annotate add_fst_column.py @ 0:4d28d3295ac3 default tip

Uploaded
author rico
date Fri, 06 Apr 2012 13:46:42 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
1 #!/usr/bin/env python
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
2
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
3 # <command interpreter="python">
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
4 # add_fst_column.py "$input" "$p1_input" "$p2_input" "$data_source.choice" "$data_source.min_value" "$retain" "$discard_fixed" "$biased" "$output"
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
5 # #for $individual, $individual_col in zip($input.dataset.metadata.individual_names, $input.dataset.metadata.individual_columns)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
6 # #set $arg = '%s:%s' % ($individual_col, $individual)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
7 # "$arg"
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
8 # #end for
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
9 # </command>
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
10
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
11 import sys
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
12 import subprocess
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
13 from Population import Population
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
14
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
15 ################################################################################
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
16
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
17 if len(sys.argv) < 12:
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
18 print >> sys.stderr, "Usage"
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
19 sys.exit(1)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
20
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
21 input, p1_input, p2_input, genotypes, min_reads, min_qual, retain, discard_fixed, biased, output = sys.argv[1:11]
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
22 individual_metadata = sys.argv[11:]
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
23
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
24 p_total = Population()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
25 p_total.from_tag_list(individual_metadata)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
26
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
27 p1 = Population()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
28 p1.from_population_file(p1_input)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
29 if not p_total.is_superset(p1):
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
30 print >> sys.stderr, 'There is an individual in population 1 that is not in the SNP table'
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
31 sys.exit(1)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
32
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
33 p2 = Population()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
34 p2.from_population_file(p2_input)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
35 if not p_total.is_superset(p2):
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
36 print >> sys.stderr, 'There is an individual in population 2 that is not in the SNP table'
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
37 sys.exit(1)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
38
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
39 ################################################################################
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
40
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
41 prog = 'Fst_column'
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
42
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
43 args = []
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
44 args.append(prog)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
45 args.append(input)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
46 args.append(genotypes)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
47 args.append(min_reads)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
48 args.append(min_qual)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
49 args.append(retain)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
50 args.append(discard_fixed)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
51 args.append(biased)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
52
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
53 columns = p1.column_list()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
54 for column in columns:
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
55 args.append('{0}:1'.format(column))
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
56
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
57 columns = p2.column_list()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
58 for column in columns:
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
59 args.append('{0}:2'.format(column))
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
60
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
61 fh = open(output, 'w')
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
62
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
63 #print "args:", ' '.join(args)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
64 p = subprocess.Popen(args, bufsize=-1, stdin=None, stdout=fh, stderr=sys.stderr)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
65 rc = p.wait()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
66 fh.close()
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
67
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
68 sys.exit(0)
4d28d3295ac3 Uploaded
rico
parents:
diff changeset
69