  
  if !query.empty?
    response = JSON.parse(open(query, 'Accept' => 'application/json').read)['response']
    if !response['docs'].empty?
      
      response['docs'].each do |e|
        self.add_to_set(:xrefs, "urn:miriam:hgnc:#{e["hgnc_id"]}")
        self.add_to_set(:xrefs, "urn:miriam:hgnc.symbol:#{e["symbol"]}")
      end
      
      self.save!
      
    end
    
  end
  
end

#has_sequence? ⇒ Boolean

Check if Generef has sequence

Returns:

  • (Boolean) —

    Return true if there is a sequence



205
206
207
 
# File 'lib/exodb/datamodel/reference.rb', line 205

def has_sequence?()
  return self[:sequence] ? true : false
end

#has_splices? ⇒ Boolean

Check that this gene has any splice variant

Returns:

  • (Boolean) —

    true if has any splices



198
199
200
 
# File 'lib/exodb/datamodel/reference.rb', line 198

def has_splices?
  return self.isoforms.exists?
end

#longest_splice ⇒ Object

return longest splice of this gene



180
181
182
183
184
185
186
187
188
189
190
191
192
193
 
# File 'lib/exodb/datamodel/reference.rb', line 180

def longest_splice()
  length = 0
  longest = nil
  self.isoforms.each do |e|
    
    if e.prot_len > length
      length = e.prot_len
      longest = e
    end
    
  end
  
  return longest
end

#symbol ⇒ String

Get gene symbol

Returns:

  • (String) —

    Return gene symbol or any id from xrefs or ‘nosymbol’



219
220
221
222
223
224
225
226
227
228
229
 
# File 'lib/exodb/datamodel/reference.rb', line 219

def symbol
  
  if self.get_xref('urn:miriam:hgnc.symbol')
    return self.get_xref('urn:miriam:hgnc.symbol').id
  elsif self.xrefs && !self.xrefs.empty?
    return self.xrefs.sort[0].id
  else
    return 'nosymbol'
  end
  
end

#to_seq ⇒ Bio::Sequence

return sequence as Bio::Sequence object

Returns:

  • (Bio::Sequence) —

    the contents reversed lexically



175
176
177
 
# File 'lib/exodb/datamodel/reference.rb', line 175

def to_seq
  return self.sequence ? Bio::Sequence.auto(self.sequence) : Bio::Sequence.auto("")
end