comparison trimmomatic.xml @ 11:59054f086eca draft

Version 0.36.6 (output logs and handle fastqillumina and fastqsolexa inputs)

10:dfa082f84068

<tool id="trimmomatic" name="Trimmomatic" version="0.36.5">
  <description>flexible read trimming tool for Illumina NGS data</description>
  <macros>
    <import>trimmomatic_macros.xml</import>
  </macros>
  <requirements>

  #else
    ln -s '$fastq_in' fastq_in.'$fastq_in.extension' &&
  #end if
  java \${_JAVA_OPTIONS:--Xmx8G} -jar \$TRIMMOMATIC_JAR_PATH/trimmomatic.jar
  #if $readtype.single_or_paired in ["pair_of_files","collection"]
    PE -threads \${GALAXY_SLOTS:-6} -phred33
      fastq_r1.'$r1_ext' fastq_r2.'$r2_ext'
      fastq_out_r1_paired.'$r1_ext' fastq_out_r1_unpaired.'$r1_ext'
      fastq_out_r2_paired.'$r2_ext' fastq_out_r2_unpaired.'$r2_ext'
  #else
    SE -threads \${GALAXY_SLOTS:-6} -phred33 fastq_in.'$fastq_in.extension' fastq_out.'$fastq_in.extension'
  #end if
  ## ILLUMINACLIP option
  #if $illuminaclip.do_illuminaclip
    #if $illuminaclip.adapter_type.standard_or_custom == "custom"
      #if $readtype.single_or_paired in ["pair_of_files","collection"]

    #end if
    #if str( $op.operation.name ) == "MAXINFO"
      MAXINFO:$op.operation.target_length:$op.operation.strictness
    #end if
  #end for
  2>&1 | tee trimmomatic.log &&
  if [ -z "\$(tail -1 trimmomatic.log | grep "Completed successfully")" ]; then echo "Trimmomatic did not finish successfully" >&2 ; exit 1 ; fi
  &&
  #if $readtype.single_or_paired  == "pair_of_files"
    mv fastq_out_r1_paired.'$r1_ext' '${fastq_out_r1_paired}' &&

         <option value="se" selected="true">Single-end</option>
         <option value="pair_of_files">Paired-end (two separate input files)</option>
         <option value="collection">Paired-end (as collection)</option>
      </param>
    <when value="se">
      <param name="fastq_in" type="data" format="fastqsanger,fastqsanger.gz" label="Input FASTQ file" />
    </when>
    <when value="pair_of_files">
      <param name="fastq_r1_in" type="data" format="fastqsanger,fastqsanger.gz"
         label="Input FASTQ file (R1/first of pair)" />
      <param name="fastq_r2_in" type="data" format="fastqsanger,fastqsanger.gz"
         label="Input FASTQ file (R2/second of pair)" />
    </when>
      <when value="collection">
        <param name="fastq_pair" format="fastqsanger,fastqsanger.gz" type="data_collection" collection_type="paired" label="Select FASTQ dataset collection with R1/R2 pair" />
      </when>
    </conditional>
    <conditional name="illuminaclip">
      <param name="do_illuminaclip" type="boolean" label="Perform initial ILLUMINACLIP step?" help="Cut adapter and other illumina-specific sequences from the read" truevalue="yes" falsevalue="no" checked="False" />
      <when value="yes">

          <param name="target_length" type="integer" label="Target read length" value="" help="The read length which is likely to allow the location of the read within the target sequence to be determined." />
          <param name="strictness" type="float" label="Strictness" value="" help="Set between zero and one - specifies the balance between preserving read length versus removal of incorrect bases; low values (&lt;0.2) favours longer reads, high values (&gt;0.8) favours read correctness." />
        </when>
      </conditional>
    </repeat>
  </inputs>
  <outputs>
    <data name="fastq_out_r1_paired" label="${tool.name} on ${readtype.fastq_r1_in.name} (R1 paired)" format_source="fastq_r1_in">
      <filter>readtype['single_or_paired'] == "pair_of_files"</filter>
    </data>

      <collection name="fastq_out_unpaired" type="paired" label="${tool.name} on ${on_string}: unpaired">
        <filter>readtype['single_or_paired'] == "collection"</filter>
        <data name="forward" label="${tool.name} on ${readtype.fastq_pair.forward.name} (R1 unpaired)" format_source="fastq_pair['forward']"/>
        <data name="reverse" label="${tool.name} on ${readtype.fastq_pair.reverse.name} (R2 unpaired)" format_source="fastq_pair['reverse']"/>
    </collection>

  </outputs>
  <tests>
    <test>
      <!-- Single-end example -->
      <param name="single_or_paired" value="se" />
      <param name="fastq_in" value="Illumina_SG_R1.fastq" ftype="fastqsanger" />
      <param name="operations_0|operation|name" value="SLIDINGWINDOW" />
      <output name="fastq_out" file="trimmomatic_se_out1.fastq" />
    </test>
    <test>
      <!-- Single-end example - gzipped -->
      <param name="single_or_paired" value="se" />
      <param name="fastq_in" value="Illumina_SG_R1.fastq.gz" ftype="fastqsanger.gz" />

      <param name="operations_0|operation|name" value="SLIDINGWINDOW" />
      <output name="fastq_out_r1_paired" file="trimmomatic_pe_r1_paired_out1.fastq" />
      <output name="fastq_out_r1_unpaired" file="trimmomatic_pe_r1_unpaired_out1.fastq" />
      <output name="fastq_out_r2_paired" file="trimmomatic_pe_r2_paired_out1.fastq" />
      <output name="fastq_out_r2_unpaired" file="trimmomatic_pe_r2_unpaired_out1.fastq" />
    </test>
    <test>
      <!-- Single-end example (cropping) -->
      <param name="single_or_paired" value="se" />
      <param name="fastq_in" value="Illumina_SG_R1.fastq" ftype="fastqsanger" />

**Credits**

This Galaxy tool has been developed within the Bioinformatics Core Facility at the
University of Manchester, with contributions from Peter van Heusden, Marius
van den Beek, Jelle Scholtalbers and Charles Girardot.

It runs the Trimmomatic program which has been developed
within Bjorn Usadel's group at RWTH Aachen university.

Trimmomatic website (including documentation):

11:59054f086eca

<tool id="trimmomatic" name="Trimmomatic" version="0.36.6">
  <description>flexible read trimming tool for Illumina NGS data</description>
  <macros>
    <import>trimmomatic_macros.xml</import>
  </macros>
  <requirements>

  #else
    ln -s '$fastq_in' fastq_in.'$fastq_in.extension' &&
  #end if
  java \${_JAVA_OPTIONS:--Xmx8G} -jar \$TRIMMOMATIC_JAR_PATH/trimmomatic.jar
  #if $readtype.single_or_paired in ["pair_of_files","collection"]
    PE -threads \${GALAXY_SLOTS:-6}
      fastq_r1.'$r1_ext' fastq_r2.'$r2_ext'
      fastq_out_r1_paired.'$r1_ext' fastq_out_r1_unpaired.'$r1_ext'
      fastq_out_r2_paired.'$r2_ext' fastq_out_r2_unpaired.'$r2_ext'
  #else
    SE -threads \${GALAXY_SLOTS:-6} fastq_in.'$fastq_in.extension' fastq_out.'$fastq_in.extension'
  #end if
  ## ILLUMINACLIP option
  #if $illuminaclip.do_illuminaclip
    #if $illuminaclip.adapter_type.standard_or_custom == "custom"
      #if $readtype.single_or_paired in ["pair_of_files","collection"]

    #end if
    #if str( $op.operation.name ) == "MAXINFO"
      MAXINFO:$op.operation.target_length:$op.operation.strictness
    #end if
  #end for
  #if $output_logs:
    -trimlog trimlog
  #end if
  2>&1 | tee trimmomatic.log &&
  if [ -z "\$(tail -1 trimmomatic.log | grep "Completed successfully")" ]; then echo "Trimmomatic did not finish successfully" >&2 ; exit 1 ; fi
  &&
  #if $readtype.single_or_paired  == "pair_of_files"
    mv fastq_out_r1_paired.'$r1_ext' '${fastq_out_r1_paired}' &&

         <option value="se" selected="true">Single-end</option>
         <option value="pair_of_files">Paired-end (two separate input files)</option>
         <option value="collection">Paired-end (as collection)</option>
      </param>
    <when value="se">
      <param name="fastq_in" type="data" format="fastqsanger,fastqsanger.gz,fastqillumina,fastqillumina.gz,fastqsolexa,fastqsolexa.gz" label="Input FASTQ file" />
    </when>
    <when value="pair_of_files">
      <param name="fastq_r1_in" type="data" format="fastqsanger,fastqsanger.gz,fastqillumina,fastqillumina.gz,fastqsolexa,fastqsolexa.gz"
         label="Input FASTQ file (R1/first of pair)" />
      <param name="fastq_r2_in" type="data" format="fastqsanger,fastqsanger.gz,fastqillumina,fastqillumina.gz,fastqsolexa,fastqsolexa.gz"
         label="Input FASTQ file (R2/second of pair)" />
    </when>
      <when value="collection">
        <param name="fastq_pair" format="fastqsanger,fastqsanger.gz,fastqillumina,fastqillumina.gz,fastqsolexa,fastqsolexa.gz" type="data_collection" collection_type="paired" label="Select FASTQ dataset collection with R1/R2 pair" />
      </when>
    </conditional>
    <conditional name="illuminaclip">
      <param name="do_illuminaclip" type="boolean" label="Perform initial ILLUMINACLIP step?" help="Cut adapter and other illumina-specific sequences from the read" truevalue="yes" falsevalue="no" checked="False" />
      <when value="yes">

          <param name="target_length" type="integer" label="Target read length" value="" help="The read length which is likely to allow the location of the read within the target sequence to be determined." />
          <param name="strictness" type="float" label="Strictness" value="" help="Set between zero and one - specifies the balance between preserving read length versus removal of incorrect bases; low values (&lt;0.2) favours longer reads, high values (&gt;0.8) favours read correctness." />
        </when>
      </conditional>
    </repeat>
    <param name="output_logs" argument="-trimlog" type="boolean" label="Output trimlog file?" truevalue="yes" falsevalue="no" checked="False" />
    <param name="output_err" type="boolean" label="Output trimmomatic log messages?" truevalue="yes" falsevalue="no" checked="False" help="these are the messages written to stderr (eg. for use in MultiQC)" />
  </inputs>
  <outputs>
    <data name="fastq_out_r1_paired" label="${tool.name} on ${readtype.fastq_r1_in.name} (R1 paired)" format_source="fastq_r1_in">
      <filter>readtype['single_or_paired'] == "pair_of_files"</filter>
    </data>

      <collection name="fastq_out_unpaired" type="paired" label="${tool.name} on ${on_string}: unpaired">
        <filter>readtype['single_or_paired'] == "collection"</filter>
        <data name="forward" label="${tool.name} on ${readtype.fastq_pair.forward.name} (R1 unpaired)" format_source="fastq_pair['forward']"/>
        <data name="reverse" label="${tool.name} on ${readtype.fastq_pair.reverse.name} (R2 unpaired)" format_source="fastq_pair['reverse']"/>
    </collection>
    <data name="log_file" format="txt" label="${tool.name} on ${on_string} (trimlog file)" from_work_dir="trimlog">
      <filter>output_logs</filter>
    </data>
    <data name="err_file" format="txt" label="${tool.name} on ${on_string} (log file)" from_work_dir="trimmomatic.log">
      <filter>output_err</filter>
    </data>
  </outputs>
  <tests>
    <test>
      <!-- Single-end example -->
      <conditional name="readtype">
        <param name="single_or_paired" value="se" />
        <param name="fastq_in" value="Illumina_SG_R1.fastq" ftype="fastqsanger" />
      </conditional>
      <param name="operations_0|operation|name" value="SLIDINGWINDOW" />
      <param name="output_logs" value="yes" />
      <param name="output_err" value="yes" />
      <output name="fastq_out" file="trimmomatic_se_out1.fastq" />
      <output name="log_file" file="trimmomatic_se_out1.log" />
      <output name="err_file" file="trimmomatic_se_out1.err" />
    </test>
    <test>
      <!-- Single-end example - gzipped -->
      <param name="single_or_paired" value="se" />
      <param name="fastq_in" value="Illumina_SG_R1.fastq.gz" ftype="fastqsanger.gz" />

      <param name="operations_0|operation|name" value="SLIDINGWINDOW" />
      <output name="fastq_out_r1_paired" file="trimmomatic_pe_r1_paired_out1.fastq" />
      <output name="fastq_out_r1_unpaired" file="trimmomatic_pe_r1_unpaired_out1.fastq" />
      <output name="fastq_out_r2_paired" file="trimmomatic_pe_r2_paired_out1.fastq" />
      <output name="fastq_out_r2_unpaired" file="trimmomatic_pe_r2_unpaired_out1.fastq" />
    </test>
    <test>
      <!-- Paired-end Illumina 1.3-1.7 quality encoding -->
      <param name="single_or_paired" value="pair_of_files" />
      <param name="fastq_r1_in" value="Illumina_SG_R1.fastqillumina" ftype="fastqillumina" />
      <param name="fastq_r2_in" value="Illumina_SG_R2.fastqillumina" ftype="fastqillumina" />
      <param name="operations_0|operation|name" value="SLIDINGWINDOW" />
      <output name="fastq_out_r1_paired" file="trimmomatic_pe_r1_paired_out1.fastqillumina" />
      <output name="fastq_out_r1_unpaired" file="trimmomatic_pe_r1_unpaired_out1.fastqillumina" />
      <output name="fastq_out_r2_paired" file="trimmomatic_pe_r2_paired_out1.fastqillumina" />
      <output name="fastq_out_r2_unpaired" file="trimmomatic_pe_r2_unpaired_out1.fastqillumina" />
    </test>
    <test>
      <!-- Paired-end Solexa quality encoding -->
      <param name="single_or_paired" value="pair_of_files" />
      <param name="fastq_r1_in" value="Illumina_SG_R1.fastqsolexa" ftype="fastqsolexa" />
      <param name="fastq_r2_in" value="Illumina_SG_R2.fastqsolexa" ftype="fastqsolexa" />
      <param name="operations_0|operation|name" value="SLIDINGWINDOW" />
      <output name="fastq_out_r1_paired" file="trimmomatic_pe_r1_paired_out1.fastqsolexa" />
      <output name="fastq_out_r1_unpaired" file="trimmomatic_pe_r1_unpaired_out1.fastqsolexa" />
      <output name="fastq_out_r2_paired" file="trimmomatic_pe_r2_paired_out1.fastqsolexa" />
      <output name="fastq_out_r2_unpaired" file="trimmomatic_pe_r2_unpaired_out1.fastqsolexa" />
    </test>
    <test>
      <!-- Single-end example (cropping) -->
      <param name="single_or_paired" value="se" />
      <param name="fastq_in" value="Illumina_SG_R1.fastq" ftype="fastqsanger" />

**Credits**

This Galaxy tool has been developed within the Bioinformatics Core Facility at the
University of Manchester, with contributions from Peter van Heusden, Marius
van den Beek, Jelle Scholtalbers, Charles Girardot, and Matthias Bernt.

It runs the Trimmomatic program which has been developed
within Bjorn Usadel's group at RWTH Aachen university.

Trimmomatic website (including documentation):