comparison umi-tools_whitelist.xml @ 14:345bdf4546fd draft

"planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/umi_tools commit bf6a3aa532e8f9d122da4c1e39f3e256ae587b79"

13:49bfe13676a1

<tool id="umi_tools_whitelist" name="UMI-tools whitelist" version="@VERSION@.0">
    <description>Extract cell barcodes from FASTQ files</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <command detect_errors="exit_code"><![CDATA[
        @COMMAND_LINK@

        umi_tools whitelist
            --bc-pattern='$bc_pattern'
            --extract-method='$extract_method'
            --subset-reads='$subset_reads'
            #if $input_type.type == 'single':
                #if $gz:
                    --stdin=input_single.gz
                #else
                    --stdin=input_single.txt
                #end if

                    --read2-in=input_read2.gz
                #else:
                    --stdin=input_read1.txt
                    --read2-in=input_read2.txt
                #end if
                #if $input_type.barcode.barcode_select == "both_reads":
                    --bc-pattern2='$input_type.barcode.bc_pattern2'
                #end if
            #end if
            #if $celloptions.use_cell_opts == "advanced":
                #if $celloptions.set_cell_number != "0":
                   --set-cell-number=$celloptions.set_cell_number
                #end if
                #if $celloptions.expect_cells != "0":
                   --expect-cells=$celloptions.expect_cells
                #end if
                --error-correct-threshold=$celloptions.error_correct_thresh
            #end if
            --method=$method
            --plot-prefix=OUT

            $prime3

            #if $log:
                --log='$out_log'
            #end if

            --log2stderr

            > '$out_whitelist' &&

            mkdir '${ out_html_report.files_path }' &&
            cp OUT_*.png '${ out_html_report.files_path }' &&

            &&
            mv OUT_cell_thresholds.tsv '$out_thresh'
    ]]></command>
    <inputs>
        <expand macro="input_types" />
        <param name="bc_pattern" argument="--bc-pattern" type="text" label="Barcode pattern for first read"
            help="Use this option to specify the format of the UMI/barcode. Use Ns to
                    represent the random positions and Xs to indicate the bc positions.
                    Bases with Ns will be extracted and added to the read name. Remaining
                    bases, marked with an X will be reattached to the read.">
            <expand macro="barcode_sanitizer" />
        </param>
        <param name="extract_method" argument="--extract-method" type="select" label="Barcode Extraction Method"
               help="If bracketed expressions are used in the above barcode pattern, then set this to 'regex'. Otherwise leave as 'string'" >
            <option value="string" selected="true" />
            <option value="regex" />
        </param>
        <param name="method" argument="--method" type="select" label="Count reads or UMIs"
               help="Many published protocols rank CBs by the number of reads the CBs appear in. However you could also use the number of unique UMIs a CB is associated with. Note that this is still and approximation to the number of transcripts captured because the same UMI could be associated with two different transcripts and be counted as independent." >
            <option value="reads" selected="true" />
            <option value="umis" />
        </param>
        <param argument="--3prime" name="prime3" type="boolean" label="Is barcode on 3' end of the read?"
            truevalue="--3prime" falsevalue=""
            help="By default the barcode is assumed to be on the 5' end of the read, but
                use this option to specify that it is on the 3' end instead." />
        <param name="subset_reads" argument="--subset-reads" type="integer" min="0" value="0" label="Use the first N reads to automatically identify the true cell barcodes." />
        <conditional name="celloptions" >
            <param name="use_cell_opts" type="select" label="Cell parameters" >
                <option value="defaults" selected="True">Use Defaults</option>
                <option value="advanced">Advanced Options</option>
            </param>
            <when value="defaults"/>
            <when value="advanced">
                <param name="set_cell_number" type="integer" min="0" value="0" label="Specify the number of cell barcodes to accept" />
                <param name="expect_cells" type="integer" min="0" value="0" label="Prior expectation on the upper limit on the number of cells sequenced" />
                <param name="error_correct_thresh" type="integer" min="0" value="0" label="Hamming distance for correction of barcodes to whitelist barcodes. Set to zero to generate no error correcting metrics." />
            </when>
        </conditional>
        <param argument="--log" type="boolean" label="Output log?" truevalue="--log" falsevalue=""
            help="Choose if you want to generate a text file containing logging information." />
    </inputs>
    <outputs>
        <data name="out_whitelist" format="tabular" label="${tool.name} on ${on_string}: Whitelist"/>
        <data name="out_log" format="txt" label="${tool.name} on ${on_string}: logfile" >
            <filter>log</filter>
        </data>
        <data name="out_html_report" format="html" label="${tool.name} on ${on_string}: Webpage" />
        <data name="out_thresh" format="tabular" label="${tool.name} on ${on_string}: TSV Cell Thresholds" />
    </outputs>
    <tests>
        <test expect_num_outputs="3">
            <conditional name="input_type" >
                <param name="type" value="single" />
                <param name="input_single" value="t_R2.fastq.gz" ftype="fastq.gz" />
            </conditional>
            <param name="bc_pattern" value="CCCCCCCCNNNNNNNN" />
            <param name="method" value="reads" />
            <param name="prime3" value="true" />
            <output name="out_whitelist" file="out_wl_single.txt" lines_diff="40" />
            <output name="out_thresh" file="out_wl_single.tresh.tab" />
            <output name="out_html_report" file="out_wl_single.html" />
        </test>
        <test expect_num_outputs="4">
            <conditional name="input_type" >
                <param name="type" value="paired" />
                <param name="input_read1" value="t_R1.fastq.gz" ftype="fastq.gz" />
                <param name="input_read2" value="t_R2.fastq.gz" ftype="fastq.gz" />
            </conditional>
            <param name="barcode_select" value="first_read_only" />
            <param name="bc_pattern" value="CCCNNNNNNNNXXXXX" />
            <param name="bc_pattern2" value="CCCCCCCCNNNNNNNN" />
            <param name="method" value="reads" />
            <param name="prime3" value="false" />
            <param name="use_cell_opts" value="advanced" />
            <param name="expect_cells" value="5" />
            <param name="error_correct_thresh" value="3" />
            <param name="log" value="true" />
            <output name="out_whitelist" file="out_wl_paired.txt" />
            <output name="out_log" file="out_wl_paired.log" lines_diff="50" />
            <output name="out_html_report" file="out_wl_paired.html" />
            <output name="out_thresh" file="out_wl_paired.tresh.tab" />
        </test>
        <test expect_num_outputs="4"> <!-- As previous, identical outputs but paired collection input -->
            <conditional name="input_type" >
                <param name="type" value="paired_collection" />
                <param name="input_readpair" >
                    <collection type="paired">
                        <element name="forward" ftype="fastq.gz" value="t_R1.fastq.gz" />
                        <element name="reverse" ftype="fastq.gz" value="t_R2.fastq.gz" />
                    </collection>
                </param>
            </conditional>
            <param name="barcode_select" value="first_read_only" />
            <param name="bc_pattern"  value="CCCNNNNNNNNXXXXX" />
            <param name="bc_pattern2" value="CCCCCCCCNNNNNNNN" />
            <param name="method" value="reads" />
            <param name="prime3" value="false" />
            <param name="use_cell_opts" value="advanced" />
            <param name="expect_cells" value="5" />
            <param name="error_correct_thresh" value="3" />
            <param name="log" value="true" />
            <output name="out_whitelist" file="out_wl_paired.txt" />
            <output name="out_log" file="out_wl_paired.log" lines_diff="50" />
            <output name="out_html_report" file="out_wl_paired.html" />
            <output name="out_thresh" file="out_wl_paired.tresh.tab" />
        </test>
        <!-- Error report on not accepting regex and lt and gt symbols -->
        <test expect_num_outputs="3">
            <conditional name="input_type" >
                <param name="type" value="single" />
                <param name="input_single" value="testYYY.40k.fastq.gz" ftype="fastq.gz" />
            </conditional>
            <param name="bc_pattern" value="(?P&#60;cell_1&#62;.{8,10})(?P&#60;discard_1&#62;ACTGGCCTGCGA){s&#60;=3}(?P&#60;cell_2&#62;.{9})(?P&#60;discard_2&#62;GGTAGCGGTGACA){s&#60;=3}(?P&#60;cell_3&#62;.{9})(?P&#60;umi_1&#62;.{8})T{3}.*" />
            <param name="extract_method" value="regex" />
            <param name="method" value="umis" />
            <param name="prime3" value="true" />
            <output name="out_whitelist" file="out_wl_user.single.txt" />
            <output name="out_thresh" file="out_wl_user.single.tresh.tab" />
            <output name="out_html_report" file="out_wl_user.single.html" />
        </test>
    </tests>
    <help><![CDATA[

-------

Extract cell barcodes and identify the most likely true barcodes using
the 'knee' method.

Options
-------

--bc-pattern
       This should be used where the barcodes are always in the same
       place in the read.

       - N = UMI position (required)
       - C = cell barcode position (optional)
       - X = sample position (optional)

       Bases with Ns and Cs will be extracted and added to the read
       name. The corresponding sequence qualities will be removed from
       the read. Bases with an X will be reattached to the read.

       E.g. If the pattern is NNNNCC,
       Then the read:
       @HISEQ:87:00000000 read1
       AAGGTTGCTGATTGGATGGGCTAG
       DA1AEBFGGCG01DFH00B1FF0B
       +
       will become:
       @HISEQ:87:00000000_TT_AAGG read1
       GCTGATTGGATGGGCTAG
       1AFGGCG01DFH00B1FF0B
       +

       where 'TT' is the cell barcode and 'AAGG' is the UMI.


--set-cell-number
        Use this option to explicity set the number of cell barcodes
        which should be accepted. Note that the exact number of cell
        barcodes in the outputted whitelist may be slightly less than
        this if there are multiple cells observed with the same
        frequency at the threshold between accepted and rejected cell
        barcodes.

--expect-cells=[EXPECTED_CELLS]
        An upper limit estimate for the number of inputted cells. The knee
        method will now select the first threshold (order ascendingly)
        which results in the number of cell barcodes accepted being <=
        EXPECTED_CELLS and > EXPECTED_CELLS * 0.1.


--bc-pattern2
       Use this option to specify the format of the UMI/barcode for
       the second read pair if required. If --bc-pattern2 is not
       supplied, this defaults to the same pattern as --bc-pattern

--3prime
       By default the barcode is assumed to be on the 5' end of the read, but
       use this option to sepecify that it is on the 3' end instead

Usage:
------

For single ended reads:
        umi_tools whitelist --bc-pattern=[PATTERN] -L extract.log
        [OPTIONS]

reads from stdin and outputs to stdout.

For paired end reads where the cell barcodes is split across the read pairs:
        umi_tools whitelist --bc-pattern=[PATTERN]
        --bc-pattern2=[PATTERN] --read2-in=[FASTQIN] -L extract.log
        [OPTIONS]

reads end one from stdin and end two from FASTQIN and outputs to stdin


Output:
-------

14:345bdf4546fd

<tool id="umi_tools_whitelist" name="UMI-tools whitelist" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
    <description>Extract cell barcodes from FASTQ files</description>
    <expand macro="bio_tools"/>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <command detect_errors="exit_code"><![CDATA[
        #import json
        @COMMAND_LINK@
        umi_tools whitelist
            
            @FASTQ_BARCODE_EXTRACTION_OPTIONS@ 

            --subset-reads='$subset_reads'
            $allow_threshold_error
            #if $ed_above_threshold
                --ed-above-threshold=$ed_above_threshold
            #end if
            --knee-method='$knee_method'
            #if $input_type_cond.input_type == 'single':
                #if $gz:
                    --stdin=input_single.gz
                #else
                    --stdin=input_single.txt
                #end if

                    --read2-in=input_read2.gz
                #else:
                    --stdin=input_read1.txt
                    --read2-in=input_read2.txt
                #end if
            #end if
            #if $celloptions.use_cell_opts == "advanced":
                #if str($celloptions.set_cell_number) != '':
                   --set-cell-number=$celloptions.set_cell_number
                #end if
                #if str($celloptions.expect_cells) != '':
                   --expect-cells=$celloptions.expect_cells
                #end if
                --error-correct-threshold=$celloptions.error_correct_threshold
            #end if
            --method=$method
            --plot-prefix=OUT

            @LOG@
            > '$out_whitelist' &&

            mkdir '${ out_html_report.files_path }' &&
            cp OUT_*.png '${ out_html_report.files_path }' &&

            &&
            mv OUT_cell_thresholds.tsv '$out_thresh'
    ]]></command>
    <inputs>
        <expand macro="input_types" />
        <expand macro="fastq_barcode_extraction_options_macro"/>

        <param argument="--method" type="select" label="Count reads or UMIs"
               help="Many published protocols rank CBs by the number of reads the CBs appear in. However you could also use the number of unique UMIs a CB is associated with. Note that this is still and approximation to the number of transcripts captured because the same UMI could be associated with two different transcripts and be counted as independent" >
            <option value="reads" selected="true" />
            <option value="umis" />
        </param>
        <!-- TODO Cannot use expect-cells with 'distance' knee method.-->
        <param argument="--knee-method" type="select" label="Method for detection of knee">
            <option value="distance" selected="true" />
            <option value="density"/>
        </param>
        <param argument="--subset-reads" type="integer" min="0" value="0" label="Use the first N reads to automatically identify the true cell barcodes" />
        <param argument="--allow-threshold-error" type="boolean" truevalue="--allow-threshold-error" falsevalue="" label="Don't select a threshold" help="Will still output the plots if requested"/> 
        <param argument="--ed-above-threshold" type="select" optional="true" label="Detect and correct CBs above the threshold" help="which may be sequence errors from another CB">
            <option value="correct">correct</option>
            <option value="discard">discard</option>
        </param>
        <conditional name="celloptions" >
            <param name="use_cell_opts" type="select" label="Cell parameters" >
                <option value="defaults" selected="True">Use Defaults</option>
                <option value="advanced">Advanced Options</option>
            </param>
            <when value="defaults"/>
            <when value="advanced">
                <param argument="--set-cell-number" type="integer" min="0" optional="true" label="Specify the number of cell barcodes to accept" />
                <param argument="--expect-cells" type="integer" min="0" optional="true" label="Prior expectation on the upper limit on the number of cells sequenced" />
                <param argument="--error-correct-threshold" type="integer" min="0" value="1" label="Hamming distance for correction of barcodes to whitelist barcodes. Set to zero to generate no error correcting metrics" />
            </when>
        </conditional>
        <expand macro="log_input_macro"/>
    </inputs>
    <outputs>
        <data name="out_whitelist" format="tabular" label="${tool.name} on ${on_string}: Whitelist"/>
        
        <data name="filtered_out" format_source="input_read1" label="${tool.name} on ${on_string}: reads not matching regex pattern">
            <filter>extract_method_cond['extract_method'] == 'regex' and extract_method_cond['filtered_out_bool'] and input_type_cond['input_type'] in ['single', 'paired']</filter>
        </data>
        <data name="filtered_out_paired" format_source="input_read2" label="${tool.name} on ${on_string}: reads not matching regex pattern">
            <filter>extract_method_cond['extract_method'] == 'regex' and extract_method_cond['filtered_out_bool'] and input_type_cond['input_type'] == 'paired'</filter>
        </data>
        <collection name="filtered_out_paired_collection" type="paired" label="${tool.name} on ${on_string}: reads not matching regex pattern">
            <data name="forward" format_source="input_readpair" />
            <data name="reverse" format_source="input_readpair" />
            <filter>extract_method_cond['extract_method'] == 'regex' and extract_method_cond['filtered_out_bool'] and input_type_cond['input_type'] == 'paired_collection'</filter>
        </collection>
        <expand macro="log_output_macro"/>
        <data name="out_html_report" format="html" label="${tool.name} on ${on_string}: Webpage" />
        <data name="out_thresh" format="tabular" label="${tool.name} on ${on_string}: TSV Cell Thresholds" />
    </outputs>
    <tests>
        <test expect_num_outputs="3">
            <conditional name="input_type_cond" >
                <param name="input_type" value="single" />
                <param name="input_read1" value="t_R2.fastq.gz" ftype="fastqsanger.gz" />
                <param name="bc_pattern" value="CCCCCCCCNNNNNNNN" />
            </conditional>
            <conditional name="extract_method_cond">
                <param name="prime3" value="true" />
            </conditional>
            <param name="method" value="reads" />
            <param name="knee_method" value="density"/>
            <output name="out_whitelist" file="out_wl_single.txt" lines_diff="40" />
            <output name="out_thresh" file="out_wl_single.tresh.tab" />
            <output name="out_html_report" file="out_wl_single.html" />
        </test>
        <test expect_num_outputs="4">
            <conditional name="input_type_cond" >
                <param name="input_type" value="paired" />
                <param name="input_read1" value="t_R1.fastq.gz" ftype="fastqsanger.gz" />
                <param name="input_read2" value="t_R2.fastq.gz" ftype="fastqsanger.gz" />
                <param name="bc_pattern" value="CCCNNNNNNNNXXXXX" />
                <!-- <param name="bc_pattern2" value="CCCCCCCCNNNNNNNN" /> -->
            </conditional>
            <param name="method" value="reads" />
            <conditional name="extract_method_cond">
                <param name="prime3" value="false" />
            </conditional>
            <param name="use_cell_opts" value="advanced" />
            <param name="knee_method" value="density"/>
            <param name="expect_cells" value="5" />
            <param name="error_correct_threshold" value="3" />
            <param name="log" value="true" />
            <output name="out_whitelist" file="out_wl_paired.txt" />
            <output name="out_log" file="out_wl_paired.log" lines_diff="50" />
            <output name="out_html_report" file="out_wl_paired.html" />
            <output name="out_thresh" file="out_wl_paired.tresh.tab" />
        </test>
        <test expect_num_outputs="4"> <!-- As previous, identical outputs but paired collection input -->
            <conditional name="input_type_cond" >
                <param name="input_type" value="paired_collection" />
                <param name="input_readpair" >
                    <collection type="paired">
                        <element name="forward" ftype="fastqsanger.gz" value="t_R1.fastq.gz" />
                        <element name="reverse" ftype="fastqsanger.gz" value="t_R2.fastq.gz" />
                    </collection>
                </param>
                <param name="bc_pattern"  value="CCCNNNNNNNNXXXXX" />
                <!-- <param name="bc_pattern2" value="CCCCCCCCNNNNNNNN" /> -->
            </conditional>
            <param name="method" value="reads" />
            <conditional name="extract_method_cond">
                <param name="prime3" value="false" />
            </conditional>
            <param name="use_cell_opts" value="advanced" />
            <param name="knee_method" value="density"/>
            <param name="expect_cells" value="5" />
            <param name="error_correct_threshold" value="3" />
            <param name="log" value="true" />
            <param name="filtered_out_bool" value="true"/>
            <output name="out_whitelist" file="out_wl_paired.txt" />
            <output name="out_log" file="out_wl_paired.log" lines_diff="50" />
            <output name="out_html_report" file="out_wl_paired.html" />
            <output name="out_thresh" file="out_wl_paired.tresh.tab" />
        </test>
        <!-- Error report on not accepting regex and lt and gt symbols -->
        <test expect_num_outputs="4">
            <conditional name="input_type_cond" >
                <param name="input_type" value="single" />
                <param name="input_read1" value="testYYY.40k.fastq.gz" ftype="fastqsanger.gz" />
            </conditional>
            <param name="bc_pattern" value="(?P&#60;cell_1&#62;.{8,10})(?P&#60;discard_1&#62;ACTGGCCTGCGA){s&#60;=3}(?P&#60;cell_2&#62;.{9})(?P&#60;discard_2&#62;GGTAGCGGTGACA){s&#60;=3}(?P&#60;cell_3&#62;.{9})(?P&#60;umi_1&#62;.{8})T{3}.*" />
            <param name="extract_method" value="regex" />
            <param name="method" value="umis" />
            <param name="knee_method" value="density"/>
            <param name="prime3" value="true" />
            <param name="filtered_out_bool" value="true"/>
            <output name="out_whitelist" file="out_wl_user.single.txt" />
            <output name="out_thresh" file="out_wl_user.single.tresh.tab" />
            <output name="out_html_report" file="out_wl_user.single.html" />
            <output name="filtered_out">
                <assert_contents>
                    <has_text text="@A00250:74:HFMFVDSXX:2:1101:1027:1016 1:N:0:GTAGAGGA"/>
                </assert_contents>
            </output>
        </test>
    </tests>
    <help><![CDATA[

-------

Extract cell barcodes and identify the most likely true barcodes using
the 'knee' method.

@FASTQ_BARCODE_EXTRACTION_HELP@



Output:
-------