Mercurial > repos > iuc > snpeff

annotate gbk2fa.py @ 29:030fe29d4c47 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
author iuc
date Wed, 03 Aug 2022 16:33:10 +0000
parents b1b327f475ac
children b7029a54f73e
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
14
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
1 import argparse
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
2 import bz2
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
3 import gzip
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
4
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
5 from Bio import SeqIO
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
6
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
7
29
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
8 def get_opener(gbk_filename):
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
9 try:
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
10 bz2.open(gbk_filename).read(1)
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
11 return bz2.open
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
12 except OSError:
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
13 pass
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
14 try:
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
15 gzip.open(gbk_filename).read(1)
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
16 return gzip.open
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
17 except OSError:
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
18 return open
14
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
19
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
20
29
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
21 parser = argparse.ArgumentParser()
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
22 parser.add_argument(
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
23 "genbank_file",
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
24 help="GenBank input file. Can be compressed with gzip or bzip2"
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
25 )
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
26 parser.add_argument(
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
27 "fasta_file", help="FASTA output datset"
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
28 )
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
29 parser.add_argument(
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
30 "--remove_version", action="store_true",
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
31 help="Remove version number from NCBI form formatted accession numbers. "
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
32 "For example, this would convert 'B000657.2' to 'B000657'"
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
33 )
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
34 args = parser.parse_args()
14
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
35
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
36
29
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
37 gbk_open = get_opener(args.genbank_file)
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
38 with gbk_open(args.genbank_file, 'rt') as input_handle, \
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
39 open(args.fasta_file, 'w') as output_handle:
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
40 for seq_record in SeqIO.parse(input_handle, 'genbank'):
14
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
41 if args.remove_version:
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
42 seq_id = seq_record.id.split('.')[0]
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
43 else:
a7c106b938dd planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
44 seq_id = seq_record.id
28
b1b327f475ac "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit e4366b6a23223f84039a0590cf0d7079b83d8f84"
iuc
parents: 19
diff changeset
45 print('Writing FASTA record: {}'.format(seq_id))
29
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
46 print('>' + seq_id, file=output_handle)
030fe29d4c47 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 28
diff changeset
47 print(seq_record.seq, file=output_handle)