Mercurial > repos > mvdbeek > tepid_merge_deletions

comparison merge_deletions.py @ 0:ff6683f8e9a1 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ListerLab/TEPID commit 82fd0448ff5baa9822a388aee78753e4b1cd94d7
author mvdbeek
date Mon, 23 Jan 2017 10:05:02 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:ff6683f8e9a1
1 #! /usr/bin/env python
2
3 import os
4 from argparse import ArgumentParser
5
6 def create_master_dict(sample, fname):
7 with open(fname, 'r') as masterfile:
8 x = 0
9 master_dict = {}
10 for line in masterfile:
11 field = line.rsplit()
12 if not line[0] == 'ins_chr':
13 coords = '\t'.join(field[:5])
14 master_dict[x] = {'coords': coords, 'accessions': [sample]}
15 x += 1
16 return master_dict
17
18
19 def merge_deletions(master, fname, sample):
20 with open(fname, 'r') as infile:
21 for line in infile:
22 field = line.rsplit()
23 coords = '\t'.join(field[:5])
24 i = len(master)-1
25 x = 0
26 while x <= i:
27 if master[x]['coords'] == coords:
28 master[x]['accessions'].append(sample)
29 break
30 elif x == i:
31 master[x+1] = {'coords': coords, 'accessions': [sample]}
32 break
33 else:
34 x += 1
35
36
37 def save_deletions(master, outf):
38 with open(outf, 'w+') as outfile:
39 for key, value in master.iteritems():
40 accessions = set(value['accessions'])
41 outfile.write('{c}\t{a}\n'.format(c=value['coords'], a=','.join(accessions)))
42
43 def get_name_from_filename(filename):
44 return os.path.basename(filename).rsplit('.', 1)[0]
45
46 if __name__ == "__main__":
47
48 parser = ArgumentParser(description='Merge TE deletions calls')
49 parser.add_argument('-o', '--output', help="File to write merged deletions to.", required=True)
50 parser.add_argument('-i', '--input', help='all files that should be merged', nargs="+", required=True)
51 options = parser.parse_args()
52
53 first_file = options.input[0]
54 first_samplename = get_name_from_filename(first_file)
55 master_dictionary = create_master_dict(first_samplename, first_file)
56 for filename in options.input[1:]:
57 samplename = get_name_from_filename(filename)
58 merge_deletions(master_dictionary, filename, samplename)
59 save_deletions(master_dictionary, options.output)