0
|
1 <tool id="determine_ipl_threshold" name="Determine IPL Threshold for Consensus Clustering" force_history_refresh="True">
|
|
2 <command interpreter="python">determine.IPL.threshold.py
|
|
3 -d ${data_fname}
|
|
4 -f ${filter_type}
|
|
5 -p ${percentage_pass}
|
|
6 -s ${selection_criteria}
|
|
7 -h $report
|
|
8 -r ${report.files_path}
|
|
9 -o ${output_rdata}
|
|
10 </command>
|
|
11 <inputs>
|
|
12 <param name="data_fname" type="data" format="tabular" label="Paradigm Results File (contains both reals and nulls)"/>
|
|
13 <param name="filter_type" type="select" label="Activity Filter" help="Specify the filter type to use (see help below)">
|
|
14 <option value="active">Active</option>
|
|
15 <option value="inactive">Inactive</option>
|
|
16 <option value="modulated" selected='true' >Modulated</option>
|
|
17 </param>
|
|
18 <param name="percentage_pass" type="float" label="% of Samples Passing (value in 0-1 range; >= 1 to indicate exact number of samples)" value="0.33" help="Specify the percentage of samples thatmust pass the threshold (see help below)"/>
|
|
19 <param name="selection_criteria" type="select" label="Selection Criteria" help="Specify the test statistic to use to select the threshold (see help below)">
|
|
20 <option value="binomial">Binomial P-value</option>
|
|
21 <option value="chisq">Chi-Squared P-value</option>
|
|
22 <option value="max_diffs" selected='true' >Overall Max Number of Differences</option>
|
|
23 </param>
|
|
24 </inputs>
|
|
25 <outputs>
|
|
26 <data format="html" name="report" label="Determine IPL Threshold (HTML)"/>
|
|
27 <data format="rdata" name="output_rdata" label="Determine IPL Threshold (RData)"/>
|
|
28 </outputs>
|
|
29 <help>
|
|
30
|
|
31 .. class:: infomark
|
|
32
|
|
33 **Determines a statistically sound threshold to use for a given IPL result**
|
|
34
|
|
35 **Parameters**
|
|
36 - **Paradigm Results File** Output from Paradigm (tab-delimited and contains both the 'real' and 'null' samples)
|
|
37
|
|
38 - **Activity Filter** Specify the filter type to use. Choice of:
|
|
39
|
|
40 * Activity - Features must exceed the user-specified threshold
|
|
41 * Inactivity - Features must fall below the user-specified threshold
|
|
42 * Modulated - Absolute value of the features must exceed the specified threshold
|
|
43
|
|
44 - **Percentage of Samples Passing** Percent of samples with an IPL that passes the threshold. Choice of:
|
|
45
|
|
46 * Real Value in [0,1] - indicate the percentage of samples that pass the threshold
|
|
47 * Integer Value - indicate the exact number of samples that pass the threshold
|
|
48
|
|
49 - **Selection Criteria** Specify the test statistic to use to select the threshold. Choice of:
|
|
50
|
|
51 * Binomial P-value - Select the threshold with the largest -log p-value (calculated as a binomial)
|
|
52 * Chi-Squared P-value - Select the threshold with the largest -log p-value (calculated as a Chi-squared)
|
|
53 * Overall Max Number of Differences - Select the threshold with the largest overall number of differences between the real and null distributions
|
|
54
|
|
55
|
|
56 </help>
|
|
57 </tool>
|