<tool id="hicexplorer_hicvalidatelocations" name="@BINARY@" version="@WRAPPER_VERSION@.0">
    <description>validate detected loops with protein peaks.</description>
    <macros>
        <token name="@BINARY@">hicValidateLocations</token>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <command detect_errors="exit_code"><![CDATA[
   
        @BINARY@
            --data '$data'
            --protein '$protein'
            --method loops
            --resolution '$resolution'
            $addChrPrefixLoops
            $addChrPrefixProtein
            --outFileName outfile
    ]]></command>
    <inputs>
        <param argument="--data" type="data" format='bed'
                        label="Loop file"
                        help="The loop file from hicDetectLoops. To use files from other sources, '
                                'please follow \'chr start end chr start end\' format." />
        <param argument="--protein" type="data" format='bed'
                        label="Loop file"
                        help="The loop file from hicDetectLoops. To use files from other sources, '
                                'please follow \'chr start end chr start end\' format." />
        <param argument="--resolution" type="integer" value=""
                label="Used resolution"
                help='The to be binned resolution.'/>    

        <param name='addChrPrefixLoops' type='boolean' truevalue='--addChrPrefixLoops'  falsevalue="" checked="false" label="Add chr prefix to loops."
                    help="Adding a \'chr\'-prefix to chromosome name of the loops."/>
        
        <param name='addChrPrefixProtein' type='boolean' truevalue='--addChrPrefixLoops'  falsevalue="" checked="false" label="Add chr prefix to protein."
                    help="Adding a \'chr\'-prefix to chromosome name of the protein."/>
   
    </inputs>
    <outputs>
        <data name="statistic" from_work_dir="outfile_statistics" format="bed" label="${tool.name} on [${on_string}]: Statistic"/>
        <data name="mergedLoops" from_work_dir="outfile_matched_locations" format="bed" label="${tool.name} on [${on_string}]: Matched locations"/>
    </outputs>
    <tests>
        <test>
            <param name="data" ftype='bed' value="hicValidateLocations/loops_1.bedgraph"/>
            <param name="protein" ftype='bed' value="hicValidateLocations/GSM935376_hg19_Gm12878_Smc3.narrowPeak"/>
            <param name="resolution" value="10000"/>
            <param name="addChrPrefixLoops" value="True"/>
            <output name="statistic" file="hicValidateLocations/overlap_smc3_statistics" ftype="bed" compare="sim_size" delta='40000'/>
            <output name="mergedLoops" file="hicValidateLocations/overlap_smc3_matched_locations" ftype="bed" compare="sim_size" delta='40000'/>
        </test>
    </tests>
    <help><![CDATA[

Validate locations
==================

This script overlaps the loop locations with protein locations to determine the accuracy of the loop detection.
Loops need to have format as follows:

`chr start end chr start end`

The protein peaks need to be in narrowPeaks or broadPeak format.

A protein match is successfull if at the bin of the x and y location a protein peak is overlapped.
A bin is assumed to have a protein if one or more protein peaks falling within the bin region.
The value of the protein is not considered, only match or non-match.

For more information about HiCExplorer please consider our documentation on readthedocs.io_

.. _readthedocs.io: http://hicexplorer.readthedocs.io/en/latest/index.html
]]></help>
    <expand macro="citations" />
</tool>