0
|
1 <tool id="ipl_feature_selection" name="IPL Feature Selection" force_history_refresh="True">
|
|
2 <command interpreter="python">ipl.feature.selection.py
|
|
3 -d $dataset
|
|
4 ${genes_only}
|
|
5 -o ${output}
|
|
6
|
|
7 #if str($thresh_format.format) == 'manual':
|
|
8 -f ${thresh_format.filter}
|
|
9 -t ${thresh_format.threshold}
|
|
10 -p ${thresh_format.perc_pass}
|
|
11 #end if
|
|
12 #if str($thresh_format.format) == 'empirical':
|
|
13 -e ${empirical_fname}
|
|
14 #end if
|
|
15
|
|
16 </command>
|
|
17 <inputs>
|
1
|
18 <param name="dataset" type="data" format='tabular' label="IPL Data Set"/>
|
0
|
19 <param name="genes_only" type="boolean" label="Genes Only (check to set yes)" truevalue="-g" falsevalue="" checked="False" />
|
|
20
|
|
21 <conditional name="thresh_format" >
|
|
22 <param name="format" type='select' label="Input Format for Threshold Used" >
|
|
23 <option value="empirical" selected='true'>Output from the Determine Threshold Tool</option>
|
|
24 <option value="manual">Manually Specify the Threshold</option>
|
|
25 </param>
|
|
26 <when value="manual" >
|
|
27 <param name="filter" type="select" label="Activity Filter" >
|
|
28 <option value="modulated" select='true' >Modulated</option>
|
|
29 <option value="active">Active</option>
|
|
30 <option value="inactive">Inactive</option>
|
|
31 </param>
|
|
32 <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"/>
|
|
33 <param name="threshold" type="float" label="Activity Threshold" value="0.25"/>
|
|
34 </when>
|
|
35 <when value="empirical" >
|
|
36 <param name="empirical_fname" type="data" format='rdata' label="Activity Threshold File" />
|
|
37 </when>
|
|
38 </conditional>
|
|
39 </inputs>
|
|
40 <outputs>
|
|
41 <data format="tabular" name="output" label="Filtered IPLs"/>
|
|
42 </outputs>
|
|
43 <help>
|
|
44 .. class:: infomark
|
|
45
|
|
46 **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
|
|
47
|
|
48 **OUTPUT:** A new matrix containing only the feaures that pass the user-specified filter
|
|
49
|
|
50 ----
|
|
51
|
|
52 **Parameters**
|
|
53
|
|
54 - **Genes Only** Check to limit the new matrix to only gene features
|
|
55
|
|
56 - **Input Format for Threshold Used:** - Specify the format to specify the Threshold. Choice of:
|
|
57 * Output from the 'Determine Threshold Tool'
|
|
58 * Manually Specify the Threshold
|
|
59
|
|
60 - **IF Output from the 'Determine Threshold Tool' IS SELECTED:**
|
|
61 * **Activity Threshold File** Specify the result file from the 'Determine IPL Threshold for Consensus Clustering" tool
|
|
62
|
|
63 - **IF Manually Specify the Threshold IS SELECTED:**
|
|
64 * **Activity Filter** Specify the filter type to use. Choice of:
|
|
65
|
|
66 * Activity - Features must exceed the user-specified threshold
|
|
67 * Inactivity - Features must fall below the user-specified threshold
|
|
68 * Modulated - Absolute value of the features must exceed the specified threshold
|
|
69
|
|
70 * **Percentage of Samples Passing** Percent of samples with an IPL that passes the threshold. Choice of:
|
|
71
|
|
72 * Real Value in [0,1] - indicate the percentage of samples that pass the threshold
|
|
73 * Integer Value - indicate the exact number of samples that pass the threshold
|
|
74
|
|
75 * **Selection Criteria** Specify the test statistic to use to select the threshold. Choice of:
|
|
76
|
|
77 * Binomial P-value - Select the threshold with the largest -log p-value (calculated as a binomial)
|
|
78 * Chi-Squared P-value - Select the threshold with the largest -log p-value (calculated as a Chi-squared)
|
|
79 * Overall Max Number of Differences - Select the threshold with the largest overall number of differences between the real and null distributions
|
|
80
|
|
81
|
|
82 </help>
|
|
83 </tool>
|