--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/GALAXY_FILES/tools/EMBER/Integrate_Data.xml	Wed Feb 29 15:03:33 2012 -0500
@@ -0,0 +1,50 @@
+<tool id="integrate_data" name="Integrate Data" version="1.3">
+  <description>Assigns potential targets to binding sites</description>
+  <command interpreter="perl">Integrate_Data.pl -b $binding_data -e $expression_data -o $output -d $dist -dt $dtype -v n</command>
+  <inputs>
+    <param format="txt" name="binding_data" type="data" label="Binding data"/>
+    <param format="txt" name="expression_data" type="data" label="Discretized expression data"/>
+    <param name="dist" type="integer" min="0" label="Max distance (kbp)" value="100" optional="true"/>
+    <param name="dtype" type="select" label="Distance type">
+        <option value="1" selected="true">To gene boundaries</option>
+        <option value="2">To TSS</option>
+    </param>
+  </inputs>
+  <outputs>
+    <data format="txt" name="output"/>
+  </outputs>
+
+  <tests>
+    <test>
+      <param name="binding_data" value="EMBER/peaks.txt"/>
+      <param name="expression_data" value="EMBER/expression_profiles.txt"/>
+      <param name="dist" value="100"/>
+      <param name="dtype" value="1"/>
+      <output name="output" file="EMBER/integrated.txt"/>
+    </test>
+  </tests>
+
+  <help>
+
+This tool combines binding data with annotated gene/probe sets to assign potential targets to each binding site.
+
+-----
+
+Description of inputs:
+
+*Binding Data*:
+
+   Binding data in bed-like format (note only the first coordinate after [chr] is used, so if you have the regular bed format, you may want to add a new second column with the average of the start and end coordinates). [other information] is retained throughout the analysis, and may contain peak ID, peak enrichment, etc.
+
+   *Format (at least 2 columns)*: [chr] [peak posn] [other information]
+
+*Discretized Expression Data*: output of PreProcess Expression Data.
+
+*Max Distance*: maximum distance from a peak in order to consider a gene a potential target (in kbp).
+
+*Distance Type*: definition of distance (to gene boundaries or to TSS). If "To gene boundaries" is chosen, peaks lying within a gene's coordinates have a distance of 0.
+
+  </help>
+
+</tool>
+