|
0
|
1 <tool id="crispr_recognition_tool" name="detect CRISPR sequences" version="0.1.5">
|
|
|
2 <description>(minced)</description>
|
|
|
3 <parallelism method="multi" split_inputs="input" split_mode="to_size" split_size="50" merge_outputs="output,output_region,output_ggf_full,output_ggf,output_fasta" />
|
|
|
4 <requirements>
|
|
|
5 <requirement type="package" version="0.1.5">minced</requirement>
|
|
|
6 <requirement type="set_environment">MINCED_SCRIPT_PATH</requirement>
|
|
|
7 </requirements>
|
|
|
8 <stdio>
|
|
|
9 <exit_code range=":-1" level="fatal" />
|
|
|
10 <exit_code range="1:" level="fatal" />
|
|
|
11 </stdio>
|
|
|
12 <command>
|
|
|
13 temp_output=\$(mktemp);
|
|
|
14 minced
|
|
|
15 -minNR $minNR
|
|
|
16 -minRL $minRL
|
|
|
17 -maxRL $maxRL
|
|
|
18 -minSL $minSL
|
|
|
19 -maxSL $maxSL
|
|
|
20 #if str($searchWL) != 'none':
|
|
|
21 -searchWL $searchWL
|
|
|
22 #end if
|
|
|
23 $input
|
|
|
24 \$temp_output
|
|
|
25 #if '-spacers' in str($outputs).split():
|
|
|
26 -spacers
|
|
|
27 #end if
|
|
|
28 #if '-gffFull' in str($outputs).split():
|
|
|
29 -gffFull
|
|
|
30 #end if
|
|
|
31 #if '-gff' in str($outputs).split():
|
|
|
32 -gff
|
|
|
33 #end if
|
|
|
34 ;
|
|
|
35
|
|
|
36 python \$MINCED_SCRIPT_PATH/reformat.py \$temp_output results.bed results.tab;
|
|
|
37 #if '-spacers' in str($outputs).split():
|
|
|
38 mv \$temp_output'_spacers.fa' $output_fasta;
|
|
|
39 #end if
|
|
|
40 </command>
|
|
|
41 <inputs>
|
|
|
42 <param format="fasta" name="input" type="data" label="Input sequence as FASTA"/>
|
|
|
43
|
|
|
44 <param name="minNR" type="integer" value="3" label="Minimum number of repeats a CRISPER must have"/>
|
|
|
45 <param name="minRL" type="integer" value="19" label="Minimum length of a CRISPER repeated region"/>
|
|
|
46 <param name="maxRL" type="integer" value="38" label="Maximum length of a CRISPER repeated region"/>
|
|
|
47 <param name="minSL" type="integer" value="19" label="Minimum length of CRISPER non repeated region" help="or spacer region"/>
|
|
|
48 <param name="maxSL" type="integer" value="48" label="Maximum length of CRISPER non repeated region" help="or spacer region"/>
|
|
|
49 <param name="searchWL" type="select" label="Length of search window to used to discover CRISPERs">
|
|
|
50 <option value="none" selected="true">default</option>
|
|
|
51 <option value="6">6</option>
|
|
|
52 <option value="7">7</option>
|
|
|
53 <option value="8">8</option>
|
|
|
54 <option value="9">9</option>
|
|
|
55 </param>
|
|
|
56
|
|
|
57 <param name="outputs" type="select" multiple="True" label="Additional output format" help="Please select a output format.">
|
|
|
58 <option value="-gff">summary results in gff format</option>
|
|
|
59 <option value="BED" selected="true">summary results in BED format</option>
|
|
|
60 <option value="-gffFull">detailed results in gff format</option>
|
|
|
61 <option value="TABULAR" selected="true">detailed results in tabular format</option>
|
|
|
62 <option value="-spacers">fasta formatted file containing the spacers</option>
|
|
|
63 <validator type="no_options" message="Please select at least one output file." />
|
|
|
64 </param>
|
|
|
65
|
|
|
66 </inputs>
|
|
|
67 <outputs>
|
|
|
68 <data format="tabular" name="output" from_work_dir="results.tab" label="${tool.name} on ${on_string}">
|
|
|
69 <filter>'TABULAR' in outputs</filter>
|
|
|
70 </data>
|
|
|
71 <data format="bed" name="output_region" from_work_dir="results.bed" label="${tool.name} on ${on_string} - BED file">
|
|
|
72 <filter>'BED' in outputs</filter>
|
|
|
73 </data>
|
|
|
74 <data format="gff3" name="output_ggf_full" from_work_dir="results_full.gff" label="${tool.name} on ${on_string} - GFF file extended">
|
|
|
75 <filter>'-gffFull' in outputs</filter>
|
|
|
76 </data>
|
|
|
77 <data format="gff3" name="output_gff" from_work_dir="results.gff" label="${tool.name} on ${on_string} - GFF file">
|
|
|
78 <filter>'-gff' in outputs</filter>
|
|
|
79 </data>
|
|
|
80 <data format="fasta" name="output_fasta" label="${tool.name} on ${on_string} - GFF file">
|
|
|
81 <filter>'-spacers' in outputs</filter>
|
|
|
82 </data>
|
|
|
83 </outputs>
|
|
|
84 <tests>
|
|
|
85 <test>
|
|
|
86 <param name="input" value="sequence.fasta" ftype="fasta" />
|
|
|
87 <param name="minNR" value="3" />
|
|
|
88 <param name="minRL" value="29" />
|
|
|
89 <param name="maxRL" value="38" />
|
|
|
90 <param name="minSL" value="19" />
|
|
|
91 <param name="maxSL" value="48" />
|
|
|
92 <param name="searchWL" value="none" />
|
|
|
93 <output name="output" file="minced.tabular" ftype="tabular" />
|
|
|
94 <output name="output_region" file="minced.bed" ftype="bed" />
|
|
|
95 </test>
|
|
|
96 </tests>
|
|
|
97 <help>
|
|
|
98 **What it does**
|
|
|
99
|
|
|
100 MinCED - Mining CRISPRs in Environmental Datasets
|
|
|
101
|
|
|
102
|
|
|
103 MinCED is a program to find Clustered Regularly Interspaced Short Palindromic
|
|
|
104 Repeats (CRISPRs) in full genomes or environmental datasets such as metagenomes,
|
|
|
105 in which sequence size can be anywhere from 100 to 800 bp. MinCED runs from the
|
|
|
106 command-line and was derived from CRT (http://www.room220.com/crt/).
|
|
|
107
|
|
|
108 https://github.com/ctSkennerton/minced
|
|
|
109
|
|
|
110 **Citation**
|
|
|
111
|
|
|
112 For the underlying tool, please cite `Bland C, Ramsey TL, Sabree F, Lowe M, Brown K, Kyrpides NC, Hugenholtz P:
|
|
|
113 CRISPR Recognition Tool (CRT): a tool for automatic detection of clustered regularly interspaced palindromic repeats. BMC Bioinformatics. 2007 Jun 18;8(1):209`
|
|
|
114
|
|
|
115 If you use this tool in Galaxy, please cite Gruening, BA et al. https://github.com/bgruening/galaxytools
|
|
|
116
|
|
|
117 </help>
|
|
|
118 </tool>
|
|
|
119
|
