Mercurial > repos > artbio > blast_unmatched

annotate blast_unmatched.py @ 0:f3b63b59a1ea draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
author artbio
date Tue, 03 Oct 2017 07:19:17 -0400
parents
children 50c1fa95a076
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
1 #!/usr/bin/env python3
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
2
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
3 import optparse
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
4
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
5
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
6 def parse_options():
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
7 """
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
8 Parse the options guiven to the script
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
9 """
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
10 parser = optparse.OptionParser(description='Get unmatched blast queries')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
11 parser.add_option('-f','--fasta', dest='fasta_file', help='Query fasta file\
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
12 used during blast')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
13 parser.add_option('-b','--blast', dest='blast_file', help='Blast tabular\
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
14 output (queries in 1rst column)')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
15 parser.add_option('-o','--output', dest='output_file', help='Output file name')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
16 (options, args) = parser.parse_args()
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
17 if len(args) > 0:
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
18 parser.error('Wrong number of arguments')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
19 return options
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
20
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
21 def get_matched(blast_file):
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
22 """
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
23 Get a dictionary of all the queries that got a match
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
24 """
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
25 matched = dict()
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
26 blast_file_handle = open(blast_file, 'r')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
27 for line in blast_file_handle.readlines():
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
28 fields = line.split("\t")
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
29 query_id = fields[0]
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
30 matched[query_id] = 1
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
31 blast_file_handle.close()
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
32 return matched
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
33
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
34 def get_unmatched(output_file, fasta_file, matched):
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
35 """
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
36 Compares matched queries to query fasta file and print unmatched to ouput
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
37 """
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
38 output_file_handle = open(output_file, 'w')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
39 fasta_file_handle = open(fasta_file, 'r')
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
40 unmatched = False
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
41 for line in fasta_file_handle.readlines():
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
42 if line.startswith('>'):
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
43 subline = line[1:100].rstrip() #qid are 100chars long in blast
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
44 if subline not in matched:
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
45 output_file_handle.write(line)
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
46 unmatched = True
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
47 else:
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
48 unmatched = False
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
49 elif unmatched:
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
50 output_file_handle.write(line)
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
51 fasta_file_handle.close()
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
52 output_file_handle.close()
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
53
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
54 def __main__():
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
55 opts = parse_options()
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
56 matched = get_matched(opts.blast_file)
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
57 get_unmatched(opts.output_file, opts.fasta_file, matched)
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
58
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
59 if __main__():
f3b63b59a1ea planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/blast_unmatched commit 5bd801feb838592fbb1f6dd68b5f1a480042da40
artbio
parents:
diff changeset
60 __main__()