Source Code for Module Bio.Translate

  1  """Code to translate DNA or RNA into proteins (DEPRECATED). 
  2   
  3  Instead of Bio.Translate, for translation you are now encouraged to use the 
  4  Seq object's translate method, or the translate function in the Bio.Seq 
  5  module.  Translate-to-stop functionality is via an optional argument. 
  6   
  7  Bio.Seq does not offer any back-translation function like the one here. It 
  8  was concluded that a since a simple back-translation giving a Seq or python 
  9  string could only capture some of the possible back translations, there were 
 10  no practical uses for such a method/function. 
 11   
 12  This module is now deprecated, and will be removed in a future release of 
 13  Biopython. 
 14  """ 
 15  import warnings 
 16  import Bio 
 17  warnings.warn("Bio.Translate and Bio.Transcribe are deprecated, and will be "\ 
 18                "removed in a future release of Biopython. Please use the "\ 
 19                "functions or object methods defined in Bio.Seq instead "\ 
 20                "(described in the tutorial). If you want to continue to use "\ 
 21                "this code, please get in contact with the Biopython developers "\ 
 22                "via the mailing lists to avoid its permanent removal from " 
 23                +"Biopython.", \ 
 24                Bio.BiopythonDeprecationWarning) 
 25   
 26  from Bio import Alphabet, Seq 
 27  from Bio.Data import CodonTable 
 28   
 29 -class Translator: 
 30 -    def __init__(self, table): 
 31          self.table = table 
 32          self._encoded = {} 
 33   
 34 -    def __str__(self): 
 35          return "Translator object\n" + str(self.table) 
 36   
 37 -    def translate(self, seq, stop_symbol = "*"): 
 38          #Allow different instances of the same class to be used: 
 39          assert seq.alphabet.__class__ == \ 
 40                 self.table.nucleotide_alphabet.__class__, \ 
 41                 "cannot translate from given alphabet (have %s, need %s)" %\ 
 42                 (seq.alphabet, self.table.nucleotide_alphabet) 
 43          s = seq.data 
 44          letters = [] 
 45          append = letters.append 
 46          table = self.table 
 47          get = table.forward_table.get 
 48          n = len(seq) 
 49          for i in range(0, n-n%3, 3): 
 50              append(get(s[i:i+3], stop_symbol)) 
 51   
 52          # return with the correct alphabet encoding (cache the encoding) 
 53          try: 
 54              alphabet = self._encoded[stop_symbol] 
 55          except KeyError: 
 56              alphabet = Alphabet.HasStopCodon(table.protein_alphabet, 
 57                                               stop_symbol) 
 58              self._encoded[stop_symbol] = alphabet 
 59   
 60          return Seq.Seq("".join(letters), alphabet) 
 61                              
 62 -    def translate_to_stop(self, seq): 
 63          # This doesn't have a stop encoding 
 64   
 65          #Allow different instances of the same class to be used: 
 66          assert seq.alphabet.__class__ == \ 
 67                 self.table.nucleotide_alphabet.__class__, \ 
 68                 "cannot translate from given alphabet (have %s, need %s)" %\ 
 69                 (seq.alphabet, self.table.nucleotide_alphabet) 
 70          s = seq.data 
 71          letters = [] 
 72          append = letters.append 
 73          table = self.table.forward_table 
 74          n = len(seq) 
 75          try: 
 76              for i in range(0, n-n%3, 3): 
 77                  append(table[s[i:i+3]]) 
 78          except KeyError: 
 79              # Stop at the first codon failure 
 80              pass 
 81          return Seq.Seq("".join(letters), self.table.protein_alphabet) 
 82   
 83 -    def back_translate(self, seq): 
 84          # includes the stop codon 
 85          if not isinstance(seq.alphabet, Alphabet.HasStopCodon): 
 86              return self._back_translate_no_stop(seq) 
 87          assert seq.alphabet.alphabet == self.table.protein_alphabet, \ 
 88                 "cannot back translate from the given alphabet (%s)" % \ 
 89                 seq.alphabet.alphabet 
 90          s = seq.data 
 91          letter = seq.alphabet.stop_symbol 
 92          letters = [] 
 93          append = letters.append 
 94          table = self.table.back_table 
 95          for c in seq.data: 
 96              if c == letter: 
 97                  append(table[None]) 
 98              else: 
 99                  append(table[c]) 
100          return Seq.Seq("".join(letters), 
101                         self.table.nucleotide_alphabet) 
102   
103 -    def _back_translate_no_stop(self, seq): 
104          # does not allow a stop codon 
105          assert seq.alphabet == self.table.protein_alphabet, \ 
106                 "cannot back translate from the given alphabet (%s)" % \ 
107                 seq.alphabet 
108          s = seq.data 
109          letters = [] 
110          append = letters.append 
111          table = self.table.back_table 
112          for c in seq.data: 
113              append(table[c]) 
114          return Seq.Seq("".join(letters), 
115                         self.table.nucleotide_alphabet) 
116   
117  unambiguous_dna_by_name = {} 
118  for key, value in CodonTable.unambiguous_dna_by_name.items(): 
119      unambiguous_dna_by_name[key] = Translator(value) 
120  unambiguous_dna_by_id = {} 
121  for key, value in CodonTable.unambiguous_dna_by_id.items(): 
122      unambiguous_dna_by_id[key] = Translator(value) 
123   
124  unambiguous_rna_by_name = {} 
125  for key, value in CodonTable.unambiguous_rna_by_name.items(): 
126      unambiguous_rna_by_name[key] = Translator(value) 
127  unambiguous_rna_by_id = {} 
128  for key, value in CodonTable.unambiguous_rna_by_id.items(): 
129      unambiguous_rna_by_id[key] = Translator(value) 
130   
131  # XXX Ambiguous - can be done the same except for stop codons! 
132  ambiguous_dna_by_name = {} 
133  for key, value in CodonTable.ambiguous_dna_by_name.items(): 
134      ambiguous_dna_by_name[key] = Translator(value) 
135  ambiguous_dna_by_id = {} 
136  for key, value in CodonTable.ambiguous_dna_by_id.items(): 
137      ambiguous_dna_by_id[key] = Translator(value) 
138   
139  ambiguous_rna_by_name = {} 
140  for key, value in CodonTable.ambiguous_rna_by_name.items(): 
141      ambiguous_rna_by_name[key] = Translator(value) 
142  ambiguous_rna_by_id = {} 
143  for key, value in CodonTable.ambiguous_rna_by_id.items(): 
144      ambiguous_rna_by_id[key] = Translator(value) 
145