Class: BetterSam

Inherits:
Object
  • Object
show all
Defined in:
lib/bettersam.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(line = nil) ⇒ BetterSam

Returns a new instance of BetterSam.



26
27
28
29
 
# File 'lib/bettersam.rb', line 26

def initialize(line=nil)
  @tags = {}
  parse_line(line) unless line.nil?
end

Instance Attribute Details

#chrom ⇒ Object

Returns the value of attribute chrom.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def chrom
  @chrom
end

#cigar ⇒ Object

Returns the value of attribute cigar.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def cigar
  @cigar
end

#cigar_list ⇒ Object (readonly)

Returns the value of attribute cigar_list.



24
25
26
 
# File 'lib/bettersam.rb', line 24

def cigar_list
  @cigar_list
end

#flag ⇒ Object

Returns the value of attribute flag.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def flag
  @flag
end

#insert ⇒ Object

Returns the value of attribute insert.



22
23
24
 
# File 'lib/bettersam.rb', line 22

def insert
  @insert
end

#length ⇒ Object

Returns the value of attribute length.



22
23
24
 
# File 'lib/bettersam.rb', line 22

def length
  @length
end

#mapq ⇒ Object

Returns the value of attribute mapq.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def mapq
  @mapq
end

#mchrom ⇒ Object

Returns the value of attribute mchrom.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def mchrom
  @mchrom
end

#mpos ⇒ Object

Returns the value of attribute mpos.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def mpos
  @mpos
end

#name ⇒ Object

Returns the value of attribute name.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def name
  @name
end

#pos ⇒ Object

Returns the value of attribute pos.



21
22
23
 
# File 'lib/bettersam.rb', line 21

def pos
  @pos
end

#qual ⇒ Object

Returns the value of attribute qual.



22
23
24
 
# File 'lib/bettersam.rb', line 22

def qual
  @qual
end

#seq ⇒ Object

Returns the value of attribute seq.



22
23
24
 
# File 'lib/bettersam.rb', line 22

def seq
  @seq
end

#snp ⇒ Object

Returns the value of attribute snp.



23
24
25
 
# File 'lib/bettersam.rb', line 23

def snp
  @snp
end

#tags ⇒ Object

Returns the value of attribute tags.



22
23
24
 
# File 'lib/bettersam.rb', line 22

def tags
  @tags
end

Instance Method Details

#both_mapped? ⇒ Boolean

pair convenience methods

Returns:

  • (Boolean) 


119
120
121
 
# File 'lib/bettersam.rb', line 119

def both_mapped?
  !(self.read_unmapped? && self.mate_unmapped?)
end

#contains_snp?(snp) ⇒ Boolean

snp storing

Returns:

  • (Boolean) 


179
180
181
 
# File 'lib/bettersam.rb', line 179

def contains_snp?(snp)
  snp >= @pos and snp < self.endpos
end

#edit_distance ⇒ Object 



132
133
134
 
# File 'lib/bettersam.rb', line 132

def edit_distance
  @tags[:NM]
end

#endpos ⇒ Object 



147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
 
# File 'lib/bettersam.rb', line 147

def endpos
  if !@cigar_list
    self.parse_cigar
  end
  e = @pos
  @cigar_list.each do |h|
    a = h.to_a
    bases = a[0][0]
    match = a[0][1]
    if match =~ /[MD]/
      e += bases
    end
  end
  return e
end

#exact_match? ⇒ Boolean

cigar parsing methods

Returns:

  • (Boolean) 


143
144
145
 
# File 'lib/bettersam.rb', line 143

def exact_match?
  @tags[:NM]==0
end

#first_in_pair? ⇒ Boolean

Returns:

  • (Boolean) 


97
98
99
 
# File 'lib/bettersam.rb', line 97

def first_in_pair?
  @flag & $flags[7] != 0
end

#get_base_at(p) ⇒ Object 



255
256
257
 
# File 'lib/bettersam.rb', line 255

def get_base_at(p)
  @seq[p]
end

#mark_snp(snp) ⇒ Object 



183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
 
# File 'lib/bettersam.rb', line 183

def mark_snp(snp)
  if self.contains_snp?(snp)
    if !@cigar_list
      self.parse_cigar
    end
    p = @pos
    s = snp
    @cigar_list.each do |h|
      if p > s and s >= @pos
        @snp = s - @pos
      else
        a = h.to_a
        bases = a[0][0]
        match = a[0][1]
        if match == "M"
          p += bases
        elsif match == "I"
          s += bases
        elsif match == "D"
          s -= bases
        end
      end
    end
    if p > s and s >= @pos
      @snp = s - @pos
    end
  end
  @snp
end

#mate_reverse_strand? ⇒ Boolean

Returns:

  • (Boolean) 


93
94
95
 
# File 'lib/bettersam.rb', line 93

def mate_reverse_strand?
  @flag & $flags[6] != 0
end

#mate_unmapped? ⇒ Boolean

Returns:

  • (Boolean) 


85
86
87
 
# File 'lib/bettersam.rb', line 85

def mate_unmapped?
  @flag & $flags[4] != 0
end

#pair_opposite_strands? ⇒ Boolean

Returns:

  • (Boolean) 


123
124
125
126
 
# File 'lib/bettersam.rb', line 123

def pair_opposite_strands?
  (!self.read_reverse_strand? && self.mate_reverse_strand?) ||
    (self.read_reverse_strand? && !self.mate_reverse_strand?)
end

#pair_same_strand? ⇒ Boolean

Returns:

  • (Boolean) 


128
129
130
 
# File 'lib/bettersam.rb', line 128

def pair_same_strand?
  !self.pair_opposite_strands?
end

#parse_cigar ⇒ Object 



163
164
165
166
167
168
169
170
171
172
173
174
175
 
# File 'lib/bettersam.rb', line 163

def parse_cigar
  str = @cigar
  l = str.length
  @cigar_list = []
  while str.length>0
    if str =~ /([0-9]+[MIDNSHPX=]+)/
      @cigar_list << {$1[0..-2].to_i => $1[-1]}
      str = str.slice($1.length, l)
    else
      puts str
    end
  end
end

#parse_line(line) ⇒ Object 



31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
 
# File 'lib/bettersam.rb', line 31

def parse_line(line)
  return false if line[0] == "@"

  f = line.chomp.split("\t", -1)
  raise "SAM lines must have at least 11 fields (had #{f.size})" if f.size < 11

  # colnames = %w(1:name 2:flag 3:chr 4:pos 5:mapq 6:cigar 7:mchr 8:mpos 9:insrt 10:seq 11:qual)

  @name = f[0]
  @flag = int_or_raw(f[1])
  @chrom = f[2]
  @pos = int_or_neg1(f[3])
  @mapq = int_or_neg1(f[4])
  @cigar = f[5]
  @mchrom = f[6]
  @mpos = int_or_neg1(f[7])
  @insert = int_or_raw(f[8])
  @seq = f[9]
  @qual = f[10]
  @length = nil

  @tags = {}
  i = 11
  while i < f.size
    tag = f[i]
    i += 1
    a = tag.split(":")
    raise line if a.length != 3
    if a[1]=="i"
      @tags[a[0].to_sym] = a[2].to_i
    elsif a[1]=="Z"
      @tags[a[0].to_sym] = a[2]
    else
      @tags[a[0].to_sym] = a[2]
    end
  end

  return true;
end

#pcr_duplicate? ⇒ Boolean

Returns:

  • (Boolean) 


113
114
115
 
# File 'lib/bettersam.rb', line 113

def pcr_duplicate?
  @flag & $flags[11] != 0
end

#primary_aln? ⇒ Boolean

Returns:

  • (Boolean) 


105
106
107
 
# File 'lib/bettersam.rb', line 105

def primary_aln?
  (@flag & $flags[9]) == 0
end

#put_snp ⇒ Object

find the location of a snp on the genome



222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
 
# File 'lib/bettersam.rb', line 222

def put_snp # find the location of a snp on the genome
  if @snp
    if !@cigar_list
      self.parse_cigar
    end
    s = @snp
    p = 0
    @cigar_list.each do |h|
      if p > s
        return s+@pos
      else
        a = h.to_a
        bases = a[0][0]
        match = a[0][1]
        if match=="M"
          p += bases
        elsif match=="D"
          s += bases
        elsif match=="I"
          s -= bases
        end
      end
    end
    if p > s
      return s+@pos
    end
  else
    puts "need to run mark_snp and transfer_snp first"
    return nil
  end
  return -1
end

#quality_fail? ⇒ Boolean

Returns:

  • (Boolean) 


109
110
111
 
# File 'lib/bettersam.rb', line 109

def quality_fail?
  @flag & $flags[10] != 0
end

#read_paired? ⇒ Boolean

flag parsing convenience methods

Returns:

  • (Boolean) 


73
74
75
 
# File 'lib/bettersam.rb', line 73

def read_paired?
  @flag & $flags[1] != 0
end

#read_properly_paired? ⇒ Boolean

Returns:

  • (Boolean) 


77
78
79
 
# File 'lib/bettersam.rb', line 77

def read_properly_paired?
  @flag & $flags[2] != 0
end

#read_reverse_strand? ⇒ Boolean

Returns:

  • (Boolean) 


89
90
91
 
# File 'lib/bettersam.rb', line 89

def read_reverse_strand?
  @flag & $flags[5] != 0
end

#read_unmapped? ⇒ Boolean

Returns:

  • (Boolean) 


81
82
83
 
# File 'lib/bettersam.rb', line 81

def read_unmapped?
  @flag & $flags[3] != 0
end

#second_in_pair? ⇒ Boolean

Returns:

  • (Boolean) 


101
102
103
 
# File 'lib/bettersam.rb', line 101

def second_in_pair?
  @flag & $flags[8] !=0
end

#transfer_snp(bs) ⇒ Object

load in another bettersam object



213
214
215
216
217
218
219
220
 
# File 'lib/bettersam.rb', line 213

def transfer_snp(bs) # load in another bettersam object
  if !self.read_unmapped? and !bs.read_unmapped?
    if (self.read_reverse_strand? and bs.read_reverse_strand?) or (!self.read_reverse_strand? and !bs.read_reverse_strand?)
      @snp = bs.snp
    else
    end
  end
end