Mercurial > repos > vipints > fml_gff3togtf

annotate GFFtools-GX/bed_to_gff.py @ 3:ff2c2e6f4ab3

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded version 2.0.0 of gfftools ready to import to local instance
author vipints
date Wed, 11 Jun 2014 16:29:25 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
3
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
1 #!/usr/bin/env python
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
2 """
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
3 Convert genome annotation data in a 12 column BED format to GFF3.
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
4
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
5 Usage: python bed_to_gff.py in.bed > out.gff
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
6
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
7 Requirement:
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
8 helper.py : https://github.com/vipints/GFFtools-GX/blob/master/helper.py
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
9
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
10 Copyright (C)
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
11 2009-2012 Friedrich Miescher Laboratory of the Max Planck Society, Tubingen, Germany.
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
12 2012-2014 Memorial Sloan Kettering Cancer Center New York City, USA.
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
13 """
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
14
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
15 import re
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
16 import sys
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
17 import helper
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
18
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
19 def __main__():
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
20 """
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
21 main function
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
22 """
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
23
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
24 try:
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
25 bed_fname = sys.argv[1]
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
26 except:
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
27 print __doc__
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
28 sys.exit(-1)
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
29
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
30 bed_fh = helper.open_file(bed_fname)
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
31
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
32 for line in bed_fh:
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
33 line = line.strip( '\n\r' )
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
34
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
35 if not line or line[0] in ['#']:
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
36 continue
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
37
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
38 parts = line.split('\t')
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
39 assert len(parts) >= 12, line
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
40
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
41 rstarts = parts[-1].split(',')
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
42 rstarts.pop() if rstarts[-1] == '' else rstarts
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
43
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
44 exon_lens = parts[-2].split(',')
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
45 exon_lens.pop() if exon_lens[-1] == '' else exon_lens
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
46
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
47 if len(rstarts) != len(exon_lens):
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
48 continue # checking the consistency col 11 and col 12
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
49
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
50 if len(rstarts) != int(parts[-3]):
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
51 continue # checking the number of exons and block count are same
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
52
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
53 if not parts[5] in ['+', '-']:
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
54 parts[5] = '.' # replace the unknown strand with '.'
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
55
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
56 # bed2gff result line
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
57 print '%s\tbed2gff\tgene\t%d\t%s\t%s\t%s\t.\tID=Gene:%s;Name=Gene:%s' % (parts[0], int(parts[1])+1, parts[2], parts[4], parts[5], parts[3], parts[3])
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
58 print '%s\tbed2gff\ttranscript\t%d\t%s\t%s\t%s\t.\tID=%s;Name=%s;Parent=Gene:%s' % (parts[0], int(parts[1])+1, parts[2], parts[4], parts[5], parts[3], parts[3], parts[3])
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
59
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
60 st = int(parts[1])
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
61 for ex_cnt in range(int(parts[-3])):
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
62 start = st + int(rstarts[ex_cnt]) + 1
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
63 stop = start + int(exon_lens[ex_cnt]) - 1
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
64 print '%s\tbed2gff\texon\t%d\t%d\t%s\t%s\t.\tParent=%s' % (parts[0], start, stop, parts[4], parts[5], parts[3])
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
65
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
66 bed_fh.close()
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
67
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
68
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
69 if __name__ == "__main__":
ff2c2e6f4ab3 Uploaded version 2.0.0 of gfftools ready to import to local instance
vipints
parents:
diff changeset
70 __main__()