Mercurial > repos > artbio > bamparse

annotate bamparse.py @ 0:b3107800eea7 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/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
author artbio
date Mon, 17 Jul 2023 01:02:01 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
1 #!/usr/bin/env python
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
2 import argparse
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
3 from collections import defaultdict
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
4
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
5 import pysam
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
6
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
7
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
8 def Parser():
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
9 the_parser = argparse.ArgumentParser()
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
10 the_parser.add_argument('--output', nargs='+', action='store', type=str,
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
11 help='Count tables')
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
12 the_parser.add_argument('--alignments', nargs='+',
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
13 help="bam alignments files")
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
14 the_parser.add_argument('--labels', nargs='+', help="Alignments labels")
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
15 the_parser.add_argument('--number',
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
16 choices=["unique", "multiple"],
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
17 help="output is a single table or multiple tables")
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
18 args = the_parser.parse_args()
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
19 return args
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
20
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
21
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
22 def get_counts(bamfile):
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
23 """
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
24 Takes an AlignmentFile object and returns a dictionary of counts for sense,
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
25 antisense, or both sense and antisense bam alignments to the references,
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
26 depending on the pre-treatment performed by sambamba in the xml wrapper
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
27 """
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
28 counts = defaultdict(int)
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
29 for ref_name in bamfile.references:
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
30 counts[ref_name] = 0
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
31 for ref_name in bamfile.references:
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
32 counts[ref_name] = bamfile.count(reference=ref_name)
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
33 return counts
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
34
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
35
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
36 def writetable(diclist, labels, output, number):
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
37 ''' diclist is a list of count dictionnaries '''
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
38 countlists = []
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
39 for dic in diclist:
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
40 counts = sorted(dic.items())
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
41 counts = [j for (i, j) in counts]
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
42 countlists.append(counts)
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
43 if number == "unique":
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
44 out = open("outputdir/table.tabular", "w")
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
45 out.write("gene\t%s\n" % "\t".join(labels))
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
46 for countline in zip(sorted(diclist[0]), *countlists):
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
47 line = [str(i) for i in countline]
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
48 out.write("%s\n" % "\t".join(line))
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
49 out.close()
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
50 else:
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
51 for i, (dic, label) in enumerate(zip(diclist, labels)):
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
52 out = open("outputdir/table" + str(i) + ".tabular", "w")
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
53 out.write("gene\t%s\n" % label)
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
54 for gene in sorted(dic):
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
55 out.write("%s\t%s\n" % (gene, dic[gene]))
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
56 out.close()
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
57
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
58
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
59 def main(alignments, labels, output, number):
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
60 diclist = []
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
61 for file in alignments:
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
62 bam_object = pysam.AlignmentFile(file, 'rb')
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
63 diclist.append(get_counts(bam_object))
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
64 writetable(diclist, labels, output, number)
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
65
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
66
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
67 if __name__ == "__main__":
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
68 args = Parser()
b3107800eea7 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/bamparse commit e8a19ac6ada887e6daa0a2e2abc9ba69392cdb8a
artbio
parents:
diff changeset
69 main(args.alignments, args.labels, args.output, args.number)