root/galaxy-central/tools/fastx_toolkit/fastq_quality_filter.xml @ 3

<tool id="cshl_fastq_quality_filter" name="Filter by quality">
        <description></description>
        <requirements><requirement type="package">fastx_toolkit</requirement></requirements>

        <command>zcat -f '$input' | fastq_quality_filter -q $quality -p $percent -v -o $output</command>

        <inputs>
                <param format="fastqsolexa,fastqsanger" name="input" type="data" label="Library to filter" />

                <param name="quality" size="4" type="integer" value="20">
                        <label>Quality cut-off value</label>
                </param>

                <param name="percent" size="4" type="integer" value="90">
                        <label>Percent of bases in sequence that must have quality equal to / higher than cut-off value</label>
                </param>
        </inputs>

        <tests>
                <test>
                        <!-- Test1:  100% of bases with quality 33 or higher (pretty steep requirement...) -->
                        <param name="input" value="fastq_qual_filter1.fastq" />
                        <param name="quality" value="33"/>
                        <param name="percent" value="100"/>
                        <output name="output" file="fastq_qual_filter1a.out" />
                </test>
                <test>
                        <!-- Test2:  80% of bases with quality 20 or higher -->
                        <param name="input" value="fastq_qual_filter1.fastq" />
                        <param name="quality" value="20"/>
                        <param name="percent" value="80"/>
                        <output name="output" file="fastq_qual_filter1b.out" />
                </test>
        </tests>

        <outputs>
                <data format="input" name="output" metadata_source="input" />
        </outputs>

        <help>
**What it does**

This tool filters reads based on quality scores.

.. class:: infomark

Using **percent = 100** requires all cycles of all reads to be at least the quality cut-off value.

.. class:: infomark

Using **percent = 50** requires the median quality of the cycles (in each read) to be at least the quality cut-off value.

--------

Quality score distribution (of all cycles) is calculated for each read. If it is lower than the quality cut-off value - the read is discarded.


**Example**::

    @CSHL_4_FC042AGOOII:1:2:214:584
    GACAATAAAC
    +CSHL_4_FC042AGOOII:1:2:214:584
    30 30 30 30 30 30 30 30 20 10

Using **percent = 50** and **cut-off = 30** - This read will not be discarded (the median quality is higher than 30).

Using **percent = 90** and **cut-off = 30** - This read will be discarded (90% of the cycles do no have quality equal to / higher than 30).

Using **percent = 100** and **cut-off = 20** - This read will be discarded (not all cycles have quality equal to / higher than 20).

------

This tool is based on `FASTX-toolkit`__ by Assaf Gordon.

 .. __: http://hannonlab.cshl.edu/fastx_toolkit/            
        </help>
</tool>
<!-- FASTQ-Quality-Filter is part of the FASTX-toolkit, by A.Gordon (gordon@cshl.edu) -->