comparison cluster.tools/ipl.feature.selection.xml @ 0:0decf3fd54bc draft

Uploaded

0:0decf3fd54bc

<tool id="ipl_feature_selection" name="IPL Feature Selection" force_history_refresh="True">
    <command interpreter="python">ipl.feature.selection.py
-d $dataset 
${genes_only}
-o ${output}

#if str($thresh_format.format) == 'manual':
-f ${thresh_format.filter} 
-t ${thresh_format.threshold} 
-p ${thresh_format.perc_pass} 
#end if
#if str($thresh_format.format) == 'empirical':
-e ${empirical_fname} 
#end if

</command>
    <inputs>
    	<param name="dataset" type="data" format='tabular' label="Data Set"/>
    	<param name="genes_only" type="boolean" label="Genes Only (check to set yes)" truevalue="-g" falsevalue="" checked="False" />
    	
	<conditional name="thresh_format" >
	  <param name="format" type='select' label="Input Format for Threshold Used" >
	    <option value="empirical" selected='true'>Output from the Determine Threshold Tool</option>
	    <option value="manual">Manually Specify the Threshold</option>
	  </param>
	  <when value="manual" >
	    <param name="filter" type="select" label="Activity Filter" >
	      <option value="modulated" select='true' >Modulated</option>
	      <option value="active">Active</option>
	      <option value="inactive">Inactive</option>
	    </param>
	    <param name="perc_pass" type="float" label="% of Samples Passing (value in 0-1 range; >= 1 to indicate exact number of samples)" value="0.33"/>
	    <param name="threshold" type="float" label="Activity Threshold" value="0.25"/>
	  </when>
	  <when value="empirical" >
	    <param name="empirical_fname" type="data" format='rdata' label="Activity Threshold File" />
	  </when>
	</conditional>
    </inputs>
    <outputs>
        <data format="tabular" name="output" label="Filtered IPLs"/>
    </outputs>
<help>
.. class:: infomark
     
**IPL Feature Selection** - Tool to filter an IPL matrix to contain only those features that exceed a given threshold for a specified percentage of samples

**OUTPUT:**  A new matrix containing only the feaures that pass the user-specified filter

----

**Parameters**

- **Genes Only** Check to limit the new matrix to only gene features

- **Input Format for Threshold Used:** - Specify the format to specify the Threshold.  Choice of:
         * Output from the 'Determine Threshold Tool'
         * Manually Specify the Threshold

- **IF Output from the 'Determine Threshold Tool' IS SELECTED:**
         * **Activity Threshold File** Specify the result file from the 'Determine IPL Threshold for Consensus Clustering" tool

- **IF Manually Specify the Threshold IS SELECTED:**
         * **Activity Filter** Specify the filter type to use.  Choice of:

                 * Activity -  Features must exceed the user-specified threshold
                 * Inactivity -  Features must fall below the user-specified threshold
                 * Modulated - Absolute value of the features must exceed the specified threshold

         * **Percentage of Samples Passing** Percent of samples with an IPL that passes the threshold. Choice of:

                 * Real Value in [0,1] - indicate the percentage of samples that pass the threshold
                 * Integer Value       - indicate the exact number of samples that pass the threshold

         * **Selection Criteria** Specify the test statistic to use to select the threshold.  Choice of:

                 * Binomial P-value - Select the threshold with the largest -log p-value (calculated as a binomial)
                 * Chi-Squared P-value - Select the threshold with the largest -log p-value (calculated as a Chi-squared)
                 * Overall Max Number of Differences - Select the threshold with the largest overall number of differences between the real and null distributions


</help>
</tool>