view cat_summarise.xml @ 0:db1360aae489 draft

"planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit 863ad85836c80811d1d6b82eaf3ce903b273368a"

<tool id="cat_summarise" name="CAT summarise" version="@VERSION@.0">
    <description>the number of contigs or bins assigned to each taxonomic name</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <expand macro="version_command" />
    <command detect_errors="exit_code"><![CDATA[
    CAT summarise
        #if $contigs_fasta
        -c '$contigs_fasta'
        #end if
        -i '$input'
        -o 'output_names_summary.txt'
    && @TXT2TSV@ -i 'output_names_summary.txt' -o '$output'
    ]]></command>
    <inputs>
        <param argument="--input" type="data" format="tabular" label="classification.names.txt output from CAT add_names"
            help="The classication must be made with --only_official names"/>
        <param argument="--contigs_fasta" type="data" format="fasta" optional="true" label="contigs.fasta used for CAT contigs" 
            help="Required if summarising a contig2classification.names output derived from CAT contigs"/>
    </inputs>
    <outputs>
        <data name="output" format="tabular" label="#set name = $input.name.replace('.txt','.summary.txt')# ${tool.name} on ${on_string} $name"/>
    </outputs>
    <tests>
        <test>
            <param name="input" ftype="tabular" value="test_contig.contig2classification.names.txt"/>
            <param name="contigs_fasta" ftype="fasta" value="contigs.fasta"/>
            <output name="output">
                <assert_contents>
                    <has_text text="Firmicutes" />
                </assert_contents>
            </output>
        </test>
    </tests>
    <help><![CDATA[
**CAT summarise** 

Summarise taxonomic asignemts from a CAT or BAT classification file that has official taxonomic names added by CAT add_names.

**INPUT**

::

  # bin   classification  reason  lineage lineage scores  superkingdom    phylum  class   order   family  genus   species
  genome2.fna     classified      based on 6/17 ORFs      1;131567;2;1224;1236;135623;641;662;666 1.00;1.00;1.00;0.89;0.89;0.89;0.89;0.89;0.69    Bacteria: 1.00  Proteobacteria: 0.89    Gammaproteobacteria: 0.89       Vibrionales: 0.89       Vibrionaceae: 0.89      Vibrio: 0.89    Vibrio cholerae: 0.69
  genome3.fna     classified      based on 4/9 ORFs       1;131567;2;1783272;1239;91061;1385;186820;1637;1639     1.00;1.00;1.00;0.76;0.76;0.76;0.76;0.76;0.76;0.76       Bacteria: 1.00  Firmicutes: 0.76        Bacilli: 0.76   Bacillales: 0.76        Listeriaceae: 0.76      Listeria: 0.76  Listeria monocytogenes: 0.76


**OUTPUT**

::

  # total number of bins is 2, of which 2 (100.00%) are classified.
  #
  # rank  clade   number of bins
  superkingdom    Bacteria        2
  phylum  Proteobacteria  1
  phylum  Firmicutes      1
  class   Gammaproteobacteria     1
  class   Bacilli 1
  order   Vibrionales     1
  order   Bacillales      1
  family  Vibrionaceae    1
  family  Listeriaceae    1
  genus   Vibrio  1
  genus   Listeria        1
  species Vibrio cholerae 1
  species Listeria monocytogenes  1


Required arguments:
  -i, --input_file     Path to named CAT contig classification file or BAT
                       bin classification file. Currently only official ranks
                       are supported, and only classification files
                       containing a single classification per contig / bin.

Optional arguments:
  -c, --contigs_fasta
                        Path to contigs fasta file. This is required if you
                        want to summarise a contig classification file.

@COMMON_HELP@
    ]]></help>
    <expand macro="citations" />
</tool>