diff falco-call.xml @ 39:a793ffd0cd04 draft

Uploaded

--- a/falco-call.xml	Fri Oct 10 11:28:57 2014 -0400
+++ b/falco-call.xml	Tue Dec 23 11:26:33 2014 -0500
@@ -6,48 +6,16 @@
     <requirement type="package" version="3.1.0">R</requirement>
     <requirement type="package" version="2">bcftools</requirement> 
     <!--
-    <requirement type="package" version="3.4">snpeff</requirement>
-    <requirement type="package" version="2">varscan</requirement> 
-    
+    <requirement type="set_environment">TOOL_DIR</requirement>
     -->
-    <!--
-    
-    <requirement type="set_environment">JAVA_JAR_PATH</requirement>
-    <requirement type="set_environment">SNPEFF_JAR_PATH</requirement>
-    
-    <requirement type="set_environment">JAR_PATH1</requirement>
-    <requirement type="set_environment">JAR_PATH2</requirement>
-    -->
-    <requirement type="set_environment">TOOL_DIR</requirement>
     
   </requirements>
 
   <description>FALCO: Amplicon Analysis Pipeline</description>
-  
-  <!-- command block to call script with specific R installation (set full paths) -->
-  <!--
-  <command>
-    /full/path/to/Rscript 
-    /full/path/to/QDNAseq.R 
-    $qdnaseq_cfg
-  </command>
-  -->
 
-  <!-- command block to call script with default R environment -->
   <command interpreter="bash">
-    <!--falco.sh $bam $bam.name $html_output $vcf_output ${html_output.files_path} &amp;&amp;; echo "ECHO PRINT"; -->
-    <!-- falco.sh $falco_cfg \$JAVA_JAR_PATH \$SNPEFF_JAR_PATH; -->
     falco-call.sh $falco_cfg;
   </command>
-  <!--
-  <command interpreter="bash">
-    falco/amplicon-VcAn-TSACP.sh $bam $bam.name;
-    echo "DONE"
-  </command>
-  <command interpreter="perl">
-    falco.pl $htmlFile ${htmlFile.files_path} $bam $bam.name;
-  </command>
-  -->
 
   <stdio>
     <!-- Anything higher than 0 means the bash script didnt finish -->
@@ -58,14 +26,31 @@
     <!-- ==================== -->
     <!-- General inputs -->
     <!-- ==================== -->
+    <conditional name="genomeSource">
+      <param name="refGenomeSource" type="select" label="You can select a reference genome from your history or use a built-in index">
+        <option value="indexed">Use a built-in index</option>
+        <option value="history">Use one from history</option>
+      </param>
+      <when value="indexed">
+        <param name="indices" type="select" label="Select a reference genome">
+          <options from_data_table="all_fasta">
+            <filter type="sort_by" column="2" />
+            <validator type="no_options" message="No indexes are available in table" />
+          </options>
+        </param>
+      </when>
+      <when value="history">
+        <param name="ownFile" type="data" format="fasta" metadata_name="dbkey" label="Select a reference from history" />
+      </when>
+    </conditional>
+
     <param name="jobName" type="text" optional="false" label="Analysis/ouput name" help="Supply a name for the outputs to remind you what they contain" value="TEST">
       <validator type="empty_field" />
       <validator type="regex" message="This field may contain only non-whitespace characters">\S+</validator>
     </param>
     <param name="bam" type="data" multiple="false" optional="false" format="bam" label="Input BAM" help="Select BAM file" />
 
-    <param format="fasta" name="reference" type="data" metadata_name="dbkey" label="Reference File"/>
-    <param format="bed" name="design_file" optional="True" type="data" label="BED File" help="When left empty, default Cancer Panel is used"/>
+    <param format="tsv" name="filter_file" optional="True" type="data" label="FILTER File" help="When left empty, default list of genomic positions is used"/>
     <param format="txt" name="manifest_file" optional="True" type="data" label="MANIFEST File" help="When left empty, default Cancer Panel is used"/>
 
     <!-- ==================== -->
@@ -79,27 +64,32 @@
   <!-- ==================== -->
   <configfiles>
     <configfile name="falco_cfg">
-      design_file=$design_file
+      filter_file=$filter_file
       manifest_file=$manifest_file
       bam_file=$bam
       bam_name=$bam.name
       html_out=$html_output
       vcf_out=$vcf_output
-      out_path=${vcf_output.files_path}
+      out_path=${html_output.files_path}
+      ## reference source
+      REF_SOURCE="${genomeSource.refGenomeSource}"
+      #if $genomeSource.refGenomeSource == "history":
+        ##build index on the fly
+        REF_FILE="${genomeSource.ownFile}"
+        DB_KEY="${dbkey}"
+      #else:
+        ##use precomputed indexes
+        REF_FILE="${genomeSource.indices.fields.path}"
+      #end if
     </configfile>
   </configfiles>
-      <!--
-      jar_path1=$JAR_PATH1
-      jar_path2=$JAR_PATH2
-      jar_path3=$JAVA_JAR_PATH
-      jar_path4=$SNPEFF_JAR_PATH
-    -->
+
   <!-- ==================== -->
   <!-- Main output is an html based report, additional on request -->
   <!-- ==================== -->
   <outputs>
-    <data format="html" name="html_output" label="FALCO AAP (${jobName})" />
-    <data format="vcf" name="vcf_output" label="FALCO AAP (${jobName}): VCF" />
+    <data format="html" name="html_output" label="FALCO-calling (${jobName})" />
+    <data format="vcf" name="vcf_output" label="FALCO-calling (${jobName}): VCF" />
   </outputs>
 
   <!-- ==================== -->
@@ -120,17 +110,23 @@
 
 **Introduction**
 
-This tool is a wrapper for the Amplicon Analysis Pipeline (AAP) FALCO_
+This tool is a wrapper for the variant caller FALCO_ which is part of the Amplicon Analysis Pipeline (AAP)
 
-.. _FALCO: https://bitbucket.org/daoud/falco
+.. _FALCO: https://github.com/tgac-vumc/falco/
 
 Calls and annotates genomic variants for each amplicon in a design.
 
+**Notes**
+
+Because each amplicon is considered separate test, if two amplicons overlap this can cause the same variant position to be present twice in the output VCF. This is intentional, you can use this to evaluate the quality of the variant call beyond the amplicon.
+
 -----
 
 **Citation**
 
-For the underlying tool please cite: Daoud Sie et al. Performance of amplicon-based next generation DNA sequencing for diagnostic gene mutation profiling in oncopathology (Cell Oncol 2014 Oct;37(5):353-61).
+For the underlying tool please cite: Daoud Sie et al. Performance of amplicon-based next generation DNA sequencing for diagnostic gene mutation profiling in oncopathology (Cell Oncol 2014 Oct;37(5):353-61). [Pubmed]
+
+.. _Pubmed: http://www.ncbi.nlm.nih.gov/pubmed/25209392
 
   </help>