<tool id="netboxr" name="netboxr" version="1.6.0">
    <description>enables automated discovery of biological process modules by network analysis.</description>
    <requirements>
        <requirement type="package" version="1.6.0">bioconductor-netboxr</requirement>
    </requirements>
    <command detect_errors="exit_code"><![CDATA[
Rscript '$__tool_directory__/netboxr_r.R'
--geneList '$geneList'
--cutoff '$cutoff'
--community $community
--globalModel $globalModel
--globalIterations '$globalIterations'
--globalNumber '$globalNumber'
--localModel $localModel
--localIterations '$localIterations'
--networkPlot $networkPlot
--plotWidth '$plotWidth'
--outputSIF $outputSIF
--neighborList $neighborList
--modmem $modmem
--nt $nt
	]]></command>
    <inputs>
        <param argument="--geneList" type="data" format="txt" label="Gene list" help="A newline-delimited text file that contains a list of genes of interest" />
        <param argument="--cutoff" type="float" value="0.05" label="P-Value Cut-Off" help="The corrected p-value (or q-value) cut-off to apply for the graph output. Default: 0.05" />
        <param argument="--community" type="select" label="Community detection method" help="Community detection methods include edge betweenness score (EBC), leading eigenvector method (LEC), Louvain method (LOUV), or Leiden method (LEID). Default: EBC">
            <option value="ebc">EBC</option>
            <option value="lec">LEC</option>
            <option value="louvain">LOUV</option>
            <option value="leiden">LEID</option>
        </param>
        <param argument="--globalModel" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Global network null model" help="Perform global network null model analysis. The global network null model calculates the empirical p-value as the number of times (over a set of iterations) the size of the largest connected component (the giant component) in the network coming from the same number of randomly selected genes equals or exceeds the size of the largest connected component in the observed network. Default: Yes" />
        <param argument="--globalIterations" type="integer" value="10" label="Global network null model iterations" help="Global network null model iterations. Recommended: 1000" />
        <param argument="--globalNumber" type="integer" value="274" label="Global network model number of genes" help="Global network null model number of genes." />
        <param argument="--localModel" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Local network null model" help="Perform local network null model analysis. The local network null model evaluates the deviation of modularity in the observed network from modularity distribution in the random network. Default: Yes" />
        <param argument="--localIterations" type="integer" value="10" label="Local network null model iterations" help="Local network null model iterations. Recommended: 1000" />
        <param argument="--networkPlot" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Network plot" help="Output the annotated graph of netboxr output. Default: Yes" />
        <param argument="--plotWidth" type="integer" value="8" label="Network plot width" help="Plot width. Recommended: 9" />
        <param argument="--outputSIF" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Network in SIF format" help="Output the network in SIF format. Default: Yes" />
        <param argument="--neighborList" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Neighbor list" help="Output a table containing the information of all neighbor nodes. Default: Yes" />
        <param argument="--modmem" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Module membership" help="Output a table containing the identified pathway module numbers. Default: Yes" />
        <param argument="--nt" type="boolean" truevalue="True" falsevalue="False" checked="True" label="Node type" help="Output a table containing the information of node types. Default: Yes" />
    </inputs>
    <outputs>
        <data name="output_pdf" format="pdf" from_work_dir="network_plot.pdf" label="${tool.name} on ${on_string}: Network plot">
            <filter>output_pdf is True</filter>
        </data>
        <data name="localModel_pdf" format="pdf" from_work_dir="localModel_histogram.pdf" label="${tool.name} on ${on_string}: Local model histogram">
            <filter>localModel is True</filter>
        </data>
        <data name="output_sif" format="sif" from_work_dir="network.sif" label="${tool.name} on ${on_string}: SIF network">
            <filter>outputSIF is True</filter>
        </data>
        <data name="output_neighbor" format="tabular" from_work_dir="neighbor_data.txt" label="${tool.name} on ${on_string}: Neighbor data">
            <filter>output_neighbor is True</filter>
        </data>
        <data name="moduleMembership" format="tabular" from_work_dir="community.membership.txt" label="${tool.name} on ${on_string}: Module membership">
            <filter>moduleMembership is True</filter>
        </data>
        <data name="nodeType" format="tabular" from_work_dir="nodeType.txt" label="${tool.name} on ${on_string}: Node type">
            <filter>nodeType is True</filter>
        </data>
        <data name="metadata" format="txt" from_work_dir="metadata.txt" label="${tool.name} on ${on_string}: Metadata" />
    </outputs>
    <tests>
        <test>
            <param name="geneList" value="netbox2010_data.txt" />
            <param name="cutoff" value="0.05" />
            <param name="community" value="ebc" />
            <param name="globalModel" value="True" />
            <param name="globalIterations" value="10" />
            <param name="globalNumber" value="274" />
            <param name="localModel" value="True" />
            <param name="localIterations" value="10" />
            <param name="networkPlot" value="True" />
            <param name="plotWidth" value="8" />
            <param name="outputSIF" value="True" />
            <param name="neighborList" value="True" />
            <param name="modmem" value="True" />
            <param name="nt" value="True" />
            <output name="output_pdf" file="network_plot.pdf" compare="sim_size" />
            <output name="localModel_pdf" file="localModel_histogram.pdf" compare="sim_size" />
            <output name="output_sif" file="network.sif">
                <assert_contents>
                    <has_text_matching expression="CBL\tPP\tPIK3R1" />
                    <has_text_matching expression="PIK3R1\tIN_SAME_COMPONENT\tAGAP2" />
                </assert_contents>
            </output>
            <output name="output_neighbor" file="neighbor_data.txt">
                <assert_contents>
                    <has_text_matching expression="idx\tname\tlocalDegree\tglobalDegree\tpValueRaw\toddsRatio\tpValueFDR" />
                    <has_text_matching expression="1712\tCRK\t11\t81\t2.39208791593171e-05\t1.70873218314362\t0.0186673141177578" />
                    <has_text_matching expression="6187\tRPS27\t2\t186\t0.975923212389053\t-0.825808759394024\t0.975923212389053" />
                </assert_contents>
            </output>
            <output name="moduleMembership" file="community.membership.txt">
                <assert_contents>
                    <has_text_matching expression="CBL\t1" />
                    <has_text_matching expression="CCT6A\t10" />
                </assert_contents>
            </output>
            <output name="nodeType" file="nodeType.txt">
                <assert_contents>
                    <has_text_matching expression="CDKN2B\tcandidate" />
                    <has_text_matching expression="PTPN11\tlinker" />
                </assert_contents>
            </output>
        </test>
    </tests>
    <help><![CDATA[

NetBox is a network-based approach that combines prior knowledge with a
network clustering algorithm. The algorithm allows for the identification
of functional modules and allows for combining multiple data types, such as
mutations and copy number alterations. NetBox performs network analysis on
human interaction networks, and comes pre-loaded with a Human Interaction
Network (HIN) derived from four literature curated data sources, including
the Human Protein Reference Database (HPRD), Reactome, NCI-Nature Pathway
Interaction (PID) Database, and the MSKCC Cancer Cell Map.

-----

**INPUTS**

**Gene list**
A newline-delimited text file that contains a list of genes of interest is
required.

Example of text file contents:

EGFR
TP53
ACTB
GAPDH

**Cutoff value**
The corrected p-value (or q-value) cut-off to apply for the graph output.
The default p-value is 0.05.


**OUTPUTS**

* A PDF with a plot of the edge annotated graph (optional)

* A text file with tabular data containing information of all neighbor nodes (optional)

  	]]></help>
    <citations>
        <citation type="doi">10.1371/journal.pone.0234669</citation>
    </citations>
</tool>