|
0
|
1 #!/usr/bin/python
|
|
|
2 import argparse
|
|
|
3 import sys
|
|
|
4 import os
|
|
|
5 import tempfile
|
|
|
6 import shutil
|
|
|
7 import subprocess
|
|
|
8 import ntpath
|
|
|
9
|
|
|
10 """
|
|
|
11 -h, --help show this help message and exit
|
|
|
12 -o PATH, --output-path PATH
|
|
|
13 Directory where output files will be written
|
|
|
14 -n NAME Analysis name
|
|
|
15 --output-format FORMAT
|
|
|
16 The FORMAT for the output file
|
|
|
17 -N NUMBER, --samplings NUMBER
|
|
|
18 Number of samplings to compute the FM bias pvalue
|
|
|
19 -e ESTIMATOR, --estimator ESTIMATOR
|
|
|
20 Test estimator for computation.
|
|
|
21 --gt THRESHOLD, --gene-threshold THRESHOLD
|
|
|
22 Minimum number of mutations per gene to compute the FM
|
|
|
23 bias
|
|
|
24 --pt THRESHOLD, --pathway-threshold THRESHOLD
|
|
|
25 Minimum number of mutations per pathway to compute the
|
|
|
26 FM bias
|
|
|
27 -s SLICES, --slices SLICES
|
|
|
28 Slices to process separated by commas
|
|
|
29 -m PATH, --mapping PATH
|
|
|
30 File with mappings between genes and pathways to be
|
|
|
31 analysed
|
|
|
32 --save-data The input data matrix will be saved
|
|
|
33 --save-analysis The analysis results will be saved
|
|
|
34 -j CORES, --cores CORES
|
|
|
35 Number of cores to use for calculations. Default is 0
|
|
|
36 that means all the available cores
|
|
|
37 -D KEY=VALUE Define external parameters to be saved in the results
|
|
|
38 -L LEVEL, --log-level LEVEL
|
|
|
39 Define log level: debug, info, warn, error, critical,
|
|
|
40 notset
|
|
|
41 """
|
|
|
42 def stop_err( msg ):
|
|
|
43 sys.stderr.write( '%s\n' % msg )
|
|
|
44 sys.exit()
|
|
|
45 def main(params):
|
|
|
46 parser = argparse.ArgumentParser()
|
|
|
47 ##TAKEN directly from the source code
|
|
|
48 parser.add_argument("-N", "--samplings", dest="num_samplings", type=int, default=10000, metavar="NUMBER",
|
|
|
49 help="Number of samplings to compute the FM bias pvalue")
|
|
|
50 parser.add_argument("-e", "--estimator", dest="estimator", metavar="ESTIMATOR",
|
|
|
51 choices=["mean", "median"], default="mean",
|
|
|
52 help="Test estimator for computation.")
|
|
|
53 parser.add_argument("--gt", "--gene-threshold", dest="mut_gene_threshold", type=int, default=2, metavar="THRESHOLD",
|
|
|
54 help="Minimum number of mutations per gene to compute the FM bias")
|
|
|
55 parser.add_argument("--pt", "--pathway-threshold", dest="mut_pathway_threshold", type=int, default=10, metavar="THRESHOLD",
|
|
|
56 help="Minimum number of mutations per pathway to compute the FM bias")
|
|
|
57 parser.add_argument("-s", "--slices", dest="slices", metavar="SLICES",
|
|
|
58 help="Slices to process separated by commas")
|
|
|
59 parser.add_argument("-m", "--mapping", dest="mapping", metavar="PATH",
|
|
|
60 help="File with mappings between genes and pathways to be analysed")
|
|
|
61 parser.add_argument("-f", "--filter", dest="filter", metavar="PATH",
|
|
|
62 help="File containing the features to be filtered. By default labels are includes,"
|
|
|
63 " labels preceded with - are excludes.")
|
|
|
64 #parser.add_argument("-o", "--output_path", type=str, required=True, help="Directory where output files will be written")
|
|
|
65 parser.add_argument("-o1", "--output1", type=str, dest="output1", required=True)
|
|
|
66
|
|
|
67 parser.add_argument("-o2", "--output2", type=str, dest="output2", required=False)
|
|
|
68 parser.add_argument("-n", "--analysis_name", type=str, required=False, help="Analysis name")
|
|
|
69 #parser.add_argument("-e", "--estimator", type=str, required=False, choices=["mean-empirical","median-empirical","mean-zscore","median-zscore"], help="Test estimator for computation")
|
|
|
70 parser.add_argument("--output-format", dest="output_format", required=False,
|
|
|
71 metavar="FORMAT",
|
|
|
72 choices=["tsv", "tsv.gz", "tsv.bz2"],
|
|
|
73 default="tsv",
|
|
|
74 help="The FORMAT for the output file")
|
|
|
75 parser.add_argument("-j", "--cores", dest="num_cores", type=int,
|
|
|
76 metavar="CORES",
|
|
|
77 help="Number of cores to use for calculations.\
|
|
|
78 Default is 0 that means all the available cores")
|
|
|
79 parser.add_argument("-D", dest="defines", metavar="KEY=VALUE", action="append", help="Define external parameters to be saved in the results")
|
|
|
80 parser.add_argument("-L", "--log-level", dest="log_level", metavar="LEVEL", default=None,
|
|
|
81 choices=["debug", "info", "warn", "error", "critical", "notset"],
|
|
|
82 help="Define log level: debug, info, warn, error, critical, notset")
|
|
|
83 parser.add_argument("-i", "--input", dest="input_path", required=True, type=str, help="Path to input file")
|
|
|
84 args = vars(parser.parse_args(params))
|
|
|
85 try:
|
|
|
86 mapping_path = args["mapping_path"]
|
|
|
87 except KeyError:
|
|
|
88 mapping_path = "no_mapping_path"
|
|
|
89 #if mapping_path=="no_mapping_path":
|
|
|
90 #params.remove(mapping_path)
|
|
|
91 #params.remove("-m")
|
|
|
92 output_dir = tempfile.mkdtemp()
|
|
|
93 params.append("-o")
|
|
|
94 params.append(output_dir)
|
|
|
95 params.append(args["input_path"])
|
|
|
96 cmd = "oncodrivefm "
|
|
|
97 i=0
|
|
|
98 while i<len(params):
|
|
|
99 p=params[i]
|
|
|
100 if p=="-i" or p=="-o1" or p=="-o2":
|
|
|
101 i+=2
|
|
|
102 else:
|
|
|
103 i+=1
|
|
|
104 cmd += " "+p
|
|
|
105 cmd += " 2>&1 "
|
|
|
106 #tmp = tempfile.NamedTemporaryFile( dir=output_dir ).name
|
|
|
107 #tmp_stderr = open( tmp, 'wb' )
|
|
|
108 print cmd
|
|
|
109 proc = subprocess.Popen(args=cmd, shell=True)
|
|
|
110 returncode = proc.wait()
|
|
|
111 #tmp_stderr.close()
|
|
|
112
|
|
|
113 if args['analysis_name'] is not None:
|
|
|
114 prefix = args["analysis_name"]
|
|
|
115 else:
|
|
|
116 ##refer: http://stackoverflow.com/a/8384788/756986
|
|
|
117 prefix = ntpath.basename(args["input_path"]).split(".")[0]
|
|
|
118 if args["mapping"] is not None:
|
|
|
119 pathway_file = prefix+"-pathways"
|
|
|
120 else:
|
|
|
121 pathway_file = None
|
|
|
122 output_format = args["output_format"]
|
|
|
123 genes_output_file_name = os.path.join(output_dir, prefix+"-genes."+output_format)
|
|
|
124 shutil.move(genes_output_file_name,args["output1"])
|
|
|
125 if pathway_file:
|
|
|
126 pathway_output_file_name = os.path.join(output_dir, pathway_file+"."+output_format)
|
|
|
127 shutil.move(pathway_output_file_name,args["output2"])
|
|
|
128 if os.path.exists( output_dir ):
|
|
|
129 shutil.rmtree( output_dir )
|
|
|
130 if __name__=="__main__":
|
|
|
131 main(sys.argv[1:])
|
