comparison dsCompareCurves.xml @ 7:ef6b589a3715 draft default tip

More documentation to match the Wiki

7:ef6b589a3715

<tool id="deepstats_compare_curves" name="dsCompareCurves" version="@WRAPPER_VERSION@.0" profile="18.01">
    <description>
        compares multiple genomic scores at multiple regions sets by bootstraps and per-bin distribution test
    </description>
    <macros>
        <token name="@BINARY@">dsCompareCurves</token>
        <import>deepStats_macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <command>
     <![CDATA[
      @BINARY@
        --input '$input'
        --output 'output'
        --comparison '$comparison'
        --signalName '$signalName'
        --scoreLabels '$nameLabels.scoreLabels'
        --regionLabels '$nameLabels.regionLabels'

        #if $scaleRegions.showScaleRegionsOpt == "yes":
          --firstRegionName '$scaleRegions.firstRegionName'
          --secondRegionName '$scaleRegions.secondRegionName'
        #else
          --firstRegionName '$scaleRegions.firstRegionName'
        #end if

        --bootstraps '$statsOptions.bootstraps'
        --bootstrapsCI '$statsOptions.bootstrapsCI'
        --wilcoxThreshold '$statsOptions.wilcoxThreshold'
        --bootPlotShareY '$plotOptions.bootPlotShareY'
        --bootPlotRatio '$plotOptions.bootPlotRatio'
        --bootPlotWidth '$plotOptions.bootPlotWidth'
        --bootPlotHeight '$plotOptions.bootPlotHeight'
        --wilcoxPlotWidth '$plotOptions.wilcoxPlotWidth'
        --wilcoxPlotHeight '$plotOptions.wilcoxPlotHeight'
        --CPU '$advancedOptions.CPU'

        #if $fontOptions.showFontOptions == "yes":
          --font '$fontOptions.font_file'
        #end if

        #if $colorsOptions.showColorsOptions == "yes":
          --bootPlotColors '$colorsOptions.bootPlotColors'
        #end if
    ]]>
    </command>
    <inputs>
      <param argument="--input" format="tabular" type="data"
        label="deepTools computeMatrix output file."/>

      <param argument="--comparison" type="select"
        label="What do you want to compare?">
        <option value="scores" selected="true">scores</option>
        <option value="regions">regions</option>
      </param>

      <param argument="--signalName" type="text" value= "Normalized ATAC-seq counts"
        label="Name of the genomic signal" help="Name given to the signal on the plots, \n can be added for linebreaks."/>

      <conditional name="scaleRegions">
          <param name="showScaleRegionsOpt" type="select" label="Which mode was used in deepTools computeMatrix?">
            <option value="reference-point" selected="true">reference-point</option>
            <option value="scale-regions">scale-regions</option>
          </param>
          <when value="scale-regions">
            <param argument="--firstRegionName" type="text" value="TSS" size="10"
                      label="Label for the region start"
                      help ="Label shown in the plot
                      for the start of the region. Default is TSS (transcription start site),
                      but could be changed to anything, e.g. &quot;peak start&quot;." />
            <param argument="--secondRegionName" type="text" value="TES" size="10"
                      label="Label for the region end"
                      help ="Label shown in the plot
                      for the end of the region. Default is TES (transcription end site),
                      but could be changed to anything, e.g. &quot;peak end&quot;." />
          </when>
          <when value="reference-point">
            <param argument="--firstRegionName" type="text" value="TSS"
                      label="Label for the region start"
                      help ="Label shown in the plot
                      for the start of the region. Default is TSS (transcription start site),
                      but could be changed to anything, e.g. &quot;peak start&quot;." />
          </when>
      </conditional>

      <section name="nameLabels" title="Customize score and region labels" expanded="false">
          <param argument="--scoreLabels" optional="true" type="text" value= ""
          label="Name of the scores" help="For plot display. It must be provided as text seperated
          by semi-colons, i.e. 'Score A;Score B;Score C' as in the example.">
            <sanitizer>
              <valid initial="string.printable"/>
            </sanitizer>
          </param>

          <param argument="--regionLabels" optional="true" type="text" value= ""
          label="Name of the regions" help="For plot display. It must be provided as text seperated
          by semi-colons, i.e. 'Regions A;Regions B;Regions C' as in the example.">
          <sanitizer>
            <valid initial="string.printable"/>
          </sanitizer>
        </param>
      </section>

      <section name="statsOptions" title="Statistics options" expanded="false">
          <param argument="--bootstraps" type="integer" value="1000"
          label="--bootstraps" help="Number of bootstraps to perform. Default is 1000,
          increasing this number will increase the computing time."/>
          <param argument="--bootstrapsCI" type="float" value="0.95"
          label="--bootstrapsCI" help="Bootstraps confidence interval. Default is 0.95,
          increasing this will make the analysis more stringent by increasing            the min and max of the shadowed area on the bootstrap plot."/>
          <param argument="--wilcoxThreshold" type="float" value="0.05"
          label="--wilcoxThreshold" help="Threshold used to define significant
          bins on the Wilcoxon rank-sum test plot (grey box). Default is 0.05"/>
      </section>

      <section name="plotOptions" title="Plotting options" expanded="false">
        <param argument="--bootPlotShareY" type="select" label="--bootPlotShareY">
          <option value="TRUE" selected="true">TRUE</option>
          <option value="FALSE">FALSE</option>
        </param>
        <param argument="--bootPlotRatio" type="float" value="0.85" label="--bootPlotRatio"
        help="Changes the aspect ratio of the plot. A value inferior to 1 results in a wide plot,
        a value above 1 results in a narrow plot. Default: 0.85."/>

        <param argument="--bootPlotWidth" type="float" value="5.2" label="--bootPlotWidth"
        help="How large the bootstraps plot should be."/>

        <param argument="--bootPlotHeight" type="float" value="3.7" label="--bootPlotHeight"
        help="How tall the bootstraps plot should be."/>

        <param argument="--wilcoxPlotWidth" type="float" value="4.6" label="--wilcoxPlotWidth"
        help="How large the Wilcoxon rank-sum test plot should be."/>

        <param argument="--wilcoxPlotHeight" type="float" value="4.6" label="--wilcoxPlotHeight"
        help="How tall the Wilcoxon rank-sum test  plot should be."/>
      </section>

      <section name="advancedOptions" title="Computing options" expanded="false">
        <param argument="--CPU" type="integer" value="4" label="--CPU"
        help="Number of CPU to use."/>
      </section>

      <conditional name="fontOptions">
        <param name="showFontOptions" type="select" label="Customize font?">
          <option value="No" selected="true">No</option>
          <option value="Yes">Yes</option>
        </param>
        <when value="No"/>
        <when value="Yes">
          <param argument="--font" type="file" format="ttf" value=" " label="--font" optional="true"
          help="Font used for plotting, given a TTF file. Default is usually Helvetica."/>
        </when>
      </conditional>

      <conditional name="colorsOptions">
        <param name="showColorsOptions" type="select" label="Customize colors?">
          <option value="No" selected="true">No</option>
          <option value="Yes">Yes</option>
        </param>
        <when value="No"/>
        <when value="Yes">
          <param argument="--bootPlotColors" type="file" format="tabular" value=" " label="--bootPlotColors"
          help="Change the bootstraps plot color palette to a user-provided one. The file must
          be tab-delimited and contain for each line two HTML color codes ( #3366CC   #769EF2 ).
          The first column corresponds to the mean color, the second column corresponds to the
          color of the bootstrap confidence interval shadowed area. The default color scale
          contains 6 colors that are color blind friendly using the dichromat R package."/>
        </when>
      </conditional>

    </inputs>
    <outputs>
      <data format="pdf" name="pdf1" label="${tool.name} on ${on_string}: Bootstrap Plot" from_work_dir="output_bootstraps.pdf"/>
      <data format="pdf" name="pdf2" label="${tool.name} on ${on_string}: Wilcox Plot" from_work_dir="output_wilcoxon.pdf"/>
    </outputs>
    <tests>
      <test>
      </test>
    </tests>
    <help><![CDATA[

.. role:: raw-html(raw)
   :format: html

.. image:: https://raw.githubusercontent.com/gtrichard/deepStats/master/docs/images/deepStats_logo.png
   :width: 400

:raw-html:`<br />`

dsCompareCurves: Comparing multiple genomic scores at multiple regions sets
===========================================================================

**deepStats** aims at providing statistical analyses and streamlining the production of high quality, color-blind friendly, and fully customisable plots (up to the fonts!) for your classic genomic datasets (.bed, .bigwig, gene lists). The goal of deepStats is thus to significantly decrease the amount of time spent in Inkscape/Illustrator to get publication ready plots, and decreasing the research time alloted to finding proper statistical analyses for your genomic signals and datasets.

``dsCompareCurves`` assesses if multiple genomics signals (ChIP-seq, ATAC-seq, etc as .bigwig files) are significantly different or not for multiple region sets (.bed files), and specifically on which bins. ``dsCompareCurves`` uses two methods to do so: bootstraps and per-bin Bonferroni corrected Wilcoxon Rank-sum tests. The input for ``dsCompareCurves`` comes from deepTools ``computeMatrix --outFileNameMatrix``, that thus must be used prior to ``dsCompareCurves``. 

In a nutshell, ``dsCompareCurves`` is replacing deepTools ``plotProfile``, but with customisable, scalable and publication ready plots, in addition to providing statistical analyses.

``dsCompareCurves`` compares regions for each signal (``--comparison regions``), or compares all signals for each region set (``--comparison scores``).

Concerning the plots and as the other deepStats tools, a lot of parameters can be customized: plot colors (although the basic color scale is color-blind friendly), fonts, plot size and size ratio.


_______________


Example usage
-------------

Below you can find an example of regions comparison.

We consider two set of regions, conserved genes, and unconserved genes in Drosophila melanogaster. We want to know which histone marks from modENCODE are differentially enriched between these two regions sets.

Prior to ``dsCompareCurves``, we used **deepTools** ``computeMatrix scale-regions`` with the two bed files as regions, and the three histone marks bigwig tracks as scores. We also added the optional parameter ``outFileNameMatrix`` in order to get the output as a tab-delimited file (.txt).

For more information about deepTools computeMatrix: https://deeptools.readthedocs.io/en/develop/content/tools/computeMatrix.html

The latter file `deeptools_computeMatrix_output.txt ` is then used as input for the present tool in "regions" mode, in order to compare the conserved and unconserved genes for each one of the three histone marks scores.

You can find an image describing the outputs below using the said data set and the parameters used for the same analysis on Galaxy:

.. class:: bold

`Example parameters`_

.. _Example parameters: https://raw.githubusercontent.com/gtrichard/deepStats/master/docs/images/dsCompareCurves_galaxy.png

:raw-html:`<br />`

Output description
++++++++++++++++++

.. image:: https://raw.githubusercontent.com/gtrichard/deepStats/master/docs/images/dsCompareCurves_doc.png
   :width: 800

      ]]></help>
    <expand macro="citations" />
</tool>