<?xml version="1.0"?>
<tool id="spolpred" name="SpolPred" version="1.0.0">
  <description>with options and commands</description>
  <requirements>
    <requirement type="package" version="1.0.0">spolpred</requirement>
  </requirements>
  <command interpreter="bash">

#set $output=$input_file.name

spolpred.sh "$input_file.name" $input_file

-l $read_length -b $type_reads -d $more_details -s $screening_options.stop_screening

#if $screening_options.stop_screening == "on":
  -a $screening_options.screening_threshold
#end if

-m $matching_threshold

  </command>
  <inputs>
    <param name="input_file" type="data" format="fastqsanger" label="FASTQ input file"/>

    <param name="read_length" type="integer" label="Read length [35, 1000]" value="75">
      <validator type="in_range" min="35" max="1000" message="Must be between 35 and 1000 (inclusive)"/>
    </param>

    <param name="type_reads" type="select" label="Type of input reads">
        <option value="d">Direct</option>
        <option value="r">Reverse</option>
    </param>

    <param name="more_details" type="select" label="Level of processing output detail"
      help="If set on, processing details are output to the job's STDOUT, including
      number of processed reads and number of spacer sequences found">
        <option value="on">High</option>
        <option value="off">Normal</option>
    </param>

    <conditional name="screening_options">
      <param name="stop_screening" type="select" label="Read screening"
        help="Used to end read processing when Screening Threshold is reached">
          <option value="on">Perform read screening</option>
          <option value="off">Do not perform read screening</option>
      </param>
      <when value="on">
        <param name="screening_threshold" type="integer" label="Screening threshold" value="50"
          help="Average number of spacer occurrences used to stop screening">
          <validator type="in_range" min="0" max="inf" message="Must be at least 0"/>
        </param>
      </when>
      <when value="off"/>
    </conditional>

    <param name="matching_threshold" type="integer" label="Matching threshold" value="4"
      help="Minimum number of spacer occurrences below which spacer absence is assigned">
      <validator type="in_range" min="1" max="inf" message="Must be at least 1"/>
    </param>

  </inputs>
  <outputs>
    <data name="outfile" format="tabular" from_work_dir="output.txt"/>
  </outputs>

  <tests>
    <test>
      <param/>
      <output/>
    </test>
  </tests>

<help>
  **Frequently Asked Questions**

  **SpolPred only accepts one FASTQ file, what if I have got paired-end reads?**

  Forward and reverse read files can be merged into one by making use of the Perl script
  shuffleSequences_fastq.pl provided in Velvet software suite. SpolPred run will therefore take longer
  than using only forward or reverse reads. In our dataset (read Methods for more details), the forward file
  had enough reads to find all present spacers and infer the octal code for 49 out of 51 samples. That
  decision will have to be made depending on the sample coverage depth.



  **What if I have a FASTA file?**

  SpolPred has been particularly designed to process raw reads and therefore only supports sequence
  files in FASTQ format.



  **What is the point of stopping the read screening?**

  By default, all reads in the FASTQ file will be processed. Nevertheless, we have observed that a point is
  reached when no more reads are needed to infer the octal code, in other words, the number of spacer
  occurrences is high enough and steady to assume that all present spacers have already been found.
  Therefore, stopping the program at this point would save time and computer resources. If low coverage
  is the case, stopping the scanning is not advisable.



  **How do I choose the Screening threshold?**

  If you have decided to scan the whole input file there is no need to set such threshold. The Screening
  threshold is used to let the program know when the screening should stop. Such value will depend on
  read coverage. Running the software and looking at the number of times all spacers are detected will
  provide insight into both the coverage and the most appropriate threshold value.



  **Why is a Matching threshold required? Are spacers not supposed to occur uniquely?**

  The number of times each spacer is found is tracked during the screening and absence assigned when
  such number does not reach a user-defined threshold (4 times by default). This threshold, here called
  Matching threshold, has had to be implemented because for some absent spacers, a few spurious
  matches were found. Those false positives are likely to be related with bad-quality issues, like
  sequencing errors. In our data set, no more than 3 false matches were detected for absent spacers, in
  contrast to 50-150 found per present spacer.



  **Should I be worried then about false positive matches?**

  As long as proper pre-filtering steps are carried out to the raw reads, no important issues are expected
  to come up.



  **Can I change the number of allowed SNPs when querying the spacers?**

  This option has not been implemented. Spacer sequences are conserved and only one SNP has been
  reported to occur at the most.



  **Why are exact matches output as well?**

  The number of read-spacer exact matches, i.e. without allowing SNPs, will enable the easily
  identification of SNPs on spacer sequences. When inferring the octal code, exact matches are not
  employed.

  Wrapper Author: Mark Iskander
</help>
<citations>
  <citation type="doi">10.1093/bioinformatics/bts544</citation>
</citations>
</tool>