diff tools/annovar/annovar.xml @ 5:4600be69b96f draft

Added databases 1000g2015aug, SPIDEX, avsnp138, avsnp142, exac03

--- a/tools/annovar/annovar.xml	Thu Apr 10 09:31:26 2014 -0400
+++ b/tools/annovar/annovar.xml	Thu Oct 01 04:24:45 2015 -0400
@@ -1,4 +1,4 @@
-<tool id="AnnovarShed" name="ANNOVAR" version="2013aug">
+<tool id="AnnovarShed" name="ANNOVAR" version="2015may">
 	<description> Annotate a file using ANNOVAR </description>
 	
 	<requirements>		
@@ -8,6 +8,9 @@
 	<command interpreter="bash">
 		annovar.sh		
 		--esp ${esp}
+		--gonl ${gonl}
+		--exac03 ${exac03}
+		--spidex ${spidex}
 		--gerp ${gerp}
 		--cosmic61 ${cosmic61}
 		--cosmic63 ${cosmic63}	
@@ -67,8 +70,9 @@
 		
 	<inputs>
 		<param name="dorun" type="hidden" value="Y"/> <!-- will add tool in future to filter on annovar columns, then will call annovar.sh with dorun==N -->
-		<param name="reference" type="select" label="Reference">
+		<param name="reference" type="select" label="Reference">			        
 			<options from_data_table="annovar_loc" />				
+			<filter type="data_meta" ref="infile" key="dbkey" column="0"/>			
 		</param>
 				
 		<param name="infile" type="data" label="Select file to annotate" help="Must be CG varfile or a tab-separated file with a 1 line header"/>
@@ -117,7 +121,7 @@
 
 
 		<!-- filter-based annotation -->
-		<param name="verdbsnp" type="select" label="Select dbSNP version(s) to annotate with" multiple="true" display="checkboxes"  optional="true" help="SNPs in dbSNP may be flagged as Clinically Associated, Select the NonFlagged version if you do not wish to annotate with these SNPs ">			
+		<param name="verdbsnp" type="select" label="Select dbSNP version(s) to annotate with" multiple="true" display="checkboxes"  optional="true" help="avSNP are reformatted dbSNP databases with one variant per line and left-normalized indels (for a more detailed discussion read this article: http://annovar.openbioinformatics.org/en/latest/articles/dbSNP/). Flagged SNPs include SNPs less than 1% minor allele frequency (MAF) (or unknown), mapping only once to reference assembly, flagged in dbSnp as clinically associated">			
 			<option value="snp128"          > 128            (hg18/hg19) </option>
 			<option value="snp128NonFlagged"> 128 NonFlagged  </option>
 			<option value="snp129"          > 129            (hg18/hg19) </option>
@@ -134,9 +138,13 @@
 			<option value="snp137NonFlagged"> 137 NonFlagged  </option>
 			<option value="snp138"          > 138            (hg19 only) </option>				
 			<option value="snp138NonFlagged"> 138 NonFlagged  </option>				
+			<option value="avsnp138"          > 138            (avSNP, hg19 only) </option>
+			<option value="avsnp142"          > 142            (avSNP, hg19 only) </option>
 		</param>	
 
 		<param name="ver1000g" type="select" label="Select 1000Genomes Annotation(s)" multiple="true" display="checkboxes"  optional="true" help="2012april database for ALL populations was converted to hg18 using the UCSC liftover program">			
+			<option value="1000g2015aug"> 2015aug (hg19) (6 populations: AMR,AFR,EUR,EAS,SAS,ALL) </option>			
+			<option value="1000g2014oct"> 2014oct (hg18/hg19) (6 populations: AMR,AFR,EUR,EAS,SAS,ALL) </option>
 			<option value="1000g2012apr"> 2012apr (hg18/hg19) (5 populations: AMR,AFR,ASN,CEU,ALL) </option>
 			<option value="1000g2012feb"> 2012feb (hg19) (1 population: ALL) </option>
 			<option value="1000g2010nov"> 2010nov (hg19) (1 population: ALL) </option>
@@ -159,21 +167,28 @@
 			<option value="esp5400_aa"          > ESP5400   African Americans  </option>			
 		</param>	
 
+        <param name="exac03" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with ExAC 03? (The Exome Aggregation Consortium)" help=" The Exome Aggregation Consortium (ExAC) is a coalition of investigators seeking to aggregate and harmonize exome sequencing data from a wide variety of large-scale sequencing projects, and to make summary data available for the wider scientific community. The data set provided on this website spans 60,706 unrelated individuals sequenced as part of various disease-specific and population genetic studies. See http://exac.broadinstitute.org/faq for more information."/> 
 
+        <param name="gonl" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with GoNL data? (hg19 only)"/> 
+        <param name="spidex" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with SPIDEX database? (hg19 only)" help="This dataset provides machine-learning prediction on how genetic variants affect RNA splicing. (Xiong et al, Science 2015)"/> 
+
+                
 		<param name="gerp" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="GERP++ Annotation?" help="GERP identifies constrained elements in multiple alignments by quantifying substitution deficits (see http://mendel.stanford.edu/SidowLab/downloads/gerp/ for details) This option annotates those variants having GERP++>2 in human genome, as this threshold is typically regarded as evolutionarily conserved and potentially functional"/>
 	
 		<param name="clinvar" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="CLINVAR Annotation? (hg19 only)" help="version 2014-02-11. Annotations include Variant Clinical Significance (unknown, untested, non-pathogenic, probable-non-pathogenic, probable-pathogenic, pathogenic, drug-response, histocompatibility, other) and Variant disease name."/>
 		<param name="nci60" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with NCI60? (hg19 only)" help="NCI-60 exome allele frequency data"/>
 		<param name="cgfortysix" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Complete Genomics 46 Genomes?" help="Diversity Panel; 46 unrelated individuals"/>
 		<param name="cgsixtynine" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Complete Genomics 69 Genomes?" help="Diversity Panel, Pedigree, YRI trio and PUR trio"/>
-		<param name="cosmic61" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC61? (hg19 only)"/>
-		<param name="cosmic63" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC63? (hg19 only)"/>
-		<param name="cosmic64" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC64? (hg19 only)"/>
-		<param name="cosmic65" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC65? (hg19 only)"/>
-		<param name="cosmic67" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC67? (hg19 only)"/>
-		<param name="cosmic68" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC68? (hg19 only)"/>
+		<param name="cosmic61" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 61? (hg19 only)"/>
+		<param name="cosmic63" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 63? (hg19 only)"/>
+		<param name="cosmic64" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 64? (hg19 only)"/>
+		<param name="cosmic65" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 65? (hg19 only)"/>
+		<param name="cosmic67" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 67? (hg19 only)"/>
+		<param name="cosmic68" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 68? (hg19 only)"/>
+		<param name="cosmic70" type="boolean" checked="False" truevalue="Y" falsevalue="N" label="Annotate with COSMIC 70? (hg19 only)"/>
+		
+       
 
-		
 		<param name="newimpactscores" type="select" label="Select functional impact scores (LJB2)" multiple="true" display="checkboxes" optional="true" help="LJB refers to Liu, Jian, Boerwinkle paper in Human Mutation, pubmed ID 21520341. ">						
 			<option value="ljb2_sift"> SIFT score </option>
 			<option value="ljb2_pp2hdiv"> PolyPhen2 HDIV score </option>