view bed_to_bigbed.xml @ 0:2c58335f1ec8 draft default tip

planemo upload

<?xml version="1.0"?>
<tool id="bed_to_bigbed" name="bedToBigBed" version="1.0">
    <description>Converts a BED file to a bigBed file</description>

    <macros>
        <import>ucsc_macros.xml</import>
    </macros>

    <requirements>
        <requirement type="package" version="340">ucsc_bigbed</requirement>
    </requirements>

    <command detect_errors="exit_code">
<![CDATA[
    @OPTIONAL_PARAM_FUNC@

    #if str($bed_clip) == "bedClip":
        bedClip -verbose=2
            "${bed_input}" "${chrominfo_input}" stdout |
    #else
        cat ${bed_input} |
    #end if

        sort -k 1,1 -k 2,2n  > "${bed_input}".sorted &&

        bedToBigBed

            ## Allow users to include spaces when specifying the extraIndex parameter
            $optional_param("-extraIndex", $extra_index.replace(" ", ""))

            $optional_param("-blockSize", $adv.block_size)
            $optional_param("-itemsPerSlot", $adv.items_per_slot)
            ${adv.no_compression}
            ${adv.tab_separated}

            ## Options for bedPlus format
            #if str($bedplus.bedplus_selector) == "yes":

                #set asql_src = $bedplus.autosql_source
                #set asql_selector = str($asql_src['autosql_source_selector'])

                #if asql_selector == "reference_schema":
                    #set asql_key = str($asql_src['autosql_ref'])
                    #set (asql_schema, asql_type) = asql_key.split("__")

                    #import os.path
                    #set asql_path = os.path.join($__tool_directory__, "schemas", asql_schema)

                    -as=${asql_path}
                    -type=${asql_type}

                #elif asql_selector == "local_schema":
                    #set asql_local = $asql_src['autosql_local']

                    -as=${asql_local.fields.path}
                    -type=${asql_local.fields.type}

                #else
                    #set asql_type = "bed%d+%d" % ($asql_src['num_bed_fields'], $asql_src['num_extra_fields'])

                    -as=$asql_src['autosql_input']
                    -type=${asql_type}
                #end if
            #end if

            "${bed_input}".sorted "${chrominfo_input}" "${bigbed_output}"
]]>
    </command>

    <expand macro="environment_LC_COLLATE" />

    <inputs>
        <param name="bed_input" type="data" format="bed" label="BED file" />

        <param name="chrominfo_input" type="data" format="tabular"
                label="Chromosomes sizes file" />

        <param name="bed_clip" type="boolean" checked="false"
                truevalue="bedClip" falsevalue=""
                label="Remove BED items that extend beyond the scaffolds"
                help="bedClip" />

        <conditional name="bedplus">
            <param name="bedplus_selector" type="select"
                    label="BED file includes extra fields (in bedPlus format)?">
                <option value="yes">Yes</option>
                <option value="no" selected="true">No</option>
            </param>

            <when value="yes">
                <conditional name="autosql_source">
                    <param name="autosql_source_selector" type="select"
                            label="Source of AutoSql schemas"
                            help="The AutoSql schema defines each field in the bedPlus file">
                        <option value="reference_schema">Reference schemas</option>
                        <option value="local_schema">Locally installed schemas</option>
                        <option value="custom_schema">Custom schema</option>
                    </param>

                    <when value="reference_schema">
                        <param name="autosql_ref" type="select"
                                label="Select a reference schema from the list">
                            <option value="bigGenePred.as__bed12+8" selected="true">bigGenePred</option>
                            <option value="bigPsl.as__bed12+12">bigPsl</option>
                            <option value="cytoband.as__bed4+1">Cytoband</option>
                            <option value="rmsk16.as__bed6+10">RepeatMasker</option>
                            <option value="spliceJunctions.as__bed12+1">Splice junctions (regtools)</option>
                            <option value="trf_simpleRepeat.as__bed4+12">Simple repeats (trfBig)</option>
                            <option value="tRNAscan.as__bed4+1">tRNAscan-SE</option>
                        </param>
                    </when>

                    <when value="local_schema">
                        <param name="autosql_local" type="select"
                                label="Select a locally installed AutoSql schema from the list">
                            <options from_data_table="autosql_table">
                                <filter type="sort_by" column="2" />
                                <validator type="no_options" message="No AutoSql schemas available" />
                            </options>
                        </param>
                    </when>

                    <when value="custom_schema">
                        <param name="autosql_input" type="data" format="txt"
                            label="Specify a custom AutoSql file from History" />

                        <param name="num_bed_fields" type="integer" min="3" max="12"
                                value="3" label="Number of BED fields" />

                        <param name="num_extra_fields" type="integer" min="1"
                                value="1" label="Number of extra fields" />
                    </when>
                </conditional>
            </when>

            <when value="no"></when>
        </conditional>

        <param name="extra_index" type="text" label="Index fields" optional="true" value=""
            help="Use a comma-separated list to specify the fields to index (-extraIndex)" />

        <section name="adv" title="Advanced options" expanded="false">
            <param name="block_size" type="integer" label="Block size"
                    min="1" optional="true"
                    help="Number of items to bundle in r-tree" />

            <param name="items_per_slot" type="integer" label="Items per slot"
                    min="1" optional="true"
                    help="Number of data points bundled at the lowest level" />

            <param name="no_compression" type="boolean" checked="false"
                truevalue="-unc" falsevalue=""
                label="Do not use compression"
                help="-unc" />

            <param name="tab_separated" type="boolean" checked="false"
                truevalue="-tab" falsevalue=""
                label="Expect fields to be tab-separated"
                help="-tab" />
        </section>
    </inputs>
    <outputs>
        <data name="bigbed_output" format="bigbed" />
    </outputs>
    <tests>
       <test>
            <!-- Test bedToBigBed with standard bed columns -->
            <param name="bed_input" value="contigs_gaps.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <output name="bigbed_output" file="contigs_gaps.bb" />
        </test>
        <test>
            <!-- Test bedToBigBed with name index that contains extra spaces -->
            <param name="bed_input" value="contigs_gaps.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <param name="extra_index" value="  name  " />
            <output name="bigbed_output" file="contigs_gaps_index.bb" />
        </test>
        <test>
            <!-- Test bedToBigBed with advanced options -->
            <param name="bed_input" value="contigs_gaps.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <param name="block_size" value="200" />
            <param name="items_per_slot" value="500" />
            <param name="no_compression" value="-unc" />
            <param name="tab_separated" value="-tab" />
            <output name="bigbed_output" file="contigs_gaps_advanced.bb" />
        </test>
        <test>
            <!-- Test bedToBigBed with bedClip -->
            <param name="bed_input" value="contigs_clip.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <param name="bed_clip" value="bedClip" />
            <output name="bigbed_output" file="contigs_gaps.bb" />
        </test>
        <test>
            <!-- Test bedToBigBed with bedPlus and G-OnRamp schema -->
            <param name="bed_input" value="contigs_cytoband.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <param name="bedplus_selector" value="yes" />
            <param name="autosql_source_selector" value="reference_schema" />
            <param name="autosql_ref" value="cytoband.as__bed4+1" ftype="txt" />
            <output name="bigbed_output" file="contigs_cytoband.bb" />
        </test>
        <test>
            <!-- Test bedToBigBed with bedPlus and local schema -->
            <param name="bed_input" value="contigs_narrowPeak.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <param name="bedplus_selector" value="yes" />
            <param name="autosql_source_selector" value="local_schema" />
            <param name="autosql_local" value="narrowPeak.as" />
            <output name="bigbed_output" file="contigs_narrowPeak.bb" />
        </test>
        <test>
            <!-- Test bedToBigBed with bedPlus and custom schema -->
            <param name="bed_input" value="contigs_cytoband.bed" ftype="bed" />
            <param name="chrominfo_input" value="contigs_chromInfo.tab" ftype="tabular" />
            <param name="bedplus_selector" value="yes" />
            <param name="autosql_source_selector" value="custom_schema" />
            <param name="autosql_input" value="cytoband.as" ftype="txt" />
            <param name="num_bed_fields" value="4" />
            <param name="num_extra_fields" value="1" />
            <output name="bigbed_output" file="contigs_cytoband.bb" />
        </test>
    </tests>
    <help>
<![CDATA[
**What it does**

This tool converts a `BED <https://genome.ucsc.edu/FAQ/FAQformat.html#format1>`_
file into a `bigBed <https://genome.ucsc.edu/goldenpath/help/bigBed.html>`_ file.

----

.. class:: infomark

**Chromosomes sizes file**

This tool requires a chromosomes sizes (``chrom.sizes``) file, which lists the size of
each scaffold within an assembly. For genome assemblies that are hosted by UCSC,
the ``chrom.sizes`` file is available through the
`Sequence and Annotation Downloads <http://hgdownload.cse.ucsc.edu/downloads.html>`_ page.
The `twoBitInfo <https://genome.ucsc.edu/goldenpath/help/twoBit.html>`_ tool
can be used to generate the ``chrom.sizes`` for the genome assemblies that are not hosted
by UCSC.

----

.. class:: infomark

**AutoSql files**

The BED input file could contain extra fields (i.e. bedPlus). The definitions for the
fields in a bedPlus file can be specified using an
`AutoSql <http://genomewiki.ucsc.edu/index.php/AutoSql>`_ file.

A large collection of AutoSql files are available through the
`UCSC Genome Browser source tree <https://github.com/ucscGenomeBrowser/kent/tree/master/src/hg/lib>`_
(files with the .as extension).

    ]]></help>

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