comparison Mutect2.xml @ 2:646e6943bcd2 draft default tip

planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/main/tools/gatk4 commit 9c42369510cce59bf0dcb0edb440322d17b18339

1:b040adcfeefd

<tool id="mutect2" name="gatk4 Mutect2" version="@WRAPPER_VERSION@" profile="18.05">
    <description>- Call somatic SNVs and indels via local assembly of haplotypes</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements"/>

            gatk CreateSequenceDictionary --REFERENCE="reference.fa" --OUTPUT="reference.dict" &&
        #else
            #set ref_flag=''
        #end if

        #if str($mode.mode_parameters) == 'tumor_only'
            ln -s $mode.tumor tumor.bam &&
            ln -s $mode.tumor.metadata.bam_index tumor.bam.bai &&
        #else
            ln -s $mode.tumor tumor.bam &&
            ln -s $mode.tumor.metadata.bam_index tumor.bam.bai &&
            ln -s $mode.normal normal.bam &&
            ln -s $mode.normal.metadata.bam_index normal.bam.bai &&
        #end if

        #if str($outputs.output_parameters) == 'yes'

        #include source=$vcf_output_opts#
        #include source=$gatk_seqdict#
        ]]>
    </command>
    <inputs>
        <conditional name="mode">
            <param name="mode_parameters" type="select" label="Type of analysis">
                <option value="tumor_only">Tumor-only</option>
                <option value="somatic">Somatic</option>
            </param>
            <when value="tumor_only">
                <param name="tumor" argument="--tumor" type="data" format="bam" label="Input Tumor BAM file" />
            </when>
            <when value="somatic">
                <param name="tumor" argument="--tumor" type="data" format="bam" label="Input Tumor BAM file" />
                <param name="normal" argument="--normal" type="data" format="bam" label="Input Normal BAM file" />
            </when>
        </conditional>
        <expand macro="gzip_vcf_params"/>
        <expand macro="ref_sel"/>

                <param name="read_validation_stringency" argument="--read-validation-stringency" type="select" optional="true" label="Read Validation Stringency" help="Validation stringency for all SAM/BAM/CRAM/SRA files read by this program.  The default stringency value SILENT can improve performance when processing a BAM file in which variable-length data (read, qualities, tags) do not otherwise need to be decoded.">
                    <option selected="true" value="SILENT">Silent</option>
                    <option value="STRICT">Strict</option>
                    <option value="LENIENT">Lenient</option>
                </param>
                <param name="sites_only_vcf_output" argument="--sites-only-vcf-output" type="boolean" truevalue="--sites-only-vcf-output" falsevalue="" optional="true" checked="false" label="Sites Only Vcf Output" help="If true, don&amp;apos;t emit genotype fields when writing vcf file output."/>
                <param name="tumor_lod_to_emit" argument="--tumor-lod-to-emit" type="float" optional="true" value="3.0" label="Tumor Lod To Emit" help="LOD threshold to emit tumor variant to VCF."/>
            </when>
            <when value="no" />
        </conditional>
        <conditional name="advanced">

        <data format="bam" name="bam_output" label="${tool.name} on ${on_string}: Debug BAM output">
            <filter>str(outputs['output_parameters']) == 'yes' and outputs['debug_bam']</filter>
        </data>
    </outputs>
    <tests>
        <test>
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
                <param name="tumor" ftype="bam" value="Mutect2-in1.bam" />
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="optional_parameters" value="no" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out1.vcf" lines_diff="2" />
            <output name="output_vcf_stats" file="Mutect2-out1.vcf.stats" />
        </test>
        <test>
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
                <param name="tumor" ftype="bam" value="Mutect2-in2.bam" />
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="read_filter" value="AmbiguousBaseReadFilter,FirstOfPairReadFilter,GoodCigarReadFilter" />
            <param name="seqdict_source" value="history" />
            <param name="seqdict_sequence" value="Mutect2-in2.dict" />
            <param name="optional_parameters" value="no" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out2.vcf" lines_diff="2" />
        </test>
        <test>
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
                <param name="tumor" ftype="bam" value="Mutect2-in3.bam" />
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="optional_parameters" value="yes" />

            <param name="annotation_group" value="StandardMutectAnnotation" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out3.vcf" lines_diff="4" />
        </test>
        <test>
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
                <param name="tumor" ftype="bam" value="Mutect2-in4.bam" />
            </conditional>
            <param name="reference_sequence" value="hg38"/>
            <param name="gzipped_output" value="false" />
	    <param name="reference_source_selector" value="cached" />
	    <param name="reference_sequence" value="hg38"/>
            <param name="optional_parameters" value="no" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out4.vcf" lines_diff="2" />
        </test>
        <test>
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
                <param name="tumor" ftype="bam" value="Mutect2-in5.bam" />
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="optional_parameters" value="no" />

            <param name="debug_bam" value="true" />
            <output name="output_vcf" file="Mutect2-out5.vcf" lines_diff="2" />
            <output name="assembly_region_out" file="Mutect2-out5-1.tabular" />
            <output name="bam_output" file="Mutect2-out5.bam" />
        </test>
        <test>
            <conditional name="mode">
                <param name="mode_parameters" value="somatic"/>
                <param name="tumor" ftype="bam" value="tumor.bam" />
                <param name="normal" ftype="bam" value="normal.bam" />
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="chr20.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />

        </test>
    </tests>
    <help><![CDATA[Call somatic short variants via local assembly of haplotypes. Short
variants include single nucleotide (SNV) and insertion and deletion
(indel) variants. The caller combines the DREAM challenge-winning
somatic genotyping engine of the original MuTect (`Cibulskis et al.,
2013 <http://www.nature.com/nbt/journal/v31/n3/full/nbt.2514.html>`__)
with the assembly-based machinery of
`HaplotypeCaller <https://gatk.broadinstitute.org/hc/en-us/articles/360035531412-HaplotypeCaller-in-a-nutshell>`__.

This tool is featured in the *Somatic Short Mutation calling Best
Practice Workflow*. See `this article <https://gatk.broadinstitute.org/hc/en-us/articles/360035531132>`__
for an overview of what traditional somatic calling entails, with usage examples. For the
latest pipeline scripts, see the `Mutect2 WDL scripts
directory <https://github.com/broadinstitute/gatk/tree/master/scripts/mutect2_wdl>`__.
Although we present the tool for somatic calling, it may apply to other
contexts, such as mitochondrial variant calling.

Usage examples
~~~~~~~~~~~~~~

2:646e6943bcd2

<tool id="mutect2" name="gatk4 Mutect2" version="@WRAPPER_VERSION@" profile="21.01">
    <description>- Call somatic SNVs and indels via local assembly of haplotypes</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements"/>

            gatk CreateSequenceDictionary --REFERENCE="reference.fa" --OUTPUT="reference.dict" &&
        #else
            #set ref_flag=''
        #end if

        ln -s $tumor tumor.bam &&
        ln -s $tumor.metadata.bam_index tumor.bam.bai &&

        #if str($mode.mode_parameters) == 'somatic'
            ln -s $mode.normal normal.bam &&
            ln -s $mode.normal.metadata.bam_index normal.bam.bai &&
        #end if

        #if str($outputs.output_parameters) == 'yes'

        #include source=$vcf_output_opts#
        #include source=$gatk_seqdict#
        ]]>
    </command>
    <inputs>
        <param name="tumor" argument="--tumor" type="data" format="bam" label="Input Tumor BAM file" />
        <conditional name="mode">
            <param name="mode_parameters" type="select" label="Type of analysis">
                <option value="tumor_only">Tumor-only</option>
                <option value="somatic">Somatic</option>
            </param>
            <when value="tumor_only">
            </when>
            <when value="somatic">
                <param name="normal" argument="--normal" type="data" format="bam" label="Input Normal BAM file" />
            </when>
        </conditional>
        <expand macro="gzip_vcf_params"/>
        <expand macro="ref_sel"/>

                <param name="read_validation_stringency" argument="--read-validation-stringency" type="select" optional="true" label="Read Validation Stringency" help="Validation stringency for all SAM/BAM/CRAM/SRA files read by this program.  The default stringency value SILENT can improve performance when processing a BAM file in which variable-length data (read, qualities, tags) do not otherwise need to be decoded.">
                    <option selected="true" value="SILENT">Silent</option>
                    <option value="STRICT">Strict</option>
                    <option value="LENIENT">Lenient</option>
                </param>
                <param name="tumor_lod_to_emit" argument="--tumor-lod-to-emit" type="float" optional="true" value="3.0" label="Tumor Lod To Emit" help="LOD threshold to emit tumor variant to VCF."/>
            </when>
            <when value="no" />
        </conditional>
        <conditional name="advanced">

        <data format="bam" name="bam_output" label="${tool.name} on ${on_string}: Debug BAM output">
            <filter>str(outputs['output_parameters']) == 'yes' and outputs['debug_bam']</filter>
        </data>
    </outputs>
    <tests>
        <test expect_num_outputs="2">
            <param name="tumor" ftype="bam" value="Mutect2-in1.bam" />
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="optional_parameters" value="no" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out1.vcf" lines_diff="2" />
            <output name="output_vcf_stats" file="Mutect2-out1.vcf.stats" />
        </test>
        <test expect_num_outputs="2">
            <param name="tumor" ftype="bam" value="Mutect2-in2.bam" />
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="read_filter" value="AmbiguousBaseReadFilter,FirstOfPairReadFilter,GoodCigarReadFilter" />
            <param name="seqdict_sequence" value="Mutect2-in2.dict" />
            <param name="optional_parameters" value="no" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out2.vcf" lines_diff="2" />
        </test>
        <test expect_num_outputs="2">
            <param name="tumor" ftype="bam" value="Mutect2-in3.bam" />
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="optional_parameters" value="yes" />

            <param name="annotation_group" value="StandardMutectAnnotation" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out3.vcf" lines_diff="4" />
        </test>
        <test expect_num_outputs="2">
            <param name="tumor" ftype="bam" value="Mutect2-in4.bam" />
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
            </conditional>
            <param name="reference_sequence" value="hg38"/>
            <param name="gzipped_output" value="false" />
	    <param name="reference_source_selector" value="cached" />
	    <param name="reference_sequence" value="hg38"/>
            <param name="optional_parameters" value="no" />
            <param name="advanced_parameters" value="no" />
            <param name="output_parameters" value="no" />
            <output name="output_vcf" file="Mutect2-out4.vcf" lines_diff="2" />
        </test>
        <test expect_num_outputs="4">
            <param name="tumor" ftype="bam" value="Mutect2-in5.bam" />
            <conditional name="mode">
                <param name="mode_parameters" value="tumor_only"/>
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="reference.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />
            <param name="optional_parameters" value="no" />

            <param name="debug_bam" value="true" />
            <output name="output_vcf" file="Mutect2-out5.vcf" lines_diff="2" />
            <output name="assembly_region_out" file="Mutect2-out5-1.tabular" />
            <output name="bam_output" file="Mutect2-out5.bam" />
        </test>
        <test expect_num_outputs="2">
            <param name="tumor" ftype="bam" value="tumor.bam" />
            <conditional name="mode">
                <param name="mode_parameters" value="somatic"/>
                <param name="normal" ftype="bam" value="normal.bam" />
            </conditional>
            <param name="reference_sequence" ftype="fasta" value="chr20.fa" />
            <param name="gzipped_output" value="false" />
            <param name="reference_source_selector" value="history" />

        </test>
    </tests>
    <help><![CDATA[Call somatic short variants via local assembly of haplotypes. Short
variants include single nucleotide (SNV) and insertion and deletion
(indel) variants. The caller combines the DREAM challenge-winning
somatic genotyping engine of the original MuTect_ (Cibulskis et al. 2013)
with the assembly-based machinery of HaplotypeCaller (see
gatk.broadinstitute.org/hc/en-us/articles/360035531412-HaplotypeCaller-in-a-nutshell).

.. _MuTect: http://www.nature.com/nbt/journal/v31/n3/full/nbt.2514.html

This tool is featured in the *Somatic Short Mutation calling Best
Practice Workflow*. See gatk.broadinstitute.org/hc/en-us/articles/360035531132 
for an overview of what traditional somatic calling entails, with usage examples. For the
latest pipeline scripts, see the Mutect2 WDL scripts directory (see
github.com/broadinstitute/gatk/tree/master/scripts/mutect2_wdl).
Although we present the tool for somatic calling, it may apply to other
contexts, such as mitochondrial variant calling.

Usage examples
~~~~~~~~~~~~~~