Mercurial > repos > iuc > snpeff

annotate gbk2fa.py @ 29:ca2b512e8d7c draft default tip

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 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
author iuc
date Mon, 21 Oct 2024 13:56:15 +0000
parents 5c7b70713fb5
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
10
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
1 import argparse
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
2 import bz2
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
3 import gzip
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
4
5b4ac70948d2 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
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
6
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
7
25
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
8 def get_opener(gbk_filename):
29
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
9 """Determines the appropriate opener for a given file, supporting
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
10 bzip2, gzip, or standard open.
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
11 """
25
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
12 try:
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
13 bz2.open(gbk_filename).read(1)
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
14 return bz2.open
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
15 except OSError:
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
16 pass
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
17 try:
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
18 gzip.open(gbk_filename).read(1)
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
19 return gzip.open
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
20 except OSError:
5c7b70713fb5 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit a4c17efb8ec4b3a315766f5b5602effa275fade3
iuc
parents: 24
diff changeset
21 return open
10
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
22
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
23
29
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
24 def main():
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
25 parser = argparse.ArgumentParser(
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
26 description="Convert GenBank files to FASTA format. "
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
27 "Supports gzip and bzip2 compressed files."
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
28 )
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
29 parser.add_argument(
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
30 "genbank_file",
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
31 help="GenBank input file. Can be compressed with gzip or bzip2"
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
32 )
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
33 parser.add_argument(
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
34 "fasta_file",
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
35 help="FASTA output dataset"
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
36 )
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
37 parser.add_argument(
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
38 "--remove_version", action="store_true",
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
39 help="Remove version number from NCBI formatted accession numbers. "
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
40 "For example, this converts 'B000657.2' to 'B000657'."
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
41 )
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
42 args = parser.parse_args()
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
43
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
44 gbk_open = get_opener(args.genbank_file)
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
45 with gbk_open(args.genbank_file, 'rt') as input_handle, \
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
46 open(args.fasta_file, 'w') as output_handle:
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
47 for seq_record in SeqIO.parse(input_handle, 'genbank'):
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
48 if args.remove_version:
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
49 seq_id = seq_record.id.split('.')[0]
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
50 else:
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
51 seq_id = seq_record.id
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
52 print(f'Writing FASTA record: {seq_id}')
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
53 output_handle.write(f'>{seq_id}\n')
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
54 output_handle.write(f'{seq_record.seq}\n')
10
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
55
5b4ac70948d2 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit eea43430ff90fe6b13b295f6d5efb2208401a7ef
iuc
parents:
diff changeset
56
29
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
57 if __name__ == "__main__":
ca2b512e8d7c planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tool_collections/snpeff commit 4cc10cfe702828a91ecf8bb58d3f84a36b0578f7
iuc
parents: 25
diff changeset
58 main()