# HG changeset patch
# User iuc
# Date 1430362047 14400
# Node ID 26890eb5a1f11b68f9b923a4dfe8935c58006077
# Parent  cfd163368d3b097c94464bbec0288a294e20d54b
Uploaded

diff -r cfd163368d3b -r 26890eb5a1f1 snpSift_geneSets.xml
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/snpSift_geneSets.xml	Wed Apr 29 22:47:27 2015 -0400
@@ -0,0 +1,67 @@
+<tool id="snpSift_geneSets" name="SnpSift GeneSets" version="4.0.0">
+    <description>Annotating GeneSets, such as Gene Ontology, KEGG, Reactome</description>
+    <!-- 
+        You can change the amount of memory used, just change the -Xmx parameter (e.g. use -Xmx2G for 2Gb of memory)
+    -->
+    <expand macro="requirements" />
+    <macros>
+        <import>snpSift_macros.xml</import>
+    </macros>
+    <command>
+        java -Xmx2G -jar \$SNPEFF_JAR_PATH/SnpSift.jar geneSets -v
+        #if $db_opts.db_opts_selector == "db"
+          "${db_opts.database.fields.path}"
+        #elif $db_opts.db_opts_selector == "histdb"
+          "$db_opts.histdb"
+        #end if
+        
+        $input 2&gt; $log &gt; $output
+    </command>
+    <inputs>
+        <param format="vcf" name="input" type="data" label="Variant input file in VCF format"/>
+        <conditional name="db_opts">
+            <param name="db_opts_selector" type="select" label="Select Annotation database" help="">
+              <option value="db" selected="True">Locally installed database</option>
+              <option value="histdb">database from your history</option>
+            </param>
+            <when value="db">
+                <param name="database" type="select" label="Molecular Signatures Database (MSigDB)">
+                    <options from_data_table="snpeff_msigdb_database" />
+                </param>
+                <param name="histdb" type="hidden" value="" />
+            </when>
+            <when value="histdb">
+                <param name="histdb" type="data" format="txt" label="Molecular Signatures Database (MSigDB)" />
+            </when>
+        </conditional>
+    </inputs>
+    <outputs>
+        <data format="vcf" name="output" label="${tool.name} on ${on_string}: VCF" />
+        <data format="txt" name="log" label="${tool.name} on ${on_string}: log" />
+    </outputs>
+    <expand macro="stdio" />
+    <tests>
+    </tests>
+    <help>
+This tool uses `SnpSift GeneSets`_ to add annotations from `MSigDB`_, a collection of annotated gene sets from different sources including Gene Ontology (GO), KEGG, Reactome.
+
+.. _SnpSift GeneSets: http://snpeff.sourceforge.net/SnpSift.html#geneSets
+
+.. class:: warningmark
+
+The input VCF file must be annotated using SnpEff before performing GeneSets annotations. This is because the tool must know which gene the variant affects.
+
+@EXTERNAL_DOCUMENTATION@
+
+@CITATION_SECTION@
+
+For `MSigDB`_, please cite |Subramanian2005|_.
+
+.. _MSigDB: http://www.broadinstitute.org/gsea/msigdb/
+.. |Subramanian2005| replace:: Subramanian, A., *et al.* (2005) Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. *Proc. Natl. Acad. Sci. U.S.A.* 102(43), 15545-15550
+.. _Subramanian2005: http://www.pnas.org/content/102/43/15545
+    </help>
+    <expand macro="citations">
+        <citation type="doi">10.1073/pnas.0506580102</citation><!-- MSigDB citation -->
+    </expand>
+</tool>
diff -r cfd163368d3b -r 26890eb5a1f1 tool-data/snpeff_msigdb_database.loc.sample
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/tool-data/snpeff_msigdb_database.loc.sample	Wed Apr 29 22:47:27 2015 -0400
@@ -0,0 +1,4 @@
+## Molecular Signatures Database (MSigDB) for SnpEff.
+## You can download MSigDb from http://www.broadinstitute.org/gsea/msigdb
+#msigdb_id	msigdb_name path_to_database
+#msigdb_08_2014 MSigDB 08-2014	/data/msigdb/08_2014/msigdb.gmt