|
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 -->
|
|
|
5 <description>(Beta) De novo assembly of RNA-Seq data Using Trinity on PSC's Greenfield</description>
|
|
|
6 <requirements>
|
|
|
7 <!-- These are versions available as modules on Greenfield -->
|
|
|
8 <requirement type="package" version="1.1.1">bowtie</requirement>
|
|
|
9 <requirement type="package" version="1.1">samtools</requirement>
|
|
|
10 <requirement type="package" version="jre7">java</requirement>
|
|
|
11 <requirement type="package" version="2.0.6">trinity</requirement>
|
|
|
12 </requirements>
|
|
|
13 <command>
|
|
|
14 MEM=`expr "\${GALAXY_SLOTS:-15}" \* 50 - 16` ;
|
|
|
15 Trinity --max_memory "\${MEM}G"
|
|
|
16 --CPU "\${GALAXY_SLOTS:-16}"
|
|
|
17 --bflyHeapSpaceMax "\${MEM}G"
|
|
|
18
|
|
|
19 ## Inputs.
|
|
|
20 #if str($inputs.paired_or_single) == "paired":
|
|
|
21 --left $inputs.left_input --right $inputs.right_input
|
|
|
22 #if $inputs.left_input.ext == 'fa':
|
|
|
23 --seqType fa
|
|
|
24 #else:
|
|
|
25 --seqType fq
|
|
|
26 #end if
|
|
|
27 #if str($inputs.library_type) != "undefined":
|
|
|
28 --SS_lib_type $inputs.library_type
|
|
|
29 #end if
|
|
|
30 --group_pairs_distance $inputs.group_pairs_distance
|
|
|
31 #else:
|
|
|
32 --single $inputs.input
|
|
|
33 #if str($inputs.input.ext) == 'fa':
|
|
|
34 --seqType fa
|
|
|
35 #else:
|
|
|
36 --seqType fq
|
|
|
37 #end if
|
|
|
38 #if str($inputs.library_type) != "undefined":
|
|
|
39 --SS_lib_type $inputs.library_type
|
|
|
40 #end if
|
|
|
41 #end if
|
|
|
42
|
|
|
43 ## Additional parameters.
|
|
|
44 #if str($additional_params.use_additional) == "yes":
|
|
|
45 --min_kmer_cov $additional_params.min_kmer_cov --max_reads_per_graph $additional_params.max_reads_per_graph
|
|
|
46 #if $additional_params.bfly_opts != 'None':
|
|
|
47 --bfly_opts " $additional_params.bfly_opts "
|
|
|
48 #end if
|
|
|
49 #end if
|
|
|
50
|
|
|
51 ## direct to output
|
|
|
52 > $trinity_log 2>&1
|
|
|
53
|
|
|
54 ## if Trinity fails, output the end of the log to stderr for Galaxy, and touch the output file for Pulsar
|
|
|
55 || (ec=\$? ; cat $trinity_log >&2 ; mkdir -p trinity_out_dir ; touch trinity_out_dir/Trinity.fasta ; exit \$ec)
|
|
|
56
|
|
|
57 </command>
|
|
|
58 <stdio>
|
|
|
59 <exit_code range="1:" level="fatal" description="Program failed" />
|
|
|
60 <exit_code range=":-1" level="fatal" description="DRM killed job" />
|
|
|
61 </stdio>
|
|
|
62 <inputs>
|
|
|
63 <conditional name="inputs">
|
|
|
64 <param name="paired_or_single" type="select" label="Paired or Single-end data?">
|
|
|
65 <option value="paired">Paired</option>
|
|
|
66 <option value="single">Single</option>
|
|
|
67 </param>
|
|
|
68 <when value="paired">
|
|
|
69 <param format="fasta,fastq" name="left_input" type="data" label="Left/Forward strand reads" help=""/>
|
|
|
70 <param format="fasta,fastq" name="right_input" type="data" label="Right/Reverse strand reads" help=""/>
|
|
|
71 <param name="library_type" type="select" label="Strand-specific Library Type">
|
|
|
72 <option value="undefined">Not set</option>
|
|
|
73 <option value="FR">FR</option>
|
|
|
74 <option value="RF">RF</option>
|
|
|
75 </param>
|
|
|
76 <param name="group_pairs_distance" type="integer" value="500" min="1" label="Group pairs distance" help="Maximum length expected between fragment pairs"/>
|
|
|
77 <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" />
|
|
|
78 </when>
|
|
|
79 <when value="single">
|
|
|
80 <param format="fasta,fastq" name="input" type="data" label="Single-end reads" help=""/>
|
|
|
81 <param name="library_type" type="select" label="Strand-specific Library Type">
|
|
|
82 <option value="undefined">Not set</option>
|
|
|
83 <option value="F">F</option>
|
|
|
84 <option value="R">R</option>
|
|
|
85 </param>
|
|
|
86 <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" />
|
|
|
87 </when>
|
|
|
88 </conditional>
|
|
|
89
|
|
|
90 <conditional name="additional_params">
|
|
|
91 <param name="use_additional" type="select" label="Use Additional Params?">
|
|
|
92 <option value="no">No</option>
|
|
|
93 <option value="yes">Yes</option>
|
|
|
94 </param>
|
|
|
95 <when value="no">
|
|
|
96 </when>
|
|
|
97 <when value="yes">
|
|
|
98 <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" />
|
|
|
99 <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" />
|
|
|
100 <param name="bfly_opts" type="text" value="None" label="bfly_opts" help="Options to pass on to Butterfly" />
|
|
|
101 <param name="min_contig_length" type="integer" value="200" min="1" label="Minimum Contig Length" help=""/>
|
|
|
102 </when>
|
|
|
103 </conditional>
|
|
|
104 </inputs>
|
|
|
105 <outputs>
|
|
|
106 <data format="txt" name="trinity_log" label="${tool.name} on ${on_string}: log" />
|
|
|
107 <data format="fasta" name="assembled_transcripts" label="${tool.name} on ${on_string}: Assembled Transcripts" from_work_dir="trinity_out_dir/Trinity.fasta"/>
|
|
|
108 </outputs>
|
|
|
109 <tests>
|
|
|
110 </tests>
|
|
|
111 <help>
|
|
|
112 .. warning:: This version of Trinity, which runs on Greenfield_ at the `Pittsburgh Supercomputing Center`_, is a **beta** version. It may not be possible to rerun the same version of this tool, Trinity itself, or its other dependencies (Bowtie, SAMtools) in the future. **When rerunning this tool on the same data, it should, but cannot be guaranteed to reproduce the exact same results to the level of certainty as other Galaxy tools.**
|
|
|
113
|
|
|
114 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.
|
|
|
115
|
|
|
116 .. _Trinity: http://trinityrnaseq.github.io
|
|
|
117 .. _Pittsburgh Supercomputing Center: http://www.psc.edu
|
|
|
118 .. _Greenfield: http://www.psc.edu/index.php/resources-for-users/computing-resources/greenfield
|
|
|
119 </help>
|
|
|
120 </tool>
|
