cufflinks: cufflinks_wrapper.py annotate

annotate cufflinks_wrapper.py @ 9:a1ea9af8d5f4 draft

planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7

author	devteam
date	Wed, 11 Nov 2015 12:34:49 -0500
parents	64698e16f4a6
children	83bec71c5c9f

rev	line source
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	1 #!/usr/bin/env python
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	2
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	3 import optparse
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	4 import os
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	5 import shutil
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	6 import subprocess
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	7 import sys
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	8 import tempfile
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	9 from galaxy.datatypes.util.gff_util import parse_gff_attributes, gff_attributes_to_str
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	10
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	11
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	12 def stop_err( msg ):
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	13 sys.exit( "%s\n" % msg )
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	14
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	15
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	16 def __main__():
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	17 #Parse Command Line
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	18 parser = optparse.OptionParser()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	19 parser.add_option( '-1', '--input', dest='input', help=' file of RNA-Seq read alignments in the SAM format. SAM is a standard short read alignment, that allows aligners to attach custom tags to individual alignments, and Cufflinks requires that the alignments you supply have some of these tags. Please see Input formats for more details.' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	20 parser.add_option( '-I', '--max-intron-length', dest='max_intron_len', help='The minimum intron length. Cufflinks will not report transcripts with introns longer than this, and will ignore SAM alignments with REF_SKIP CIGAR operations longer than this. The default is 300,000.' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	21 parser.add_option( '-F', '--min-isoform-fraction', dest='min_isoform_fraction', help='After calculating isoform abundance for a gene, Cufflinks filters out transcripts that it believes are very low abundance, because isoforms expressed at extremely low levels often cannot reliably be assembled, and may even be artifacts of incompletely spliced precursors of processed transcripts. This parameter is also used to filter out introns that have far fewer spliced alignments supporting them. The default is 0.05, or 5% of the most abundant isoform (the major isoform) of the gene.' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	22 parser.add_option( '-j', '--pre-mrna-fraction', dest='pre_mrna_fraction', help='Some RNA-Seq protocols produce a significant amount of reads that originate from incompletely spliced transcripts, and these reads can confound the assembly of fully spliced mRNAs. Cufflinks uses this parameter to filter out alignments that lie within the intronic intervals implied by the spliced alignments. The minimum depth of coverage in the intronic region covered by the alignment is divided by the number of spliced reads, and if the result is lower than this parameter value, the intronic alignments are ignored. The default is 5%.' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	23 parser.add_option( '-p', '--num-threads', dest='num_threads', help='Use this many threads to align reads. The default is 1.' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	24 parser.add_option( '-G', '--GTF', dest='GTF', help='Tells Cufflinks to use the supplied reference annotation to estimate isoform expression. It will not assemble novel transcripts, and the program will ignore alignments not structurally compatible with any reference transcript.' )
8 64698e16f4a6 Uploaded devteam parents: 7 diff changeset	25 parser.add_option("--compatible-hits-norm", dest='compatible_hits_norm', action="store_true", help='Count hits compatible with reference RNAs only')
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	26 parser.add_option( '-g', '--GTF-guide', dest='GTFguide', help='use reference transcript annotation to guide assembly' )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	27 parser.add_option("--3-overhang-tolerance", dest='three_overhang_tolerance', help='The number of bp allowed to overhang the 3prime end of a reference transcript when determining if an assembled transcript should be merged with it (ie, the assembled transcript is not novel). The default is 600 bp.')
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	28 parser.add_option("--intron-overhang-tolerance", dest='intron_overhang_tolerance', help='The number of bp allowed to enter the intron of a reference transcript when determining if an assembled transcript should be merged with it (ie, the assembled transcript is not novel). The default is 50 bp.')
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	29 parser.add_option("--no-faux-reads", dest='no_faux_reads', help='This option disables tiling of the reference transcripts with faux reads. Use this if you only want to use sequencing reads in assembly but do not want to output assembled transcripts that lay within reference transcripts. All reference transcripts in the input annotation will also be included in the output.')
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	30 parser.add_option( '-u', '--multi-read-correct', dest='multi_read_correct', action="store_true", help='Tells Cufflinks to do an initial estimation procedure to more accurately weight reads mapping to multiple locations in the genome')
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	31
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	32 # Normalization options.
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	33 parser.add_option( "--no-effective-length-correction", dest="no_effective_length_correction", action="store_true" )
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	34 parser.add_option( "--no-length-correction", dest="no_length_correction", action="store_true" )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	35
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	36 # Wrapper / Galaxy options.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	37 parser.add_option( '-A', '--assembled-isoforms-output', dest='assembled_isoforms_output_file', help='Assembled isoforms output file; formate is GTF.' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	38
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	39 # Advanced Options:
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	40 parser.add_option( "--library-type", dest="library_type", help=' library prep used for input reads, default fr-unstranded')
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	41 parser.add_option( '-M', '--mask-file', dest='mask_file', help='Tells Cufflinks to ignore all reads that could have come from transcripts in this GTF file. \
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	42 We recommend including any annotated rRNA, mitochondrial transcripts other abundant transcripts \
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	43 you wish to ignore in your analysis in this file. Due to variable efficiency of mRNA enrichment \
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	44 methods and rRNA depletion kits, masking these transcripts often improves the overall robustness \
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	45 of transcript abundance estimates.')
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	46 parser.add_option( '-m', '--inner-mean-dist', dest='inner_mean_dist', help='This is the expected (mean) inner distance between mate pairs. \
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	47 For, example, for paired end runs with fragments selected at 300bp, \
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	48 where each end is 50bp, you should set -r to be 200. The default is 45bp.') # cufflinks: --frag-len-mean
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	49
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	50 parser.add_option( '-s', '--inner-dist-std-dev', dest='inner_dist_std_dev', help='The standard deviation for the distribution on inner distances between mate pairs. The default is 20bp.' ) # cufflinks: --frag-len-std-dev
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	51 parser.add_option( '--max-mle-iterations', dest='max_mle_iterations', help='Sets the number of iterations allowed during maximum likelihood estimation of abundances. Default: 5000' )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	52 parser.add_option( '--junc-alpha', dest='junc_alpha', help='Alpha value for the binomial test used during false positive spliced alignment filtration. Default: 0.001' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	53 parser.add_option( '--small-anchor-fraction', dest='small_anchor_fraction', help='Spliced reads with less than this percent of their length on each side of\
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	54 the junction are considered suspicious and are candidates for filtering prior to assembly. Default: 0.09.' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	55 parser.add_option( '--overhang-tolerance', dest='overhang_tolerance', help='The number of bp allowed to enter the intron of a transcript when determining if a \
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	56 read or another transcript is mappable to/compatible with it. The default is 8 bp based on the default bowtie/TopHat parameters.' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	57 parser.add_option( '--max-bundle-length', dest='max_bundle_length', help='Maximum genomic length of a given bundle" help="Default: 3,500,000bp' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	58 parser.add_option( '--max-bundle-frags', dest='max_bundle_frags', help='Sets the maximum number of fragments a locus may have before being skipped. Skipped loci are listed in skipped.gtf. Default: 1,000,000' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	59 parser.add_option( '--min-intron-length', dest='min_intron_length', help='Minimal allowed intron size. Default: 50' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	60 parser.add_option( '--trim-3-avgcov-thresh', dest='trim_three_avgcov_thresh', help='Minimum average coverage required to attempt 3prime trimming. Default: 10' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	61 parser.add_option( '--trim-3-dropoff-frac', dest='trim_three_dropoff_frac', help='The fraction of average coverage below which to trim the 3prime end of an assembled transcript. Default: 0.1' )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	62
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	63 # Bias correction options.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	64 parser.add_option( '-b', dest='do_bias_correction', action="store_true", help='Providing Cufflinks with a multifasta file via this option instructs it to run our new bias detection and correction algorithm which can significantly improve accuracy of transcript abundance estimates.')
2 da11bfc10e81 Update to the new data table specification. Dave Bouvier <dave@bx.psu.edu> parents: 0 diff changeset	65 parser.add_option( '', '--index', dest='index', help='The path of the reference genome' )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	66 parser.add_option( '', '--ref_file', dest='ref_file', help='The reference dataset from the history' )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	67
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	68 # Global model (for trackster).
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	69 parser.add_option( '', '--global_model', dest='global_model_file', help='Global model used for computing on local data' )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	70
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	71 (options, args) = parser.parse_args()
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	72
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	73 # output version # of tool
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	74 try:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	75 tmp = tempfile.NamedTemporaryFile().name
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	76 tmp_stdout = open( tmp, 'wb' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	77 proc = subprocess.Popen( args='cufflinks --no-update-check 2>&1', shell=True, stdout=tmp_stdout )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	78 tmp_stdout.close()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	79 returncode = proc.wait()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	80 stdout = None
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	81 for line in open( tmp_stdout.name, 'rb' ):
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	82 if line.lower().find( 'cufflinks v' ) >= 0:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	83 stdout = line.strip()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	84 break
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	85 if stdout:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	86 sys.stdout.write( '%s\n' % stdout )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	87 else:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	88 raise Exception
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	89 except:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	90 sys.stdout.write( 'Could not determine Cufflinks version\n' )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	91
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	92 # If doing bias correction, set/link to sequence file.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	93 if options.do_bias_correction:
2 da11bfc10e81 Update to the new data table specification. Dave Bouvier <dave@bx.psu.edu> parents: 0 diff changeset	94 if options.ref_file:
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	95 # Sequence data from history.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	96 # Create symbolic link to ref_file so that index will be created in working directory.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	97 seq_path = "ref.fa"
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	98 os.symlink( options.ref_file, seq_path )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	99 else:
2 da11bfc10e81 Update to the new data table specification. Dave Bouvier <dave@bx.psu.edu> parents: 0 diff changeset	100 if not os.path.exists( options.index ):
da11bfc10e81 Update to the new data table specification. Dave Bouvier <dave@bx.psu.edu> parents: 0 diff changeset	101 stop_err( 'Reference genome %s not present, request it by reporting this error.' % options.index )
da11bfc10e81 Update to the new data table specification. Dave Bouvier <dave@bx.psu.edu> parents: 0 diff changeset	102 seq_path = options.index
da11bfc10e81 Update to the new data table specification. Dave Bouvier <dave@bx.psu.edu> parents: 0 diff changeset	103
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	104 # Build command.
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	105
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	106 # Base; always use quiet mode to avoid problems with storing log output.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	107 cmd = "cufflinks -q --no-update-check"
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	108
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	109 # Add options.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	110 if options.max_intron_len:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	111 cmd += ( " -I %i" % int( options.max_intron_len ) )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	112 if options.min_isoform_fraction:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	113 cmd += ( " -F %f" % float( options.min_isoform_fraction ) )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	114 if options.pre_mrna_fraction:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	115 cmd += ( " -j %f" % float( options.pre_mrna_fraction ) )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	116 if options.num_threads:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	117 cmd += ( " -p %i" % int( options.num_threads ) )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	118 if options.GTF:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	119 cmd += ( " -G %s" % options.GTF )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	120 if options.compatible_hits_norm:
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	121 cmd += ( " --compatible-hits-norm" )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	122 if options.GTFguide:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	123 cmd += ( " -g %s" % options.GTFguide )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	124 cmd += ( " --3-overhang-tolerance %i" % int( options.three_overhang_tolerance ) )
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	125 cmd += ( " --intron-overhang-tolerance %i" % int( options.intron_overhang_tolerance ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	126 if options.no_faux_reads:
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	127 cmd += ( " --no-faux-reads" )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	128 if options.multi_read_correct:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	129 cmd += ( " -u" )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	130
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	131 if options.library_type and options.library_type != 'auto':
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	132 cmd += ( " --library-type %s" % options.library_type)
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	133 if options.mask_file:
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	134 cmd += ( " --mask-file %s" % options.mask_file )
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	135 if options.inner_mean_dist:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	136 cmd += ( " -m %i" % int( options.inner_mean_dist ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	137 if options.inner_dist_std_dev:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	138 cmd += ( " -s %i" % int( options.inner_dist_std_dev ) )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	139 if options.max_mle_iterations:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	140 cmd += ( " --max-mle-iterations %i" % int( options.max_mle_iterations ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	141 if options.junc_alpha:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	142 cmd += ( " --junc-alpha %f" % float( options.junc_alpha) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	143 if options.small_anchor_fraction:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	144 cmd += ( " --small-anchor-fraction %f" % float(options.small_anchor_fraction ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	145 if options.overhang_tolerance:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	146 cmd += ( " --overhang-tolerance %i" % int( options.overhang_tolerance ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	147 if options.max_bundle_length:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	148 cmd += ( " --max-bundle-length %i" % int( options.max_bundle_length ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	149 if options.max_bundle_frags:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	150 cmd += ( " --max-bundle-frags %i" % int( options.max_bundle_frags ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	151 if options.min_intron_length:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	152 cmd += ( " --min-intron-length %i" % int( options.min_intron_length ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	153 if options.trim_three_avgcov_thresh:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	154 cmd += ( " --trim-3-avgcov-thresh %i" % int( options.trim_three_avgcov_thresh ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	155 if options.trim_three_dropoff_frac:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	156 cmd += ( " --trim-3-dropoff-frac %f" % float( options.trim_three_dropoff_frac ) )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	157
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	158 if options.do_bias_correction:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	159 cmd += ( " -b %s" % seq_path )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	160 if options.no_effective_length_correction:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	161 cmd += ( " --no-effective-length-correction" )
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	162 if options.no_length_correction:
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	163 cmd += ( " --no-length-correction" )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	164
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	165 # Add input files.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	166 cmd += " " + options.input
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	167
7 5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	168 # Debugging.
5346d5eea8b1 Uploaded devteam parents: 2 diff changeset	169 print cmd
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	170
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	171 #
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	172 # Run command and handle output.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	173 #
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	174 try:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	175 #
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	176 # Run command.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	177 #
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	178 tmp_name = tempfile.NamedTemporaryFile( dir="." ).name
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	179 tmp_stderr = open( tmp_name, 'wb' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	180 proc = subprocess.Popen( args=cmd, shell=True, stderr=tmp_stderr.fileno() )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	181 returncode = proc.wait()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	182 tmp_stderr.close()
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	183
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	184 # Error checking.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	185 if returncode != 0:
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	186 raise Exception("return code = %i" % returncode)
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	187
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	188 #
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	189 # Handle output.
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	190 #
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	191
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	192 # Read standard error to get total map/upper quartile mass.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	193 total_map_mass = -1
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	194 tmp_stderr = open( tmp_name, 'r' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	195 for line in tmp_stderr:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	196 if line.lower().find( "map mass" ) >= 0 or line.lower().find( "upper quartile" ) >= 0:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	197 total_map_mass = float( line.split(":")[1].strip() )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	198 break
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	199 tmp_stderr.close()
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	200
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	201 #
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	202 # If there's a global model provided, use model's total map mass
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	203 # to adjust FPKM + confidence intervals.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	204 #
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	205 if options.global_model_file:
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	206 # Global model is simply total map mass from original run.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	207 global_model_file = open( options.global_model_file, 'r' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	208 global_model_total_map_mass = float( global_model_file.readline() )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	209 global_model_file.close()
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	210
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	211 # Ratio of global model's total map mass to original run's map mass is
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	212 # factor used to adjust FPKM.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	213 fpkm_map_mass_ratio = total_map_mass / global_model_total_map_mass
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	214
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	215 # Update FPKM values in transcripts.gtf file.
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	216 transcripts_file = open( "transcripts.gtf", 'r' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	217 tmp_transcripts = tempfile.NamedTemporaryFile( dir="." ).name
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	218 new_transcripts_file = open( tmp_transcripts, 'w' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	219 for line in transcripts_file:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	220 fields = line.split( '\t' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	221 attrs = parse_gff_attributes( fields[8] )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	222 attrs[ "FPKM" ] = str( float( attrs[ "FPKM" ] ) * fpkm_map_mass_ratio )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	223 attrs[ "conf_lo" ] = str( float( attrs[ "conf_lo" ] ) * fpkm_map_mass_ratio )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	224 attrs[ "conf_hi" ] = str( float( attrs[ "conf_hi" ] ) * fpkm_map_mass_ratio )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	225 fields[8] = gff_attributes_to_str( attrs, "GTF" )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	226 new_transcripts_file.write( "%s\n" % '\t'.join( fields ) )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	227 transcripts_file.close()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	228 new_transcripts_file.close()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	229 shutil.copyfile( tmp_transcripts, "transcripts.gtf" )
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	230
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	231 # TODO: update expression files as well.
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	232
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	233 # Set outputs. Transcript and gene expression handled by wrapper directives.
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	234 shutil.copyfile( "transcripts.gtf", options.assembled_isoforms_output_file )
0 b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	235 if total_map_mass > -1:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	236 f = open( "global_model.txt", 'w' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	237 f.write( "%f\n" % total_map_mass )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	238 f.close()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	239 except Exception, e:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	240 # Read stderr so that it can be reported:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	241 tmp_stderr = open( tmp_name, 'rb' )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	242 stderr = ''
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	243 buffsize = 1048576
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	244 try:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	245 while True:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	246 stderr += tmp_stderr.read( buffsize )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	247 if not stderr or len( stderr ) % buffsize != 0:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	248 break
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	249 except OverflowError:
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	250 pass
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	251 tmp_stderr.close()
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	252 stop_err( 'Error running cufflinks.\n%s\n%s' % ( str( e ), stderr ) )
b50aacc8ae49 Uploaded tool tarball. devteam parents: diff changeset	253
9 a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	254 if __name__ == "__main__":
a1ea9af8d5f4 planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tool_collections/cufflinks/cufflinks commit a1517c9d22029095120643bbe2c8fa53754dd2b7 devteam parents: 8 diff changeset	255 __main__()

Mercurial > repos > devteam > cufflinks

annotate cufflinks_wrapper.py @ 9:a1ea9af8d5f4 draft