|
16
|
1 <?xml version="1.0"?>
|
|
|
2 <tool name="BloomTree Manager - Create" id="btman_create" version="1.0.0">
|
|
|
3 <description>a Sequence Bloom Tree</description>
|
|
|
4 <macros>
|
|
|
5 <import>macros.xml</import>
|
|
|
6 </macros>
|
|
|
7 <expand macro="requirements" />
|
|
|
8 <command detect_errors="exit_code">
|
|
|
9 <![CDATA[
|
|
|
10 python '$__tool_directory__/create.py'
|
|
|
11
|
|
|
12 --explist '${explist}'
|
|
|
13
|
|
|
14 --qualitycontrol ${conditional_quality.quality_control}
|
|
|
15 #if $conditional_quality.quality_control == '0':
|
|
|
16 --qualitythreshold 0.0
|
|
|
17 #elif $conditional_quality.quality_control == '1':
|
|
|
18 --qualitythreshold ${conditional_quality.quality_threshold}
|
|
|
19 #end if
|
|
|
20
|
|
|
21 --klen ${kmer_len}
|
|
|
22 --minabundance ${min_abundance}
|
|
|
23
|
|
|
24 --outfile '${outfile}'
|
|
|
25 ]]>
|
|
|
26 </command>
|
|
|
27 <inputs>
|
|
|
28 <param format="tabular" name="explist" type="data" label="Select a file with the list of experiments" help="This should be a tabular file with two columns. Take a look at the tool documentation for a detailed explanation about the its content." />
|
|
|
29
|
|
|
30 <conditional name="conditional_quality">
|
|
|
31 <param name="quality_control" type="boolean" checked="false" truevalue="1" falsevalue="0" label="Apply a quality control procedure" />
|
|
|
32 <when value="1">
|
|
|
33 <param name="quality_threshold" size="1" type="float" value="0.8" min="0.0" max="1.0" label="Quality threshold" help="If the number of sequences flagged as poor quality on the total number of sequences in a file is less than this threshold, the sequence file will be excluded." />
|
|
|
34 </when>
|
|
|
35 </conditional>
|
|
|
36
|
|
|
37 <param name="kmer_len" type="integer" value="21" min="0" label="K-mer length" />
|
|
|
38 <param name="min_abundance" type="integer" value="2" min="0" label="Bloom filter minimum abundance" help="This value is the minimum abundance cutoff for the creation of the Bloom filters. It is worth noting that the same minimum abundance is used for each Bloom filter." />
|
|
|
39 </inputs>
|
|
|
40 <outputs>
|
|
|
41 <data format="txt" name="outfile" label="${tool.name} SBT: Result" from_work_dir="btman.create.txt" />
|
|
|
42 </outputs>
|
|
|
43
|
|
|
44 <help><![CDATA[
|
|
|
45 This tool is part of the BloomTree Manager Framework that allow to create a Sequence Bloom Tree starting
|
|
|
46 with a set of FASTA or FASTQ files. It allows also to control the quality of the input dataset and
|
|
|
47 exclude the files that do not reach a specified quality level.
|
|
|
48
|
|
|
49 -----
|
|
|
50
|
|
|
51 **Input file**
|
|
|
52
|
|
|
53 The input file for this tool must contain two columns with their values delimited by a tab.
|
|
|
54 The first column contains a list of SRA accessions, and the second column contains a unique identifier
|
|
|
55 for each set of SRA accessions.
|
|
|
56
|
|
|
57 The input file is structured like the example below::
|
|
|
58
|
|
|
59 SRR805782 blood
|
|
|
60 SRR837459 blood
|
|
|
61 SRR837458 blood
|
|
|
62 SRR837453 blood
|
|
|
63 SRR837456 blood
|
|
|
64 ...
|
|
|
65 SRR791048 breast
|
|
|
66 SRR553483 breast
|
|
|
67 SRR553482 breast
|
|
|
68 SRR791045 breast
|
|
|
69 ...
|
|
|
70 SRR950876 brain
|
|
|
71 SRR786621 brain
|
|
|
72
|
|
|
73 It is worth noting that for each cluster of accessions, every accession should be unique.
|
|
|
74 It is indeed possible to repeat an accession in multiple clusters.
|
|
|
75
|
|
|
76 The tool will create a Sequence Bloom Tree for each cluster of accessions.
|
|
|
77
|
|
|
78 -----
|
|
|
79
|
|
|
80 **Output**
|
|
|
81
|
|
|
82 The tool returns a single text file only. It contains the a tree identifier, one for
|
|
|
83 each cluster of accessions specified in the input file, that can be used with the
|
|
|
84 Query tool of the BloomTree Manager Suite to search for the presence of a set of
|
|
|
85 specific transcripts.
|
|
|
86
|
|
|
87 Take a look at the Query tool documentation for a detailed description about how
|
|
|
88 to query a Sequence Bloom Tree.
|
|
|
89
|
|
|
90 -----
|
|
|
91
|
|
|
92 .. class:: infomark
|
|
|
93
|
|
|
94 **Notes**
|
|
|
95
|
|
|
96 This Galaxy tool has been developed by Fabio Cumbo.
|
|
|
97
|
|
|
98 Please visit this GithHub_repository_ for more information about the BloomTree Manager
|
|
|
99
|
|
|
100 .. _GithHub_repository: https://github.com/fabio-cumbo/bloomtree-manager
|
|
|
101 ]]></help>
|
|
|
102
|
|
|
103 <expand macro="citations" />
|
|
|
104 </tool>
|
