Mercurial > repos > miller-lab > genome_diversity

comparison population_structure.py @ 0:2c498d40ecde

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author miller-lab
date Mon, 09 Apr 2012 12:03:06 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:2c498d40ecde
1 #!/usr/bin/env python
2
3 import errno
4 import os
5 import shutil
6 import subprocess
7 import sys
8 from BeautifulSoup import BeautifulSoup
9 import gd_composite
10
11 ################################################################################
12
13 def run_admixture(ped_file, populations):
14 prog = 'admixture'
15
16 args = []
17 args.append(prog)
18 args.append(input_ped_file)
19 args.append(populations)
20
21 #print "args:", ' '.join(args)
22 ofh = open('/dev/null', 'w')
23 p = subprocess.Popen(args, bufsize=-1, stdin=None, stdout=ofh, stderr=sys.stderr)
24 rc = p.wait()
25 ofh.close()
26
27 def run_r(input_file, output_file, populations):
28 prog = 'R'
29
30 args = []
31 args.append(prog)
32 args.append('--vanilla')
33 args.append('--quiet')
34 args.append('--args')
35 args.append(input_file)
36 args.append(output_file)
37 args.append(populations)
38
39 _realpath = os.path.realpath(__file__)
40 _script_dir = os.path.dirname(_realpath)
41 r_script_file = os.path.join(_script_dir, 'population_structure.r')
42
43 ifh = open(r_script_file)
44 ofh = open('/dev/null', 'w')
45 p = subprocess.Popen(args, bufsize=-1, stdin=ifh, stdout=ofh, stderr=None)
46 rc = p.wait()
47 ifh.close()
48 ofh.close()
49
50 def mkdir_p(path):
51 try:
52 os.makedirs(path)
53 except OSError, e:
54 if e.errno <> errno.EEXIST:
55 raise
56
57 def get_populations(input):
58 pops = []
59 pop_names = {}
60
61 with open(input) as fh:
62 soup = BeautifulSoup(fh)
63 misc = soup.find('div', {'id': 'gd_misc'})
64
65 return 'Populations\n{0}'.format(misc('ul')[0])
66
67 ################################################################################
68
69 if len(sys.argv) != 6:
70 print >> sys.stderr, "Usage"
71 sys.exit(1)
72
73 input_html_file, input_ped_file, output_file, extra_files_path, populations = sys.argv[1:6]
74 populations_html = get_populations(input_html_file)
75
76 run_admixture(input_ped_file, populations)
77
78 ped_base = os.path.basename(input_ped_file)
79 if ped_base.endswith('.ped'):
80 ped_base = ped_base[:-4]
81
82 p_file = '%s.%s.P' % (ped_base, populations)
83 q_file = '%s.%s.Q' % (ped_base, populations)
84
85 mkdir_p(extra_files_path)
86 numeric_output_file = os.path.join(extra_files_path, 'numeric.txt')
87 shutil.copy2(q_file, numeric_output_file)
88 os.remove(p_file)
89 os.remove(q_file)
90
91 graphical_output_file = os.path.join(extra_files_path, 'graphical.pdf')
92 run_r(numeric_output_file, graphical_output_file, populations)
93
94 ################################################################################
95
96 info_page = gd_composite.InfoPage()
97 info_page.set_title('Population structure Galaxy Composite Dataset')
98
99 display_file = gd_composite.DisplayFile()
100 display_value = gd_composite.DisplayValue()
101
102 out_pdf = gd_composite.Parameter(name='graphical.pdf', value='graphical.pdf', display_type=display_file)
103 out_txt = gd_composite.Parameter(name='numeric.txt', value='numeric.txt', display_type=display_file)
104
105 info_page.add_output_parameter(out_pdf)
106 info_page.add_output_parameter(out_txt)
107
108 in_pops = gd_composite.Parameter(description='Number of populations', value=populations, display_type=display_value)
109
110 info_page.add_input_parameter(in_pops)
111
112 misc_pops = gd_composite.Parameter(description=populations_html, display_type=display_value)
113
114 info_page.add_misc(misc_pops)
115
116
117 with open (output_file, 'w') as ofh:
118 print >> ofh, info_page.render()
119
120
121 sys.exit(0)