Mercurial > repos > nate > trinity_psc
annotate trinityrnaseq.xml @ 5:1de64ac61b4e draft default tip
Uploaded
| author | nate |
|---|---|
| date | Thu, 04 May 2017 14:16:09 -0400 |
| parents | f98f80675d60 |
| children |
| rev | line source |
|---|---|
| 0 | 1 <tool id="trinity_psc" name="Trinity" version="0.0.1"> |
| 2 | |
| 3 <!-- Written by Jeremy Goecks, now maintained here by bhaas and additional | |
| 4 modifications by Nate Coraor --> | |
| 1 | 5 <description>(Beta) De novo assembly of RNA-Seq data Using Trinity on PSC's Bridges</description> |
| 0 | 6 <requirements> |
| 1 | 7 <!-- These are versions available as modules on Bridges --> |
| 8 <requirement type="package" version="1.1.2">bowtie</requirement> | |
| 9 <requirement type="package" version="1.3">samtools</requirement> | |
| 0 | 10 <requirement type="package" version="jre7">java</requirement> |
|
4
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
11 <requirement type="package" version="5.18.4-threads">perl</requirement> |
| 1 | 12 <requirement type="package" version="2.2.0">trinity</requirement> |
| 0 | 13 </requirements> |
| 14 <command> | |
| 1 | 15 MEM=`expr "\${GALAXY_SLOTS:-16}" \* 48 - 16` ; |
| 16 | |
| 2 | 17 workdir=`pwd`; |
| 1 | 18 echo "workdir is \$workdir"; |
| 2 | 19 cd \$LOCAL; |
| 20 echo "Running Trinity from `pwd`"; | |
| 1 | 21 |
| 5 | 22 Trinity --no_version_check |
| 23 --max_memory "\${MEM}G" | |
| 0 | 24 --CPU "\${GALAXY_SLOTS:-16}" |
| 1 | 25 --bflyHeapSpaceMax "32G" |
| 26 --bflyHeapSpaceInit "2G" | |
| 27 --bflyGCThreads "6" | |
| 2 | 28 |
| 29 #if $additional_params.use_additional == "yes" and $additional_params.normalize_reads == "yes": | |
| 30 --normalize_reads | |
| 31 #end if | |
| 0 | 32 |
| 33 ## Inputs. | |
| 34 #if str($inputs.paired_or_single) == "paired": | |
| 35 --left $inputs.left_input --right $inputs.right_input | |
| 36 #if $inputs.left_input.ext == 'fa': | |
| 37 --seqType fa | |
| 38 #else: | |
| 39 --seqType fq | |
| 40 #end if | |
| 41 #if str($inputs.library_type) != "undefined": | |
| 42 --SS_lib_type $inputs.library_type | |
| 43 #end if | |
| 44 --group_pairs_distance $inputs.group_pairs_distance | |
| 45 #else: | |
| 46 --single $inputs.input | |
| 47 #if str($inputs.input.ext) == 'fa': | |
| 48 --seqType fa | |
| 49 #else: | |
| 50 --seqType fq | |
| 51 #end if | |
| 52 #if str($inputs.library_type) != "undefined": | |
| 53 --SS_lib_type $inputs.library_type | |
| 54 #end if | |
| 55 #end if | |
| 56 | |
| 57 ## Additional parameters. | |
| 58 #if str($additional_params.use_additional) == "yes": | |
| 59 --min_kmer_cov $additional_params.min_kmer_cov --max_reads_per_graph $additional_params.max_reads_per_graph | |
| 60 #if $additional_params.bfly_opts != 'None': | |
| 61 --bfly_opts " $additional_params.bfly_opts " | |
| 62 #end if | |
| 63 #end if | |
| 64 | |
| 65 ## direct to output | |
| 66 > $trinity_log 2>&1 | |
| 67 | |
|
4
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
68 ## if Trinity fails, output the end of the log to stderr for Galaxy, and touch the output file for Pulsar |
|
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
69 || (ec=\$? ; cp -p $trinity_log \$workdir; cd \$workdir; cat $trinity_log >&2 ; mkdir -p trinity_out_dir ; touch trinity_out_dir/Trinity.fasta ; exit \$ec); |
| 1 | 70 |
|
4
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
71 mkdir -p \$workdir/trinity_out_dir; |
|
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
72 cp -p trinity_out_dir/Trinity* \$workdir/trinity_out_dir; |
|
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
73 cd \$workdir; |
| 0 | 74 </command> |
| 75 <stdio> | |
| 76 <exit_code range="1:" level="fatal" description="Program failed" /> | |
| 77 <exit_code range=":-1" level="fatal" description="DRM killed job" /> | |
| 78 </stdio> | |
| 79 <inputs> | |
| 80 <conditional name="inputs"> | |
| 81 <param name="paired_or_single" type="select" label="Paired or Single-end data?"> | |
| 82 <option value="paired">Paired</option> | |
| 83 <option value="single">Single</option> | |
| 84 </param> | |
| 85 <when value="paired"> | |
| 86 <param format="fasta,fastq" name="left_input" type="data" label="Left/Forward strand reads" help=""/> | |
| 87 <param format="fasta,fastq" name="right_input" type="data" label="Right/Reverse strand reads" help=""/> | |
| 88 <param name="library_type" type="select" label="Strand-specific Library Type"> | |
| 89 <option value="undefined">Not set</option> | |
| 90 <option value="FR">FR</option> | |
| 91 <option value="RF">RF</option> | |
| 92 </param> | |
| 93 <param name="group_pairs_distance" type="integer" value="500" min="1" label="Group pairs distance" help="Maximum length expected between fragment pairs"/> | |
| 94 <param name="path_reinforcement_distance" type="integer" value="75" min="1" label="Path reinforcement distance" help="Minimum read overlap required for path extension in the graph" /> | |
| 95 </when> | |
| 96 <when value="single"> | |
| 97 <param format="fasta,fastq" name="input" type="data" label="Single-end reads" help=""/> | |
| 98 <param name="library_type" type="select" label="Strand-specific Library Type"> | |
| 99 <option value="undefined">Not set</option> | |
| 100 <option value="F">F</option> | |
| 101 <option value="R">R</option> | |
| 102 </param> | |
| 103 <param name="path_reinforcement_distance" type="integer" value="40" min="1" label="Path reinforcement distance" help="Minimum read overlap required for path extension in the graph" /> | |
| 104 </when> | |
| 105 </conditional> | |
| 106 | |
| 107 <conditional name="additional_params"> | |
| 108 <param name="use_additional" type="select" label="Use Additional Params?"> | |
| 109 <option value="no">No</option> | |
| 110 <option value="yes">Yes</option> | |
| 111 </param> | |
| 112 <when value="no"> | |
| 113 </when> | |
| 114 <when value="yes"> | |
| 115 <param name="min_kmer_cov" type="integer" value="1" min="1" label="inchworm_min_kmer_cov" help="Minimum kmer coverage required by Inchworm for initial contig construction" /> | |
| 116 <param name="max_reads_per_graph" type="integer" value="20000000" min="10000" label="chrysalis_max_reads_per_graph" help="Maximum number of reads to be anchored within each transcript graph by Chrysalis" /> | |
| 117 <param name="bfly_opts" type="text" value="None" label="bfly_opts" help="Options to pass on to Butterfly" /> | |
| 118 <param name="min_contig_length" type="integer" value="200" min="1" label="Minimum Contig Length" help=""/> | |
| 2 | 119 <param name="normalize_reads" type="boolean" truevalue="yes" falsevalue="no" help="(--normalize_reads) Normalize reads, can decrease runtime and memory requirements for datasets exceeding 300M pairs"/> |
| 0 | 120 </when> |
| 121 </conditional> | |
| 122 </inputs> | |
| 123 <outputs> | |
| 124 <data format="txt" name="trinity_log" label="${tool.name} on ${on_string}: log" /> | |
| 125 <data format="fasta" name="assembled_transcripts" label="${tool.name} on ${on_string}: Assembled Transcripts" from_work_dir="trinity_out_dir/Trinity.fasta"/> | |
| 126 </outputs> | |
| 127 <tests> | |
| 128 </tests> | |
| 129 <help> | |
|
4
f98f80675d60
Corrected perl module name and improved output handling for better performance from Phil Blood.
nate
parents:
3
diff
changeset
|
130 Trinity is a de novo transcript assembler that uses RNA-seq data as input. This tool runs all Trinity_ commands--Inchworm, Chrysalis, and Butterfly--in a single pass. This version of Trinity runs on Bridges_ at the `Pittsburgh Supercomputing Center`_ using a version of Trinity 2.2.0 optimized for the unique memory profile of that system. |
| 0 | 131 |
| 132 .. _Trinity: http://trinityrnaseq.github.io | |
| 133 .. _Pittsburgh Supercomputing Center: http://www.psc.edu | |
| 1 | 134 .. _Bridges: http://www.psc.edu/bridges |
| 0 | 135 </help> |
| 136 </tool> |