comparison msstatstmt.xml @ 4:8375a0035d79 draft

"planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/msstatstmt commit 0ed9b6032443be289192cb096c4a5694fccd3639"

3:5667ff6f7a40

<tool id="msstatstmt" name="MSstatsTMT" version="@TOOL_VERSION@+galaxy@GALAXY_VERSION@">
    <description>protein significance analysis in shotgun mass spectrometry-based proteomic experiments with tandem mass tag (TMT) labeling</description>
    <macros>
        <token name="@TOOL_VERSION@">1.8.2</token>
        <token name="@GALAXY_VERSION@">1</token>
        <xml name="input_options_shared">
            <param name="useUniquePeptide" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="true" label="Use unique peptide" help="Yes (default) removes peptides that are assigned for more than one protein. We assume to use unique peptide for each protein." />
            <param name="rmPSM_withMissing_withinRun" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="false" label="Remove PSM with missing value within run" help="Yes will remove PSM with any missing value within each run. Default is No." />
            <param name="rmPSM_withfewMea_withinRun" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="true" label="Remove PSM with few measurements within run" help="Only for 'Remove PSM with missing value within run = No'. Yes (default) will remove the features that have 1 or 2 measurements within each run." />
            <param name="rmProtein_with1Feature" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="false" label="Remove proteins with one feature" help="Yes will remove the proteins which have only 1 peptide and charge. Default is No." />

        <requirement type="package" version="@TOOL_VERSION@">bioconductor-msstatstmt</requirement>
    </requirements>
    <command detect_errors="exit_code"><![CDATA[
        cat '$msstatstmt_script' > '$out_r_script' &&
        Rscript '$msstatstmt_script'
        && cat msstats*.log > '$out_msstats_log'
        && cat msstatstmt.log > '$out_msstatstmt_log'
    ]]></command>
    <configfiles>
        <configfile name="msstatstmt_script"><![CDATA[
library(MSstatsTMT, warn.conflicts = F, quietly = T, verbose = F)
library(MSstats)

                              maxQuantileforCensored = $proteinSummarization.maxQuantileforCensored)
                              #end if

#for $plot_type in $selected_outputs
    #if $plot_type[-4:] == "Plot"
        dataProcessPlotsTMT(input,
                            quant,
                            type = '$plot_type',
                            ylimUp = $out_plots_opt.adv.ylimUp,
                            ylimDown = $out_plots_opt.adv.ylimDown,
                            x.axis.size = $out_plots_opt.adv.x_axis_size,
                            y.axis.size = $out_plots_opt.adv.y_axis_size,

                            which.Protein = unlist(read.table("$out_plots_opt.which_Protein.protein_list", sep = "\n", header = FALSE), use.names = FALSE),
                            #end if
                            originalPlot = $out_plots_opt.adv.originalPlot,
                            summaryPlot = $out_plots_opt.adv.summaryPlot)
    #elif $plot_type == "quant"
        write.table(quant,
                "Quant.tsv",
                sep = "\t",
                quote = F,
                row.names = F,
                dec = ".")

    #if $group.use_comp_matrix.select == 'true'
        comp_matrix <- read.table("$group.use_comp_matrix.comparison_matrix", sep="\t", header=TRUE, check.names=FALSE)

        comparison <- comp_matrix[,-1]
        row.names(comparison) <- as.character(comp_matrix[,1])
        comparison <- as.matrix(comparison[levels(quant\$Condition)])
    #end if

    comparisons <- groupComparisonTMT(data = quant,
                                      #if $group.use_comp_matrix.select == 'true'
                                      contrast.matrix = comparison,

                                      moderated = $group.moderated,
                                      adj.method = "$group.adj_method",
                                      remove_norm_channel = $group.remove_norm_channel,
                                      remove_empty_channel = $group.remove_empty_channel)

    write.table(comparisons,
                "ComparisonResult.tsv",
                sep = "\t",
                quote = F,
                row.names = F,
                dec = ".")

    #for $plot_type in $group.selected_group_outputs
       #if $plot_type == "VolcanoPlot" or $plot_type == "Heatmap" or $plot_type == "ComparisonPlot"
        groupComparisonPlots(data = comparisons,
                             type = "$plot_type",
                             sig = $group.comparison_plots_opt.sig,
                             #if $group.comparison_plots_opt.FCcutoff:
                             FCcutoff = $group.comparison_plots_opt.FCcutoff,
                             #end if

                    </conditional>
                </section>
            </when>
        </conditional>
        <param name="selected_outputs" type="select" multiple="true" optional="false" label="Select Outputs">
            <option value="msstats_log" selected="true">MSstats log</option>
            <option value="msstatstmt_log" selected="true">MSstatsTMT log</option>
            <option value="r_script" selected="false">MSstats Rscript</option>
            <option value="quant" selected="true">Protein abundance</option>
            <option value="ProfilePlot" selected="false">Profile Plot</option>
            <option value="QCPlot" selected="false">QC Plot</option>
        </param>
    </inputs>
    <outputs>
        <data name="out_msstats_log" from_work_dir="msstats.log" format="txt" label="${tool.name} on ${on_string}: MSstats log">
            <filter>'msstats_log' in selected_outputs</filter>
        </data>
        <data name="out_msstatstmt_log" from_work_dir="msstatstmt.log" format="txt" label="${tool.name} on ${on_string}: MSstatsTMT log">
            <filter>'msstatstmt_log' in selected_outputs</filter>
        </data>
        <data name="out_r_script" format="txt" label="${tool.name} on ${on_string}: Rscript">
            <filter>'r_script' in selected_outputs</filter>
        </data>
        <data name="out_quant" from_work_dir="Quant.tsv" format="tabular" label="${tool.name} on ${on_string}: Protein abundance">

        <data name="out_group_comp_plot" from_work_dir="MSstats_group_ComparisonPlot.pdf" format="pdf" label="${tool.name} on ${on_string}: Group Comparison - Comparison Plot">
            <filter>group['group_comparison'] == 'true' and 'ComparisonPlot' in group['selected_group_outputs']</filter>
        </data>
    </outputs>
    <tests>
        <test expect_num_outputs="5">
            <conditional name="input">
                <param name="input_src" value="MSstatsTMT"/>
                <param name="msstatstmt_input" ftype="tabular" value="input.msstatstmt.txt"/>
            </conditional>
            <param name="selected_outputs" value="msstats_log,msstatstmt_log,r_script,ProfilePlot,QCPlot"/>
            <output name="out_msstats_log">
                <assert_contents>
                    <has_text text="1 level of Isotope type labeling in this experiment" />
                </assert_contents>
            </output>
            <output name="out_msstatstmt_log">
                <assert_contents>
                    <has_text text="MSstatsTMT - proteinSummarization function" />
                </assert_contents>
            </output>
            <output name="out_r_script">
                <assert_contents>
                    <has_n_lines n="52" />
                </assert_contents>
            </output>
            <output name="out_profile_plot" file="ProfilePlot.pdf" compare="sim_size"/>
            <output name="out_qc_plot" file="QCPlot.pdf" compare="sim_size"/>
        </test>
        <test expect_num_outputs="4">
            <conditional name="input">
                <param name="input_src" value="MaxQuant"/>
                <param name="evidence" ftype="tabular" value="evidence.txt"/>
                <param name="annotation" ftype="tabular" value="annotation.txt"/>
                <param name="proteinGroups" ftype="tabular" value="proteinGroups.txt"/>

                    <param name="select" value="true"/>
                    <param name="comparison_matrix" ftype="tabular" value="comparison_matrix.txt"/>
                </conditional>
                <param name="selected_group_outputs" value="comparison_result,VolcanoPlot,ComparisonPlot"/>
            </conditional>
            <param name="selected_outputs" value="ProfilePlot"/>
            <conditional name="which_Protein">
                <param name="select" value="list"/>
                <param name="protein_list" ftype="tabular" value="proteinIDs.txt"/>
            </conditional>
            <output name="out_group_comp">

                <param name="selected_group_outputs" value="comparison_result,Heatmap"/>
            </conditional>
            <param name="selected_outputs" value="r_script"/>
            <output name="out_r_script">
                <assert_contents>
                    <has_n_lines n="58" />
                </assert_contents>
            </output>
            <output name="out_group_comp">
                <assert_contents>
                    <has_n_lines n="51" />
                    <has_n_columns n="8" />
                    <has_text text="Long_LF-Short_HF" />
                    <has_text text="sp|O35226|PSMD4_MOUSE" />
                </assert_contents>
            </output>
            <output name="out_group_heatmap" file="MSstats_group_Heatmap.pdf" compare="sim_size"/>
        </test>

    #> 2 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 3 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 4 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 5 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 6 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #>   PrecursorCharge                                             PSM Mixture
    #> 1               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 2               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 3               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 4               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 5               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3

4:8375a0035d79

<tool id="msstatstmt" name="MSstatsTMT" version="@TOOL_VERSION@+galaxy@GALAXY_VERSION@">
    <description>protein significance analysis in shotgun mass spectrometry-based proteomic experiments with tandem mass tag (TMT) labeling</description>
    <macros>
        <token name="@TOOL_VERSION@">2.0.0</token>
        <token name="@GALAXY_VERSION@">0</token>
        <xml name="input_options_shared">
            <param name="useUniquePeptide" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="true" label="Use unique peptide" help="Yes (default) removes peptides that are assigned for more than one protein. We assume to use unique peptide for each protein." />
            <param name="rmPSM_withMissing_withinRun" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="false" label="Remove PSM with missing value within run" help="Yes will remove PSM with any missing value within each run. Default is No." />
            <param name="rmPSM_withfewMea_withinRun" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="true" label="Remove PSM with few measurements within run" help="Only for 'Remove PSM with missing value within run = No'. Yes (default) will remove the features that have 1 or 2 measurements within each run." />
            <param name="rmProtein_with1Feature" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="false" label="Remove proteins with one feature" help="Yes will remove the proteins which have only 1 peptide and charge. Default is No." />

        <requirement type="package" version="@TOOL_VERSION@">bioconductor-msstatstmt</requirement>
    </requirements>
    <command detect_errors="exit_code"><![CDATA[
        cat '$msstatstmt_script' > '$out_r_script' &&
        Rscript '$msstatstmt_script'

        && cat MSstatsTMT_summarization_log*.log > '$out_msstatstmt_sum_log'
        && cat MSstatsTMT_summarization_MSstats*.log > '$out_msstatstmt_sum_ms'

        #if $input.input_src != 'MSstatsTMT':
            && cat MSstatsTMT_converter_log*.log > '$out_msstatstmt_conv_log'
        #else:
            && echo "MSstatsTMT input. No conversion needed." > '$out_msstatstmt_conv_log'
        #end if

        #if $group.group_comparison == 'true':
            && cat MSstatsTMT_log_groupComparison*.log > '$out_msstatstmt_group_log'
            && cat MSstats_log*.log > '$out_msstats_gc_log'
        #else:
            && echo "No groupComparison was performed." > '$out_msstatstmt_group_log'
            && echo "No groupComparison was performed." > '$out_msstats_gc_log'
        #end if

    ]]></command>
    <configfiles>
        <configfile name="msstatstmt_script"><![CDATA[
library(MSstatsTMT, warn.conflicts = F, quietly = T, verbose = F)
library(MSstats)

                              maxQuantileforCensored = $proteinSummarization.maxQuantileforCensored)
                              #end if

#for $plot_type in $selected_outputs
    #if $plot_type[-4:] == "Plot"
        dataProcessPlotsTMT(quant,
                            type = '$plot_type',
                            ylimUp = $out_plots_opt.adv.ylimUp,
                            ylimDown = $out_plots_opt.adv.ylimDown,
                            x.axis.size = $out_plots_opt.adv.x_axis_size,
                            y.axis.size = $out_plots_opt.adv.y_axis_size,

                            which.Protein = unlist(read.table("$out_plots_opt.which_Protein.protein_list", sep = "\n", header = FALSE), use.names = FALSE),
                            #end if
                            originalPlot = $out_plots_opt.adv.originalPlot,
                            summaryPlot = $out_plots_opt.adv.summaryPlot)
    #elif $plot_type == "quant"
        write.table(quant\$ProteinLevelData,
                "Quant.tsv",
                sep = "\t",
                quote = F,
                row.names = F,
                dec = ".")

    #if $group.use_comp_matrix.select == 'true'
        comp_matrix <- read.table("$group.use_comp_matrix.comparison_matrix", sep="\t", header=TRUE, check.names=FALSE)

        comparison <- comp_matrix[,-1]
        row.names(comparison) <- as.character(comp_matrix[,1])
        comparison <- as.matrix(comparison[levels(quant\$ProteinLevelData\$Condition)])
    #end if

    comparisons <- groupComparisonTMT(data = quant,
                                      #if $group.use_comp_matrix.select == 'true'
                                      contrast.matrix = comparison,

                                      moderated = $group.moderated,
                                      adj.method = "$group.adj_method",
                                      remove_norm_channel = $group.remove_norm_channel,
                                      remove_empty_channel = $group.remove_empty_channel)

    write.table(comparisons\$ComparisonResult,
                "ComparisonResult.tsv",
                sep = "\t",
                quote = F,
                row.names = F,
                dec = ".")

    #for $plot_type in $group.selected_group_outputs
       #if $plot_type == "VolcanoPlot" or $plot_type == "Heatmap" or $plot_type == "ComparisonPlot"

        ## Workaround for missing option "MSstatsLog" (getOption("MSstatsLog") returns NULL)
        MSstatsConvert::MSstatsLogsSettings()
        #############################################

        groupComparisonPlots(data = comparisons\$ComparisonResult,
                             type = "$plot_type",
                             sig = $group.comparison_plots_opt.sig,
                             #if $group.comparison_plots_opt.FCcutoff:
                             FCcutoff = $group.comparison_plots_opt.FCcutoff,
                             #end if

                    </conditional>
                </section>
            </when>
        </conditional>
        <param name="selected_outputs" type="select" multiple="true" optional="false" label="Select Outputs">
            <option value="msstatstmt_conv_log" selected="false">MSstatsTMT converter log</option>
            <option value="msstatstmt_sum_log" selected="true">MSstatsTMT summarization log</option>
            <option value="msstatstmt_sum_ms" selected="true">MSstatsTMT summarization MSstats</option>
            <option value="msstatstmt_group_log" selected="false">MSstatsTMT groupComparison log</option>
            <option value="msstats_gc_log" selected="false">MSstats log (groupComparison)</option>
            <option value="r_script" selected="false">MSstats Rscript</option>
            <option value="quant" selected="true">Protein abundance</option>
            <option value="ProfilePlot" selected="false">Profile Plot</option>
            <option value="QCPlot" selected="false">QC Plot</option>
        </param>
    </inputs>
    <outputs>
        <data name="out_msstatstmt_conv_log" from_work_dir="MSstatsTMT_converter_log.log" format="txt" label="${tool.name} on ${on_string}: MSstatsTMT converter log">
            <filter>'msstatstmt_conv_log' in selected_outputs</filter>
        </data>
        <data name="out_msstatstmt_sum_log" from_work_dir="MSstatsTMT_summarization_log.log" format="txt" label="${tool.name} on ${on_string}: MSstatsTMT summarization log">
            <filter>'msstatstmt_sum_log' in selected_outputs</filter>
        </data>
        <data name="out_msstatstmt_sum_ms" from_work_dir="MSstatsTMT_summarization_MSstats.log" format="txt" label="${tool.name} on ${on_string}: MSstatsTMT summarization MSstats">
            <filter>'msstatstmt_sum_ms' in selected_outputs</filter>
        </data>
        <data name="out_msstatstmt_group_log" from_work_dir="MSstatsTMT_log_groupComparison.log" format="txt" label="${tool.name} on ${on_string}: MSstatsTMT groupComparison log">
            <filter>'msstatstmt_group_log' in selected_outputs</filter>
        </data>
        <data name="out_msstats_gc_log" from_work_dir="MSstats_log.log" format="txt" label="${tool.name} on ${on_string}: MSstats log (groupComparison)">
            <filter>'msstats_gc_log' in selected_outputs</filter>
        </data>
        <data name="out_r_script" format="txt" label="${tool.name} on ${on_string}: Rscript">
            <filter>'r_script' in selected_outputs</filter>
        </data>
        <data name="out_quant" from_work_dir="Quant.tsv" format="tabular" label="${tool.name} on ${on_string}: Protein abundance">

        <data name="out_group_comp_plot" from_work_dir="MSstats_group_ComparisonPlot.pdf" format="pdf" label="${tool.name} on ${on_string}: Group Comparison - Comparison Plot">
            <filter>group['group_comparison'] == 'true' and 'ComparisonPlot' in group['selected_group_outputs']</filter>
        </data>
    </outputs>
    <tests>
        <test expect_num_outputs="6">
            <conditional name="input">
                <param name="input_src" value="MSstatsTMT"/>
                <param name="msstatstmt_input" ftype="tabular" value="input.msstatstmt.txt"/>
            </conditional>
            <param name="selected_outputs" value="msstatstmt_conv_log,msstatstmt_sum_ms,msstatstmt_sum_log,r_script,ProfilePlot,QCPlot"/>
            <output name="out_msstatstmt_sum_ms">
                <assert_contents>
                    <has_text text="MSstats - dataProcess function" />
                </assert_contents>
            </output>
            <output name="out_msstatstmt_sum_log">
                <assert_contents>
                    <has_text text="MSstatsTMT - proteinSummarization function" />
                </assert_contents>
            </output>
            <output name="out_r_script">
                <assert_contents>
                    <has_n_lines n="50" />
                </assert_contents>
            </output>
            <output name="out_profile_plot" file="ProfilePlot.pdf" compare="sim_size"/>
            <output name="out_qc_plot" file="QCPlot.pdf" compare="sim_size"/>
        </test>
        <test expect_num_outputs="6">
            <conditional name="input">
                <param name="input_src" value="MaxQuant"/>
                <param name="evidence" ftype="tabular" value="evidence.txt"/>
                <param name="annotation" ftype="tabular" value="annotation.txt"/>
                <param name="proteinGroups" ftype="tabular" value="proteinGroups.txt"/>

                    <param name="select" value="true"/>
                    <param name="comparison_matrix" ftype="tabular" value="comparison_matrix.txt"/>
                </conditional>
                <param name="selected_group_outputs" value="comparison_result,VolcanoPlot,ComparisonPlot"/>
            </conditional>
            <param name="selected_outputs" value="ProfilePlot,msstatstmt_group_log,msstats_gc_log"/>
            <conditional name="which_Protein">
                <param name="select" value="list"/>
                <param name="protein_list" ftype="tabular" value="proteinIDs.txt"/>
            </conditional>
            <output name="out_group_comp">

                <param name="selected_group_outputs" value="comparison_result,Heatmap"/>
            </conditional>
            <param name="selected_outputs" value="r_script"/>
            <output name="out_r_script">
                <assert_contents>
                    <has_n_lines n="61" />
                </assert_contents>
            </output>
            <output name="out_group_comp">
                <assert_contents>
                    <has_n_lines n="51" />
                    <has_n_columns n="8" />
                    <has_text text="Long_LF vs Short_HF" />
                    <has_text text="sp|O35226|PSMD4_MOUSE" />
                </assert_contents>
            </output>
            <output name="out_group_heatmap" file="MSstats_group_Heatmap.pdf" compare="sim_size"/>
        </test>

    #> 2 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 3 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 4 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 5 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #> 6 sp|Q60854|SPB6_MOUSE .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR
    #>   Charge                                             PSM Mixture
    #> 1               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 2               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 3               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 4               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3
    #> 5               3 .(TMT6plex)AFVEVNEEGTEAAAATAGMM(Oxidation)TVR_3       3