PageRenderTime 29ms CodeModel.GetById 15ms app.highlight 6ms RepoModel.GetById 2ms app.codeStats 0ms

/tools/gatk/realigner_target_creator.xml

https://bitbucket.org/cistrome/cistrome-harvard/
XML | 165 lines | 145 code | 20 blank | 0 comment | 0 complexity | 0daf04f37014a978b8dd20549cd01e47 MD5 | raw file
  1<tool id="gatk_realigner_target_creator" name="Realigner Target Creator" version="0.0.4">
  2  <description>for use in local realignment</description>
  3  <requirements>
  4      <requirement type="package" version="1.3">gatk</requirement>
  5      <requirement type="package">samtools</requirement>
  6  </requirements>
  7  <macros>
  8    <import>gatk_macros.xml</import>
  9  </macros>
 10  <command interpreter="python">gatk_wrapper.py
 11   --max_jvm_heap_fraction "1"
 12   --stdout "${output_log}"
 13   -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
 14   #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
 15       -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
 16   #end if
 17   -p 'java 
 18    -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
 19    -T "RealignerTargetCreator"
 20    -o "${output_interval}"
 21    -et "NO_ET" ##ET no phone home
 22    --num_threads 4 ##hard coded, for now
 23    ##-log "${output_log}" ##don't use this to log to file, instead directly capture stdout
 24    #if $reference_source.reference_source_selector != "history":
 25        -R "${reference_source.ref_file.fields.path}"
 26    #end if
 27   '
 28    #set $rod_binding_names = dict()
 29    #for $rod_binding in $rod_bind:
 30        #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom':
 31            #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name
 32        #else
 33            #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector
 34        #end if
 35        #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1
 36        -d "-known:${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[$rod_bind_name]}"
 37    #end for
 38   
 39    #include source=$standard_gatk_options#
 40    ##start analysis specific options
 41    #if $analysis_param_type.analysis_param_type_selector == "advanced":
 42        -p '
 43        --minReadsAtLocus "${analysis_param_type.minReadsAtLocus}"
 44        --windowSize "${analysis_param_type.windowSize}"
 45        --mismatchFraction "${analysis_param_type.mismatchFraction}"
 46        --maxIntervalSize "${analysis_param_type.maxIntervalSize}"
 47        '
 48    #end if
 49  </command>
 50  <inputs>
 51    <conditional name="reference_source">
 52      <expand macro="reference_source_selector_param" />
 53      <when value="cached">
 54        <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &amp;lt;input_file&amp;gt;">
 55          <validator type="unspecified_build" />
 56          <validator type="dataset_metadata_in_data_table" table_name="gatk_picard_indexes" metadata_name="dbkey" metadata_column="dbkey" message="Sequences are not currently available for the specified build." /> <!-- fixme!!! this needs to be a select -->
 57        </param>
 58        <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;" >
 59          <options from_data_table="gatk_picard_indexes">
 60            <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/>
 61          </options>
 62          <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
 63        </param>
 64      </when>
 65      <when value="history">
 66        <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &amp;lt;input_file&amp;gt;" />
 67        <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;">
 68          <options>
 69            <filter type="data_meta" key="dbkey" ref="input_bam" />
 70          </options>
 71        </param>
 72      </when>
 73    </conditional>
 74    
 75    <repeat name="rod_bind" title="Binding for reference-ordered data" help="-known,--known &amp;lt;known&amp;gt;">
 76        <conditional name="rod_bind_type">
 77          <param name="rod_bind_type_selector" type="select" label="Binding Type">
 78            <option value="dbsnp" selected="True">dbSNP</option>
 79            <option value="snps">SNPs</option>
 80            <option value="indels">INDELs</option>
 81            <option value="custom">Custom</option>
 82          </param>
 83          <when value="dbsnp">
 84              <param name="input_rod" type="data" format="vcf" label="ROD file" />
 85          </when>
 86          <when value="snps">
 87              <param name="input_rod" type="data" format="vcf" label="ROD file" />
 88          </when>
 89          <when value="indels">
 90              <param name="input_rod" type="data" format="vcf" label="ROD file" />
 91          </when>
 92          <when value="custom">
 93              <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
 94              <param name="input_rod" type="data" format="vcf" label="ROD file" />
 95          </when>
 96        </conditional>
 97    </repeat>
 98    
 99    <expand macro="gatk_param_type_conditional" />
100    
101    <expand macro="analysis_type_conditional">
102        <param name="windowSize" type="integer" value="10" label="Window size for calculating entropy or SNP clusters (windowSize)" help="-window,--windowSize &amp;lt;windowSize&amp;gt;" />
103        <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 (-mismatch,--mismatchFraction &amp;lt;mismatchFraction&amp;gt;)"/>
104        <param name="minReadsAtLocus" type="integer" value="4" label="Minimum reads at a locus to enable using the entropy calculation (minReadsAtLocus)" help="-minReads,--minReadsAtLocus &amp;lt;minReadsAtLocus&amp;gt;" />
105        <param name="maxIntervalSize" type="integer" value="500" label="Maximum interval size" help="-maxInterval,--maxIntervalSize &amp;lt;maxIntervalSize&amp;gt;" />
106    </expand>
107  </inputs>
108  <outputs>
109    <data format="gatk_interval" name="output_interval" label="${tool.name} on ${on_string} (GATK intervals)" />
110    <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
111  </outputs>
112  <tests>
113      <test>
114          <param name="reference_source_selector" value="history" />
115          <param name="ref_file" value="phiX.fasta" ftype="fasta" />
116          <param name="input_bam" value="gatk/fake_phiX_reads_1.bam" ftype="bam" />
117          <param name="rod_bind_type_selector" value="dbsnp" />
118          <param name="input_rod" value="gatk/fake_phiX_variant_locations.vcf" ftype="vcf" />
119          <param name="gatk_param_type_selector" value="basic" />
120          <param name="analysis_param_type_selector" value="advanced" />
121          <param name="windowSize" value="10" />
122          <param name="mismatchFraction" value="0.15" />
123          <param name="minReadsAtLocus" value="4" />
124          <param name="maxIntervalSize" value="500" />
125          <output name="output_interval" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.gatk_interval" /> 
126          <output name="output_log" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.log.contains" compare="contains"/>
127      </test>
128  </tests>
129  <help>
130**What it does**
131
132Emits 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.
133
134For more information on local realignment around indels using the GATK, see this `tool specific page &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Local_realignment_around_indels&gt;`_.
135
136To learn about best practices for variant detection using GATK, see this `overview &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Best_Practice_Variant_Detection_with_the_GATK_v3&gt;`_.
137
138If you encounter errors, please view the `GATK FAQ &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Frequently_Asked_Questions&gt;`_.
139
140------
141
142**Inputs**
143
144GenomeAnalysisTK: RealignerTargetCreator accepts an aligned BAM input file.
145
146
147**Outputs**
148
149The output is in GATK Interval format.
150
151
152Go `here &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Input_files_for_the_GATK&gt;`_ for details on GATK file formats.
153
154-------
155
156**Settings**::
157
158 windowSize          window size for calculating entropy or SNP clusters
159 mismatchFraction    fraction of base qualities needing to mismatch for a position to have high entropy; to disable set to &lt;= 0 or &gt; 1
160 minReadsAtLocus     minimum reads at a locus to enable using the entropy calculation
161 maxIntervalSize     maximum interval size
162
163@CITATION_SECTION@
164  </help>
165</tool>