From 47463daa73bb1b304f2504ab052e5e4ef94f0d0e Mon Sep 17 00:00:00 2001
From: Greg Von Kuster <greg@bx.psu.edu>
Date: Tue, 23 Mar 2010 15:43:58 -0400
Subject: [PATCH] Add the add_scores tool.

---
 tool-data/add_scores.loc.sample | 21 ++++++++++++
 tool_conf.xml.sample            |  1 +
 tools/evolution/add_scores.xml  | 60 +++++++++++++++++++++++++++++++++
 3 files changed, 82 insertions(+)
 create mode 100644 tool-data/add_scores.loc.sample
 create mode 100644 tools/evolution/add_scores.xml

diff --git a/tool-data/add_scores.loc.sample b/tool-data/add_scores.loc.sample
new file mode 100644
index 00000000000..b50a0dcbcaf
--- /dev/null
+++ b/tool-data/add_scores.loc.sample
@@ -0,0 +1,21 @@
+#This is a sample file distributed with Galaxy that enables tools
+#to use a directory of gzipped genome files for use with add_scores.  You will 
+#need to supply these files and then create a add_scores.loc file 
+#similar to this one (store it in this directory) that points to 
+#the directories in which those files are stored. The add_scores.loc 
+#file has this format (white space characters are TAB characters):
+#
+#<build>	<file_path>
+#
+#So, for example, if your add_scores.loc began like this:
+#
+#hg18	/afs/bx.psu.edu/depot/data/genome/hg18/misc/phyloP/
+#
+#then your /afs/bx.psu.edu/depot/data/genome/hg18/misc/phyloP/ directory
+#would need to contain the following gzipped files, among others:
+#
+#-rw-r--r-- 1 rico rico 161981190 2010-03-19 12:48 chr10.phyloP44way.primate.wigFix.gz
+#-rw-r--r-- 1 rico rico     54091 2010-03-19 12:56 chr10_random.phyloP44way.primate.wigFix.gz
+#-rw-r--r-- 1 rico rico 158621990 2010-03-19 12:46 chr11.phyloP44way.primate.wigFix.gz
+#
+hg18	/galaxy/data/hg18/misc/phyloP
diff --git a/tool_conf.xml.sample b/tool_conf.xml.sample
index 09db1b56711..a5741bb6867 100644
--- a/tool_conf.xml.sample
+++ b/tool_conf.xml.sample
@@ -162,6 +162,7 @@
     <tool file="hyphy/hyphy_dnds_wrapper.xml" />
     <tool file="evolution/mutate_snp_codon.xml" />
     <tool file="evolution/codingSnps.xml" />
+    <tool file="evolution/add_scores.xml" />
  </section>
  <section name="Metagenomic analyses" id="tax_manipulation">
     <tool file="taxonomy/gi2taxonomy.xml" />
diff --git a/tools/evolution/add_scores.xml b/tools/evolution/add_scores.xml
new file mode 100644
index 00000000000..84475bd156a
--- /dev/null
+++ b/tools/evolution/add_scores.xml
@@ -0,0 +1,60 @@
+<tool id="add_scores" name="Add scores">
+  <description>for interspecies conservation at each SNPs</description>
+  <command>
+    add_scores $input1 ${input1.metadata.dbkey} ${input1.metadata.chromCol} ${input1.metadata.startCol} ${GALAXY_DATA_INDEX_DIR}/add_scores.loc $out_file1
+  </command>
+  <inputs>
+    <param format="interval" name="input1" type="data" label="SNPs"/>
+  </inputs>
+  <outputs>
+    <data format="input" name="out_file1" />
+  </outputs>
+  <tests>
+    <test>
+	<param name="input1" value="add_scores_input1.interval" dbkey="hg18" />
+	<output name="output" file="add_scores_output1.interval" />
+    </test>
+    <test>
+	<param name="input1" value="add_scores_input2.bed" dbkey="hg18" />
+	<output name="output" file="add_scores_output2.interval" />
+    </test>
+  </tests>
+
+  <help>
+This tool adds a column that measures interspecies conservation at each SNP position, using conservation scores for primates computed by the phyloP program. It currently works only for hg18.
+
+**Example**
+
+- input file, with SNPs::
+
+    chr22	16440426	14440427	C/T
+    chr22	15494851	14494852	A/G
+    chr22	14494911	14494912	A/T
+    chr22	14550435	14550436	A/G
+    chr22	14611956	14611957	G/T
+    chr22	14612076	14612077	A/G
+    chr22	14668537	14668538	C
+    chr22	14668703	14668704	A/T
+    chr22	14668775	14668776	G
+    chr22	14680074	14680075	A/T
+    etc.
+
+- output file, showing non-synonymous substitutions in coding regions::
+
+    chr22	16440426	14440427	C/T	0.509
+    chr22	15494851	14494852	A/G	0.427
+    chr22	14494911	14494912	A/T	NA
+    chr22	14550435	14550436	A/G	NA
+    chr22	14611956	14611957	G/T	-2.142
+    chr22	14612076	14612077	A/G	0.369
+    chr22	14668537	14668538	C	0.419
+    chr22	14668703	14668704	A/T	-1.462
+    chr22	14668775	14668776	G	0.470
+    chr22	14680074	14680075	A/T	0.000
+    chr22       14680074        14680075        A/T	0.303
+    etc.
+
+"NA", means that the phyloP score was not available.
+
+</help>
+</tool>