view tools/gatk/realigner_target_creator.xml @ 1:cdcb0ce84a1b

Uploaded

<tool id="gatk_realigner_target_creator" name="Realigner Target Creator" version="0.0.1">
  <description>for use in local realignment</description>
  <command interpreter="python">gatk_wrapper.py
   --stdout "${output_log}"
   -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
   -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
   -p 'java 
    -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
    -T "RealignerTargetCreator"
    -o "${output_interval}"
    -et "NO_ET" ##ET no phone home
    ##-log "${output_log}" ##don't use this to log to file, instead directly capture stdout
    #if $reference_source.reference_source_selector != "history":
        -R "${reference_source.ref_file.fields.path}"
    #end if
   '
    #set $rod_binding_names = dict()
    #if str( $input_dbsnp_rod ) != "None":
        -d "-D" "${input_dbsnp_rod}" "${input_dbsnp_rod.ext}" "dbsnp_rod"
    #end if
    #for $rod_binding in $rod_bind:
        #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom':
            #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name
        #else
            #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector
        #end if
        #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1
        -d "-B:${rod_bind_name},%(file_type)s" "${rod_binding.rod_bind_type.input_rod}" "${rod_binding.rod_bind_type.input_rod.ext}" "input_${rod_bind_name}_${rod_binding_names[$rod_bind_name]}"
        #if str( $rod_binding.rod_bind_type.rodToIntervalTrackName ):
            -p '--rodToIntervalTrackName "${rod_bind_name}"'
        #end if
    #end for
   
    ##start standard gatk options
    #if $gatk_param_type.gatk_param_type_selector == "advanced":
        #for $sample_metadata in $gatk_param_type.sample_metadata:
            -p '--sample_metadata "${sample_metadata.sample_metadata_file}"'
        #end for
        #for $read_filter in $gatk_param_type.read_filter:
            -p '--read_filter "${read_filter.read_filter_type.read_filter_type_selector}"
            ###raise Exception( str( dir( $read_filter ) ) )
            #for $name, $param in $read_filter.read_filter_type.iteritems():
                #if $name not in [ "__current_case__", "read_filter_type_selector" ]:
                    --${name} "${param}"
                #end if
            #end for
            '
        #end for
        #if str( $gatk_param_type.input_intervals ) != "None":
            -d "-L" "${gatk_param_type.input_intervals}" "${gatk_param_type.input_intervals.ext}" "input_intervals"
        #end if
        #if str( $gatk_param_type.input_exclude_intervals ) != "None":
            -d "-XL" "${gatk_param_type.input_exclude_intervals}" "${gatk_param_type.input_exclude_intervals.ext}" "input_intervals"
        #end if

        -p '--BTI_merge_rule "${gatk_param_type.BTI_merge_rule}"'
        
        -p '--downsampling_type "${gatk_param_type.downsampling_type.downsampling_type_selector}"'
        #if str( $gatk_param_type.downsampling_type.downsampling_type_selector ) != "NONE":
            -p '--${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_type_selector} "${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_value}"'
        #end if
        -p '
        --baq "${gatk_param_type.baq}"
        --baqGapOpenPenalty "${gatk_param_type.baq_gap_open_penalty}"
        ${gatk_param_type.use_original_qualities}
        --defaultBaseQualities "${gatk_param_type.default_base_qualities}"
        --validation_strictness "${gatk_param_type.validation_strictness}"
        --interval_merging "${gatk_param_type.interval_merging}"
        '
        #if str( $gatk_param_type.read_group_black_list ) != "None":
            -d "-read_group_black_list" "${gatk_param_type.read_group_black_list}" "txt" "input_read_group_black_list"
        #end if
    #end if
    #if $reference_source.reference_source_selector == "history":
        -d "-R" "${reference_source.ref_file}" "${reference_source.ref_file.ext}" "gatk_input"
    #end if
    ##end standard gatk options
    ##start analysis specific options
    #if $analysis_param_type.analysis_param_type_selector == "advanced":
        -p '
        --minReadsAtLocus "${analysis_param_type.minReadsAtLocus}"
        --windowSize "${analysis_param_type.windowSize}"
        --mismatchFraction "${analysis_param_type.mismatchFraction}"
        --maxIntervalSize "${analysis_param_type.maxIntervalSize}"
        '
    #end if
  </command>
  <inputs>
    <conditional name="reference_source">
      <param name="reference_source_selector" type="select" label="Choose the source for the reference list">
        <option value="cached">Locally cached</option>
        <option value="history">History</option>
      </param>
      <when value="cached">
        <param name="input_bam" type="data" format="bam" label="BAM file">
          <validator type="unspecified_build" />
          <validator type="dataset_metadata_in_file" filename="picard_index.loc" metadata_name="dbkey" metadata_column="1" message="Sequences are not currently available for the specified build." /> <!-- fixme!!! this needs to be a select -->
        </param>
        <param name="ref_file" type="select" label="Using reference genome">
          <options from_data_table="picard_indexes">
            <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/>
          </options>
        </param>
      </when>
      <when value="history"> <!-- FIX ME!!!! -->
        <param name="input_bam" type="data" format="bam" label="BAM file" />
        <param name="ref_file" type="data" format="fasta" label="Using reference file" />
      </when>
    </conditional>
    
    <param name="input_dbsnp_rod" type="data" format="gatk_dbsnp" optional="True" label="dbSNP reference ordered data (ROD)" />
    <repeat name="rod_bind" title="Binding for reference-ordered data">
        <conditional name="rod_bind_type">
	      <param name="rod_bind_type_selector" type="select" label="Binding Type">
	        <option value="snps" selected="True">SNPs</option>
	        <option value="indels">INDELs</option>
	        <option value="custom">Custom</option>
	      </param>
          <when value="snps">
              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
              <param name="rodToIntervalTrackName" type="boolean" truevalue="--rodToIntervalTrackName" falsevalue="" label="Use ROD as interval List (-BTI, --rodToIntervalTrackName)" help="Only one ROD may have this option specified" />
          </when>
          <when value="indels">
              <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="ROD file" />
              <param name="rodToIntervalTrackName" type="boolean" truevalue="--rodToIntervalTrackName" falsevalue="" label="Use ROD as interval List (-BTI, --rodToIntervalTrackName)" help="Only one ROD may have this option specified" />
          </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="rodToIntervalTrackName" type="boolean" truevalue="--rodToIntervalTrackName" falsevalue="" label="Use ROD as interval List (-BTI, --rodToIntervalTrackName)" help="Only one ROD may have this option specified" />
          </when>
        </conditional>
    </repeat>
    
    <conditional name="gatk_param_type">
      <param name="gatk_param_type_selector" type="select" label="Basic or Advanced GATK options">
        <option value="basic" selected="True">Basic</option>
        <option value="advanced">Advanced</option>
      </param>
      <when value="basic">
        <!-- Do nothing here -->
      </when>
      <when value="advanced">
        <repeat name="sample_metadata" title="Sample Metadata">
            <param name="sample_metadata_file" type="data" format="txt" label="Sample file(s) in JSON format" />
        </repeat>
        <repeat name="read_filter" title="Read Filter">
            <conditional name="read_filter_type">
		      <param name="read_filter_type_selector" type="select" label="Read Filter Type">
		        <option value="MaxReadLength" selected="True">MaxReadLength</option>
		        <option value="ZeroMappingQualityRead">ZeroMappingQualityRead</option>
		      </param>
	          <when value="ZeroMappingQualityRead">
	              <!-- no extra options -->
	          </when>
	          <when value="MaxReadLength">
	              <param name="maxReadLength" type="integer" value="76" label="Max Read Length"/>
	          </when>
            </conditional>
        </repeat>
        <param name="input_intervals" type="data" format="picard_interval_list" optional="True" label="A list of genomic intervals over which to operate" />
        <param name="input_exclude_intervals" type="data" format="picard_interval_list" optional="True" label="A list of genomic intervals to exclude from processing" />
        
        <param name="BTI_merge_rule" type="select" label="BTI merge rule">
          <option value="UNION" selected="True">UNION</option>
          <option value="INTERSECTION">INTERSECTION</option>
        </param>
        
        <conditional name="downsampling_type">
          <param name="downsampling_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="Downsampling Type">
            <option value="NONE" selected="True">NONE</option>
            <option value="ALL_READS">ALL_READS</option>
            <option value="BY_SAMPLE">BY_SAMPLE</option>
          </param>
          <when value="NONE">
	          <!-- no more options here -->
	      </when>
          <when value="ALL_READS">
	          <conditional name="downsample_to_type">
	              <param name="downsample_to_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="Downsampling Type">
	                  <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
	                  <option value="downsample_to_coverage">Downsample by Coverage</option>
	              </param>
	              <when value="downsample_to_fraction">
	                  <param name="downsample_to_value" type="float" label="Fraction [0.0-1.0] of reads to downsample to" value="0.1"/>
	              </when>
	              <when value="downsample_to_coverage">
	                  <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0"/>
	              </when>
	          </conditional>
	      </when>
          <when value="BY_SAMPLE">
	          <conditional name="downsample_to_type">
	              <param name="downsample_to_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="Downsampling Type">
	                  <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
	                  <option value="downsample_to_coverage">Downsample by Coverage</option>
	              </param>
	              <when value="downsample_to_fraction">
	                  <param name="downsample_to_value" type="float" label="Fraction [0.0-1.0] of reads to downsample to" value="0.1"/>
	              </when>
	              <when value="downsample_to_coverage">
	                  <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0"/>
	              </when>
	          </conditional>
	      </when>
        </conditional>
        <param name="baq" type="select" label="Type of BAQ calculation to apply in the engine">
          <option value="OFF" selected="True">OFF</option>
          <option value="CALCULATE_AS_NECESSARY">CALCULATE_AS_NECESSARY</option>
          <option value="RECALCULATE">RECALCULATE</option>
        </param>
        <param name="baq_gap_open_penalty" type="integer" label="BAQ gap open penalty (Phred Scaled)" value="40" help="Default value is 40. 30 is perhaps better for whole genome call sets."/>
        <param name="use_original_qualities" type="boolean" truevalue="--useOriginalQualities" falsevalue="" label="Use the original base quality scores from the OQ tag" />
        <param name="default_base_qualities" type="integer" label="Value to be used for all base quality scores, when some are missing" value="-1"/>
        <param name="validation_strictness" type="select" label="How strict should we be with validation">
          <option value="STRICT" selected="True">STRICT</option>
          <option value="LENIENT">LENIENT</option>
          <option value="SILENT">SILENT</option>
        </param>
        <param name="interval_merging" type="select" label="Interval merging rule">
          <option value="ALL" selected="True">ALL</option>
          <option value="OVERLAPPING_ONLY">OVERLAPPING_ONLY</option>
        </param>
        <param name="read_group_black_list" type="data" format="txt" optional="True" label="Read group black list" />
      </when>
    </conditional>
    
    <conditional name="analysis_param_type">
      <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options">
        <option value="basic" selected="True">Basic</option>
        <option value="advanced">Advanced</option>
      </param>
      <when value="basic">
        <!-- Do nothing here -->
      </when>
      <when value="advanced">
        <param name="windowSize" type="integer" value="10" label="Window size for calculating entropy or SNP clusters (windowSize)" />
        <param name="mismatchFraction" type="float" value="0.15" label="Fraction of base qualities needing to mismatch for a position to have high entropy (mismatchFraction)" help="to disable set to &lt;= 0 or &gt; 1"/>
        <param name="minReadsAtLocus" type="integer" value="4" label="Minimum reads at a locus to enable using the entropy calculation (minReadsAtLocus)" />
        <param name="maxIntervalSize" type="integer" value="500" label="Maximum interval size" />
      </when>
    </conditional>
  </inputs>
  <outputs>
    <data format="gatk_interval" name="output_interval" label="${tool.name} on ${on_string} (GATK intervals)" />
    <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
  </outputs>
  <tests>
      <test>
          <param name="reference_source_selector" value="history" />
          <param name="ref_file" value="phiX.fasta" ftype="fasta" />
          <param name="input_bam" value="gatk/fake_phiX_reads_1.bam" ftype="bam" />
          <param name="input_dbsnp_rod"  />
          <param name="rod_bind_type_selector" value="snps" />
          <param name="rodToIntervalTrackName" />
          <param name="input_rod" value="gatk/fake_phiX_variant_locations.bed" ftype="bed" />
          <param name="gatk_param_type_selector" value="basic" />
          <param name="analysis_param_type_selector" value="basic" />
          <output name="output_interval" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.gatk_interval" /> 
          <output name="output_log" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.log.contains" compare="contains"/>
      </test>
  </tests>
  <help>
**What it does**

Emits intervals for the Local Indel Realigner to target for cleaning.  Ignores 454 reads, MQ0 reads, and reads with consecutive indel operators in the CIGAR string.

------

Please cite the website "http://addlink.here" as well as:

Add citation here 2011.

------

**Input formats**

GenomeAnalysisTK: RealignerTargetCreator accepts an aligned BAM input file.

------

**Outputs**

The output is in GATK Interval format, see http://addlink.here for more details.

-------

**Settings**::

 windowSize          window size for calculating entropy or SNP clusters
 mismatchFraction    fraction of base qualities needing to mismatch for a position to have high entropy; to disable set to &lt;= 0 or &gt; 1
 minReadsAtLocus     minimum reads at a locus to enable using the entropy calculation
 maxIntervalSize     maximum interval size

  </help>
</tool>