annotate cuffmerge_wrapper.py @ 3:3ed34b8bddfc draft

planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit 82ee6fc860c52c531b7a57bbb346ab1a67a434a5
author devteam
date Sun, 19 Feb 2017 12:11:53 -0500
parents 0ed8b7f6d506
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
1 #!/usr/bin/env python
e556721b9872 Uploaded
devteam
parents:
diff changeset
2
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
3 import optparse
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
4 import os
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
5 import shutil
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
6 import subprocess
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
7 import sys
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
8 import tempfile
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
9
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
10
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
11 def stop_err(msg):
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
12 sys.stderr.write('%s\n' % msg)
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
13 sys.exit()
e556721b9872 Uploaded
devteam
parents:
diff changeset
14
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
15
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
16 def __main__():
e556721b9872 Uploaded
devteam
parents:
diff changeset
17 parser = optparse.OptionParser()
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
18 parser.add_option('-g', dest='ref_annotation', help='An optional "reference" annotation GTF. Each sample is matched against this file, and sample isoforms are tagged as overlapping, matching, or novel where appropriate. See the refmap and tmap output file descriptions below.')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
19 parser.add_option('-s', dest='use_seq_data', action="store_true", help='Causes cuffmerge to look into for fasta files with the underlying genomic sequences (one file per contig) against which your reads were aligned for some optional classification functions. For example, Cufflinks transcripts consisting mostly of lower-case bases are classified as repeats. Note that <seq_dir> must contain one fasta file per reference chromosome, and each file must be named after the chromosome, and have a .fa or .fasta extension.')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
20 parser.add_option('-p', '--num-threads', dest='num_threads', help='Use this many threads to align reads. The default is 1.')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
21
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
22 # Wrapper / Galaxy options.
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
23 parser.add_option('', '--index', dest='index', help='The path of the reference genome')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
24 parser.add_option('', '--ref_file', dest='ref_file', help='The reference dataset from the history')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
25
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
26 # Outputs.
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
27 parser.add_option('', '--merged-transcripts', dest='merged_transcripts')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
28 parser.add_option('--min-isoform-fraction', dest='min_isoform_fraction')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
29
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
30 (options, args) = parser.parse_args()
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
31
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
32 # Set/link to sequence file.
e556721b9872 Uploaded
devteam
parents:
diff changeset
33 if options.use_seq_data:
e556721b9872 Uploaded
devteam
parents:
diff changeset
34 if options.ref_file:
e556721b9872 Uploaded
devteam
parents:
diff changeset
35 # Sequence data from history.
e556721b9872 Uploaded
devteam
parents:
diff changeset
36 # Create symbolic link to ref_file so that index will be created in working directory.
e556721b9872 Uploaded
devteam
parents:
diff changeset
37 seq_path = "ref.fa"
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
38 os.symlink(options.ref_file, seq_path)
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
39 else:
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
40 if not os.path.exists(options.index):
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
41 stop_err('Reference genome %s not present, request it by reporting this error.' % options.index)
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
42 seq_path = options.index
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
43
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
44 # Build command.
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
45
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
46 # Base.
e556721b9872 Uploaded
devteam
parents:
diff changeset
47 cmd = "cuffmerge -o cm_output "
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
48
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
49 # Add options.
e556721b9872 Uploaded
devteam
parents:
diff changeset
50 if options.num_threads:
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
51 cmd += (" -p %i " % int(options.num_threads))
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
52 if options.ref_annotation:
e556721b9872 Uploaded
devteam
parents:
diff changeset
53 cmd += " -g %s " % options.ref_annotation
e556721b9872 Uploaded
devteam
parents:
diff changeset
54 if options.use_seq_data:
e556721b9872 Uploaded
devteam
parents:
diff changeset
55 cmd += " -s %s " % seq_path
e556721b9872 Uploaded
devteam
parents:
diff changeset
56 if options.min_isoform_fraction:
e556721b9872 Uploaded
devteam
parents:
diff changeset
57 cmd += " --min-isoform-fraction %s " % (options.min_isoform_fraction)
e556721b9872 Uploaded
devteam
parents:
diff changeset
58 # Add input files to a file.
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
59 with tempfile.NamedTemporaryFile(mode='w', dir=".", delete=False) as inputs_file:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
60 for arg in args:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
61 inputs_file.write(arg + "\n")
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
62 cmd += inputs_file.name
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
63
e556721b9872 Uploaded
devteam
parents:
diff changeset
64 # Run command.
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
65 try:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
66 with tempfile.NamedTemporaryFile(dir=".") as tmp_stderr:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
67 returncode = subprocess.call(args=cmd, stderr=tmp_stderr, shell=True)
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
68
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
69 # Error checking.
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
70 if returncode != 0:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
71 # Get stderr, allowing for case where it's very large.
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
72 buffsize = 1048576
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
73 stderr = ''
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
74 with open(tmp_stderr.name, 'r') as tmp_stderr2:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
75 try:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
76 while True:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
77 stderr += tmp_stderr2.read(buffsize)
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
78 if not stderr or len(stderr) % buffsize != 0:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
79 break
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
80 except OverflowError:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
81 pass
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
82 raise Exception(stderr)
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
83
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
84 if len(open("cm_output/merged.gtf", 'r').read().strip()) == 0:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
85 raise Exception('The output file is empty, there may be an error with your input file or settings.')
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
86
0
e556721b9872 Uploaded
devteam
parents:
diff changeset
87 # Copy outputs.
2
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
88 shutil.move("cm_output/merged.gtf", options.merged_transcripts)
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
89 except Exception as e:
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
90 stop_err('Error running cuffmerge: %s' % e)
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
91
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
92
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
93 if __name__ == "__main__":
0ed8b7f6d506 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cuffmerge commit eb18f691975ef9539b5ebd4f118343c8ad967a1f
devteam
parents: 0
diff changeset
94 __main__()