Mercurial > repos > miller-lab > genome_diversity

annotate offspring_heterozygosity_pedigree.py @ 32:03c22b722882

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
remove BeautifulSoup dependency
author Richard Burhans <burhans@bx.psu.edu>
date Fri, 20 Sep 2013 13:54:23 -0400
parents a631c2f6d913
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
31
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
1 #!/usr/bin/env python
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
2
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
3 import sys
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
4 import gd_util
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
5
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
6 from Population import Population
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
7
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
8 def load_and_check_pop(file, total_pop, name):
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
9 p = Population()
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
10 p.from_population_file(file)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
11 if not total_pop.is_superset(p):
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
12 gd_util.die('There is an individual in the {0} that is not in the SNP table'.format(name))
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
13 return p
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
14
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
15 def append_breeders_from_file(the_list, filename, kind):
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
16 with open(filename) as fh:
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
17 for line in fh:
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
18 elems = line.split()
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
19 breeder = elems[0].rstrip('\r\n')
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
20 the_list.append('{0}:{1}'.format(kind, breeder))
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
21
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
22 ################################################################################
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
23
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
24 if len(sys.argv) != 9:
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
25 gd_util.die('Usage')
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
26
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
27 input, input_type, pedigree, ind_arg, founders, b1_input, b2_input, output = sys.argv[1:]
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
28
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
29 p_total = Population()
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
30 p_total.from_wrapped_dict(ind_arg)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
31
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
32 f1 = load_and_check_pop(founders, p_total, 'founders')
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
33
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
34 ################################################################################
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
35
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
36 prog = 'offspring_heterozygosity2'
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
37
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
38 args = [ prog ]
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
39 args.append(input) # a Galaxy SNP table
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
40 args.append(pedigree) # a pedigree, where the SNP table is for the founders
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
41
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
42 for tag in f1.tag_list():
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
43 column, name = tag.split(':')
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
44 if type == 'gd_genotype':
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
45 column = int(column) - 2
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
46 tag = 'founder:{0}:{1}'.format(column, name)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
47 args.append(tag)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
48
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
49 append_breeders_from_file(args, b1_input, 0)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
50 append_breeders_from_file(args, b2_input, 1)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
51
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
52 with open(output, 'w') as fh:
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
53 gd_util.run_program(prog, args, stdout=fh)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
54
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
55 ################################################################################
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
56
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
57 sys.exit(0)
a631c2f6d913 Update to Miller Lab devshed revision 3c4110ffacc3
Richard Burhans <burhans@bx.psu.edu>
parents:
diff changeset
58