comparison tools/regVariation/featureCounter.py @ 0:9071e359b9a3

Uploaded
author xuebing
date Fri, 09 Mar 2012 19:37:19 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:9071e359b9a3
1 #!/usr/bin/env python
2 #Guruprasad Ananda
3 """
4 Calculate count and coverage of one query on another, and append the Coverage and counts to
5 the last four columns as bases covered, percent coverage, number of completely present features, number of partially present/overlapping features.
6
7 usage: %prog bed_file_1 bed_file_2 out_file
8 -1, --cols1=N,N,N,N: Columns for chr, start, end, strand in first file
9 -2, --cols2=N,N,N,N: Columns for chr, start, end, strand in second file
10 """
11 from galaxy import eggs
12 import pkg_resources
13 pkg_resources.require( "bx-python" )
14 import sys, traceback, fileinput
15 from warnings import warn
16 from bx.intervals.io import *
17 from bx.cookbook import doc_optparse
18 from bx.intervals.operations import quicksect
19 from galaxy.tools.util.galaxyops import *
20
21 assert sys.version_info[:2] >= ( 2, 4 )
22
23 def stop_err(msg):
24 sys.stderr.write(msg)
25 sys.exit()
26
27 def counter(node, start, end):
28 global full, partial
29 if node.start <= start and node.maxend > start:
30 if node.end >= end or (node.start == start and end > node.end > start):
31 full += 1
32 elif end > node.end > start:
33 partial += 1
34 if node.left and node.left.maxend > start:
35 counter(node.left, start, end)
36 if node.right:
37 counter(node.right, start, end)
38 elif start < node.start < end:
39 if node.end <= end:
40 full += 1
41 else:
42 partial += 1
43 if node.left and node.left.maxend > start:
44 counter(node.left, start, end)
45 if node.right:
46 counter(node.right, start, end)
47 else:
48 if node.left:
49 counter(node.left, start, end)
50
51 def count_coverage( readers, comments=True ):
52 primary = readers[0]
53 secondary = readers[1]
54 secondary_copy = readers[2]
55
56 rightTree = quicksect.IntervalTree()
57 for item in secondary:
58 if type( item ) is GenomicInterval:
59 rightTree.insert( item, secondary.linenum, item.fields )
60
61 bitsets = secondary_copy.binned_bitsets()
62
63 global full, partial
64
65 for interval in primary:
66 if type( interval ) is Header:
67 yield interval
68 if type( interval ) is Comment and comments:
69 yield interval
70 elif type( interval ) == GenomicInterval:
71 chrom = interval.chrom
72 start = int(interval.start)
73 end = int(interval.end)
74 full = 0
75 partial = 0
76 if chrom not in bitsets:
77 bases_covered = 0
78 percent = 0.0
79 full = 0
80 partial = 0
81 else:
82 bases_covered = bitsets[ chrom ].count_range( start, end-start )
83 if (end - start) == 0:
84 percent = 0
85 else:
86 percent = float(bases_covered) / float(end - start)
87 if bases_covered:
88 root = rightTree.chroms[chrom] #root node for the chrom tree
89 counter(root, start, end)
90 interval.fields.append(str(bases_covered))
91 interval.fields.append(str(percent))
92 interval.fields.append(str(full))
93 interval.fields.append(str(partial))
94 yield interval
95
96 def main():
97 options, args = doc_optparse.parse( __doc__ )
98
99 try:
100 chr_col_1, start_col_1, end_col_1, strand_col_1 = parse_cols_arg( options.cols1 )
101 chr_col_2, start_col_2, end_col_2, strand_col_2 = parse_cols_arg( options.cols2 )
102 in1_fname, in2_fname, out_fname = args
103 except:
104 stop_err( "Data issue: click the pencil icon in the history item to correct the metadata attributes." )
105
106 g1 = NiceReaderWrapper( fileinput.FileInput( in1_fname ),
107 chrom_col=chr_col_1,
108 start_col=start_col_1,
109 end_col=end_col_1,
110 strand_col=strand_col_1,
111 fix_strand=True )
112 g2 = NiceReaderWrapper( fileinput.FileInput( in2_fname ),
113 chrom_col=chr_col_2,
114 start_col=start_col_2,
115 end_col=end_col_2,
116 strand_col=strand_col_2,
117 fix_strand=True )
118 g2_copy = NiceReaderWrapper( fileinput.FileInput( in2_fname ),
119 chrom_col=chr_col_2,
120 start_col=start_col_2,
121 end_col=end_col_2,
122 strand_col=strand_col_2,
123 fix_strand=True )
124
125
126 out_file = open( out_fname, "w" )
127
128 try:
129 for line in count_coverage([g1,g2,g2_copy]):
130 if type( line ) is GenomicInterval:
131 out_file.write( "%s\n" % "\t".join( line.fields ) )
132 else:
133 out_file.write( "%s\n" % line )
134 except ParseError, exc:
135 out_file.close()
136 fail( str( exc ) )
137
138 out_file.close()
139
140 if g1.skipped > 0:
141 print skipped( g1, filedesc=" of 1st dataset" )
142 if g2.skipped > 0:
143 print skipped( g2, filedesc=" of 2nd dataset" )
144 elif g2_copy.skipped > 0:
145 print skipped( g2_copy, filedesc=" of 2nd dataset" )
146
147 if __name__ == "__main__":
148 main()