Mercurial > repos > iuc > gatk2

changeset 2:28b85e2409ce draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author bgruening
date Sat, 30 Nov 2013 10:36:47 -0500
parents c8752e31f496
children b7ef18eaa9c9
files base_recalibrator.xml indel_realigner.xml print_reads.xml realigner_target_creator.xml tool_dependencies.xml
diffstat 5 files changed, 47 insertions(+), 47 deletions(-) [+]
line wrap: on
line diff
--- a/base_recalibrator.xml	Tue Nov 26 06:10:20 2013 -0500
+++ b/base_recalibrator.xml	Sat Nov 30 10:36:47 2013 -0500
@@ -1,5 +1,5 @@
 <tool id="gatk2_base_recalibrator" name="Base Recalibrator" version="0.0.7">
-  <description>on BAM files</description>
+  <description>calculates covariates used to recalibrate base quality scores of reads</description>
   <expand macro="requirements" />
   <macros>
     <import>gatk2_macros.xml</import>
@@ -20,6 +20,9 @@
     --num_cpu_threads_per_data_thread 8
 
     @THREADS@
+    ## we set non standards at every run and the user can choose which ones are preferred
+    ## in our select box both standard options (ContextCovariate, CycleCovariate) are selected by default
+    --no_standard_covs
 
     #if $reference_source.reference_source_selector != "history":
         -R "${reference_source.ref_file.fields.path}"
@@ -28,15 +31,11 @@
         --BQSR "${input_recal}"
     #end if
     --out "${output_recal}"
-    ${standard_covs}
     #if str( $covariates ) != "None":
         #for $cov in str( $covariates ).split( ',' ):
             -cov "${cov}"
         #end for
     #end if
-    #if $plot_pdf:
-       --plot_pdf_file $output_pdf
-    #end if
    '
     
     #set $snp_dataset_provided = False
@@ -118,26 +117,25 @@
             Please be aware that one should only run recalibration with the covariates file created on the same input bam(s).
       </help>
     </param>
-    <param name="standard_covs" type="boolean" truevalue="" falsevalue="--no_standard_covs" checked="false" label="Do not use the standard set of covariates in addition to the ones selected" help="--no_standard_covs/-noStandard" />
+
     <param name="covariates" type="select" multiple="True" display="checkboxes" label="Covariates to be used in the recalibration" help="-cov,--covariate &amp;lt;covariate&amp;gt;" >
       <!-- might we want to load the available covariates from an external configuration file, since additional ones can be added to local installs? -->
-      <option value="ReadGroupCovariate" />
+      <option value="ContextCovariate" selected="true"/>
+      <option value="CycleCovariate" selected="true"/>
+      <option value="RepeatLengthCovariate" />
+      <option value="RepeatUnitCovariate" />
+      <option value="RepeatUnitAndLengthCovariate" />
+      <!--
+      Note: ReadGroupCovariate and QualityScoreCovariate are required covariates and will 
+      be added for the user regardless of whether or not they were specified.
       <option value="QualityScoreCovariate" />
-      <option value="CycleCovariate" />
-      <option value="ContextCovariate" />
-      <!-- covariates below were pulled from list option -->
-      <option value="HomopolymerCovariate" />
-      <option value="GCContentCovariate" />
-      <option value="MappingQualityCovariate" />
-      <option value="MinimumNQSCovariate" />
-      <option value="PositionCovariate" />
-      <option value="PrimerRoundCovariate" />
-      <option value="TileCovariate" />
+      <option value="ReadGroupCovariate" />
+      -->
     </param>
     
-    <repeat name="rod_bind" title="Binding for reference-ordered data" help="-knownSites,--knownSites &amp;lt;knownSites&amp;gt;">
+    <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-knownSites,--knownSites &amp;lt;knownSites&amp;gt;)">
         <conditional name="rod_bind_type">
-          <param name="rod_bind_type_selector" type="select" label="Binding Type">
+          <param name="rod_bind_type_selector" type="select" label="Variant Type">
             <option value="dbsnp" selected="True">dbSNP</option>
             <option value="snps">SNPs</option>
             <option value="indels">INDELs</option>
@@ -145,24 +143,24 @@
             <option value="custom">Custom</option>
           </param>
           <when value="dbsnp">
-              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
           </when>
           <when value="snps">
-              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
           </when>
           <when value="indels">
-              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
           </when>
           <when value="mask">
-              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
           </when>
           <when value="custom">
-              <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
-              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
+              <param name="custom_rod_name" type="text" value="Unknown" label="Customer's variant file"/>
+              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
           </when>
         </conditional>
     </repeat>
-    <param name="plot_pdf" type="boolean" truevalue="--plot_pdf_file" falsevalue="" checked="False" label="Output recalibration plots pdf"/>
+
     <expand macro="gatk_param_type_conditional" />
     
     <conditional name="analysis_param_type">
@@ -243,9 +241,6 @@
   <outputs>
     <data format="gatk_report" name="output_recal" label="${tool.name} on ${on_string} (Covariate File)" />
     <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
-    <data format="pdf" name="output_pdf" label="${tool.name} on ${on_string} (plot_pdf_file)">
-      <filter>plot_pdf == True</filter>
-    </data>
   </outputs>
   <tests>
       <test>
--- a/indel_realigner.xml	Tue Nov 26 06:10:20 2013 -0500
+++ b/indel_realigner.xml	Sat Nov 30 10:36:47 2013 -0500
@@ -87,26 +87,26 @@
       </when>
     </conditional>
     <param name="target_intervals" type="data" format="gatk_interval,bed,picard_interval_list" label="Restrict realignment to provided intervals" help="-targetIntervals,--targetIntervals &amp;lt;targetIntervals&amp;gt;" />
-    <repeat name="rod_bind" title="Binding for reference-ordered data" help="-known,--knownAlleles &amp;lt;knownAlleles&amp;gt;">
+    <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-known,--knownAlleles &amp;lt;knownAlleles&amp;gt;)">
         <conditional name="rod_bind_type">
-          <param name="rod_bind_type_selector" type="select" label="Binding Type">
+          <param name="rod_bind_type_selector" type="select" label="Variant Type">
             <option value="dbsnp" selected="True">dbSNP</option>
             <option value="snps">SNPs</option>
             <option value="indels">INDELs</option>
             <option value="custom">Custom</option>
           </param>
           <when value="dbsnp">
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
           <when value="snps">
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
           <when value="indels">
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
           <when value="custom">
-              <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="custom_rod_name" type="text" value="Unknown" label="Customer's variant file"/>
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
         </conditional>
     </repeat>
--- a/print_reads.xml	Tue Nov 26 06:10:20 2013 -0500
+++ b/print_reads.xml	Sat Nov 30 10:36:47 2013 -0500
@@ -73,7 +73,8 @@
     #end if
   </command>
   <inputs>
-    <param name="input_recal" type="data" format="csv" optional="true" label="Covariates table recalibration file" help="-BQSR,--BQSR &amp;lt;recal_file&amp;gt;" />
+    <param name="input_recal" type="data" format="gatk_report" optional="true" label="Covariates table recalibration file" 
+        help="The input covariates table file which enables on-the-fly base quality score recalibration (intended for use with BaseRecalibrator files) (-BQSR,--BQSR)" />
     <conditional name="reference_source">
       <expand macro="reference_source_selector_param" />
       <when value="cached">
@@ -212,7 +213,7 @@
 
 **Inputs**
 
-GenomeAnalysisTK: PrintReads accepts an aligned BAM and a recalibration CSV input files.
+GenomeAnalysisTK: PrintReads accepts an aligned BAM and a recalibration (gatk_report) input files.
 
 
 **Outputs**
@@ -235,7 +236,7 @@
  exception_if_no_tile           If provided, TileCovariate will throw an exception when no tile can be found. The default behavior is to use tile = -1
  solid_recal_mode               How should we recalibrate solid bases in whichthe reference was inserted? Options = DO_NOTHING, SET_Q_ZERO, SET_Q_ZERO_BASE_N, or REMOVE_REF_BIAS (DO_NOTHING|SET_Q_ZERO|SET_Q_ZERO_BASE_N|REMOVE_REF_BIAS)
  solid_nocall_strategy          Defines the behavior of the recalibrator when it encounters no calls in the color space. Options = THROW_EXCEPTION, LEAVE_READ_UNRECALIBRATED, or PURGE_READ (THROW_EXCEPTION|LEAVE_READ_UNRECALIBRATED|PURGE_READ)
- recal_file                     Filename for the input covariates table recalibration .csv file
+ recal_file                     Filename for the input covariates table recalibration .gatk_report file
  out                            The output BAM file
  bam_compression                Compression level to use for writing BAM files
  disable_bam_indexing           Turn off on-the-fly creation of indices for output BAM files.
--- a/realigner_target_creator.xml	Tue Nov 26 06:10:20 2013 -0500
+++ b/realigner_target_creator.xml	Sat Nov 30 10:36:47 2013 -0500
@@ -74,26 +74,26 @@
       </when>
     </conditional>
     
-    <repeat name="rod_bind" title="Binding for reference-ordered data" help="-known,--known &amp;lt;known&amp;gt;">
+    <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-known,--known &amp;lt;known&amp;gt;)">
         <conditional name="rod_bind_type">
-          <param name="rod_bind_type_selector" type="select" label="Binding Type">
+          <param name="rod_bind_type_selector" type="select" label="Variant Type">
             <option value="dbsnp" selected="True">dbSNP</option>
-            <option value="snps">SNPs</option>
+            <option value="snps">SNPs</option>  
             <option value="indels">INDELs</option>
             <option value="custom">Custom</option>
           </param>
           <when value="dbsnp">
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
           <when value="snps">
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
           <when value="indels">
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
           <when value="custom">
-              <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
-              <param name="input_rod" type="data" format="vcf" label="ROD file" />
+              <param name="custom_rod_name" type="text" value="Unknown" label="Customer's variant file" />
+              <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" />
           </when>
         </conditional>
     </repeat>
--- a/tool_dependencies.xml	Tue Nov 26 06:10:20 2013 -0500
+++ b/tool_dependencies.xml	Sat Nov 30 10:36:47 2013 -0500
@@ -4,6 +4,10 @@
         <environment_variable action="set_to" name="GATK2_PATH">/please set the path to your GATK2 dir in the corresponding env.sh file/</environment_variable>
     </set_environment>
     <!-- 
+    Use GATK2_SITE_OPTIONS to set additional parameters that should be inserted in every GATK2 call.
+    The intended use case was to prohibit GATK2 to collect and send data.
+    For example:
+    
     -et "NO_ET" -K "/data/gatk2_key_file" ##ET no phone home
     -->
     <set_environment version="1.0">