|
5
|
1 #!/usr/bin/env python
|
|
|
2 """
|
|
|
3 Convert data from GFF and associated genome sequence in fasta file into GenBank.
|
|
|
4
|
|
|
5 Usage:
|
|
|
6 python gff_to_gbk.py in.gff in.fasta out.gbk
|
|
|
7
|
|
|
8 Requirements:
|
|
|
9 BioPython:- http://biopython.org/
|
|
|
10 helper.py : https://github.com/vipints/GFFtools-GX/blob/master/helper.py
|
|
|
11
|
|
|
12 Copyright (C)
|
|
|
13 2010-2012 Friedrich Miescher Laboratory of the Max Planck Society, Tubingen, Germany.
|
|
|
14 2012-2014 Memorial Sloan Kettering Cancer Center New York City, USA.
|
|
|
15 """
|
|
|
16
|
|
|
17 import sys
|
|
|
18 import helper
|
|
|
19 import gffparser_bcbio
|
|
|
20
|
|
|
21 from Bio import SeqIO
|
|
|
22 from Bio.Alphabet import generic_dna
|
|
|
23
|
|
|
24 def __main__():
|
|
|
25 """
|
|
|
26 main wrapper
|
|
|
27 """
|
|
|
28
|
|
|
29 try:
|
|
|
30 gff_fname = sys.argv[1]
|
|
|
31 fasta_fname = sys.argv[2]
|
|
|
32 gb_fname = sys.argv[3]
|
|
|
33 except:
|
|
|
34 print __doc__
|
|
|
35 sys.exit(-1)
|
|
|
36
|
|
|
37 fasta_fh = helper.open_file(fasta_fname)
|
|
|
38
|
|
|
39 fasta_rec = SeqIO.to_dict(SeqIO.parse(fasta_fh, "fasta", generic_dna))
|
|
|
40 fasta_fh.close()
|
|
|
41
|
|
|
42 gff_rec = gffparser_bcbio.parse(gff_fname, fasta_rec)
|
|
|
43
|
|
|
44 try:
|
|
|
45 gb_fh = open(gb_fname, "w")
|
|
|
46 except:
|
|
|
47 print 'file not ready for writing %s' % gb_fname
|
|
|
48 sys.exit(-1)
|
|
|
49
|
|
|
50 SeqIO.write(gff_rec, gb_fh, "genbank")
|
|
|
51 gb_fh.close()
|
|
|
52
|
|
|
53 if __name__=="__main__":
|
|
|
54 __main__()
|
