mirror of
https://github.com/galaxyproject/galaxy.git
synced 2026-10-13 08:51:38 +08:00
744 lines
45 KiB
XML
744 lines
45 KiB
XML
<tool id="gatk_depth_of_coverage" name="Depth of Coverage" version="0.0.2">
|
|
<description>on BAM files</description>
|
|
<requirements>
|
|
<requirement type="package" version="1.4">gatk</requirement>
|
|
<requirement type="package">samtools</requirement>
|
|
</requirements>
|
|
<macros>
|
|
<import>gatk_macros.xml</import>
|
|
</macros>
|
|
<command interpreter="python">gatk_wrapper.py
|
|
--max_jvm_heap_fraction "1"
|
|
--stdout "${output_log}"
|
|
#for $i, $input_bam in enumerate( $reference_source.input_bams ):
|
|
-d "-I" "${input_bam.input_bam}" "${input_bam.input_bam.ext}" "gatk_input_${i}"
|
|
#if str( $input_bam.input_bam.metadata.bam_index ) != "None":
|
|
-d "" "${input_bam.input_bam.metadata.bam_index}" "bam_index" "gatk_input_${i}" ##hardcode galaxy ext type as bam_index
|
|
#end if
|
|
#end for
|
|
-p 'java
|
|
-jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
|
|
-T "DepthOfCoverage"
|
|
##--num_threads 4 ##hard coded, for now
|
|
|
|
-et "NO_ET" ##ET no phone home
|
|
#if $reference_source.reference_source_selector != "history":
|
|
-R "${reference_source.ref_file.fields.path}"
|
|
#end if
|
|
#if str( $input_calculate_coverage_over_genes ) != "None":
|
|
--calculateCoverageOverGenes "${input_calculate_coverage_over_genes}"
|
|
#end if
|
|
#if str( $partition_type ) != "None":
|
|
#for $pt in str( $partition_type ).split( ',' ):
|
|
--partitionType "${pt}"
|
|
#end for
|
|
#end if
|
|
--out "${output_per_locus_coverage}"
|
|
|
|
#for $ct_group in $summary_coverage_threshold_group:
|
|
--summaryCoverageThreshold "${ct_group.summary_coverage_threshold}"
|
|
#end for
|
|
--outputFormat "${output_format}"
|
|
'
|
|
|
|
#include source=$standard_gatk_options#
|
|
##start analysis specific options
|
|
#if $analysis_param_type.analysis_param_type_selector == "advanced":
|
|
-p '
|
|
${analysis_param_type.ignore_deletion_sites}
|
|
${analysis_param_type.include_deletions}
|
|
--maxBaseQuality "${analysis_param_type.max_base_quality}"
|
|
--maxMappingQuality "${analysis_param_type.max_mapping_quality}"
|
|
--minBaseQuality "${analysis_param_type.min_base_quality}"
|
|
--minMappingQuality "${analysis_param_type.min_mapping_quality}"
|
|
--nBins "${analysis_param_type.n_bins}"
|
|
${analysis_param_type.omit_depth_output_at_each_base}
|
|
${analysis_param_type.omit_interval_statistics}
|
|
${analysis_param_type.omit_locus_table}
|
|
${analysis_param_type.omit_per_sample_stats}
|
|
${analysis_param_type.print_base_counts}
|
|
${analysis_param_type.print_bin_endpoints_and_exit}
|
|
--start "${analysis_param_type.start}"
|
|
--stop "${analysis_param_type.stop}"
|
|
'
|
|
#end if
|
|
##Move additional files to final location
|
|
#if str( $partition_type ) != "None":
|
|
#set $partition_types = str( $partition_type ).split( ',' )
|
|
#else:
|
|
#set $partition_types = [ 'sample' ]
|
|
#end if
|
|
#if 'sample' in $partition_types and ( str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.print_bin_endpoints_and_exit ) == "" ):
|
|
#if str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_per_sample_stats ) == "":
|
|
&& mv ${output_per_locus_coverage}.sample_summary ${output_summary_sample}
|
|
&& mv ${output_per_locus_coverage}.sample_statistics ${output_statistics_sample}
|
|
#end if
|
|
#if $gatk_param_type.gatk_param_type_selector == "advanced" and len( $gatk_param_type.input_interval_repeat ) and ( str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_interval_statistics ) == "" ):
|
|
&& mv ${output_per_locus_coverage}.sample_interval_summary ${output_interval_summary_sample}
|
|
&& mv ${output_per_locus_coverage}.sample_interval_statistics ${output_interval_statistics_sample}
|
|
#end if
|
|
#if str( $input_calculate_coverage_over_genes ) != "None":
|
|
&& mv ${output_per_locus_coverage}.sample_gene_summary ${output_gene_summary_sample}
|
|
&& mv ${output_per_locus_coverage}.sample_gene_statistics ${output_gene_statistics_sample}
|
|
#end if
|
|
#if str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_depth_output_at_each_base ) == "":
|
|
&& mv ${output_per_locus_coverage}.sample_cumulative_coverage_counts ${output_cumulative_coverage_counts_sample}
|
|
&& mv ${output_per_locus_coverage}.sample_cumulative_coverage_proportions ${output_cumulative_coverage_proportions_sample}
|
|
#end if
|
|
#end if
|
|
|
|
#if 'readgroup' in $partition_types and ( str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.print_bin_endpoints_and_exit ) == "" ):
|
|
#if str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_per_sample_stats ) == "":
|
|
&& mv ${output_per_locus_coverage}.read_group_summary ${output_summary_readgroup}
|
|
&& mv ${output_per_locus_coverage}.read_group_statistics ${output_statistics_readgroup}
|
|
#end if
|
|
#if $gatk_param_type.gatk_param_type_selector == "advanced" and len( $gatk_param_type.input_interval_repeat ) and ( str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_interval_statistics ) == "" ):
|
|
&& mv ${output_per_locus_coverage}.read_group_interval_summary ${output_interval_summary_readgroup}
|
|
&& mv ${output_per_locus_coverage}.read_group_interval_statistics ${output_interval_statistics_readgroup}
|
|
#end if
|
|
#if str( $input_calculate_coverage_over_genes ) != "None":
|
|
&& mv ${output_per_locus_coverage}.read_group_gene_summary ${output_gene_summary_readgroup}
|
|
&& mv ${output_per_locus_coverage}.read_group_gene_statistics ${output_gene_statistics_readgroup}
|
|
#end if
|
|
#if str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_depth_output_at_each_base ) == "":
|
|
&& mv ${output_per_locus_coverage}.read_group_cumulative_coverage_counts ${output_cumulative_coverage_counts_readgroup}
|
|
&& mv ${output_per_locus_coverage}.read_group_cumulative_coverage_proportions ${output_cumulative_coverage_proportions_readgroup}
|
|
#end if
|
|
#end if
|
|
|
|
#if 'library' in $partition_types and ( str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.print_bin_endpoints_and_exit ) == "" ):
|
|
#if str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_per_sample_stats ) == "":
|
|
&& mv ${output_per_locus_coverage}.library_summary ${output_summary_library}
|
|
&& mv ${output_per_locus_coverage}.library_statistics ${output_statistics_library}
|
|
#end if
|
|
#if $gatk_param_type.gatk_param_type_selector == "advanced" and len( $gatk_param_type.input_interval_repeat ) and ( str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_interval_statistics ) == "" ):
|
|
&& mv ${output_per_locus_coverage}.library_interval_summary ${output_interval_summary_library}
|
|
&& mv ${output_per_locus_coverage}.library_interval_statistics ${output_interval_statistics_library}
|
|
#end if
|
|
#if str( $input_calculate_coverage_over_genes ) != "None":
|
|
&& mv ${output_per_locus_coverage}.library_gene_summary ${output_gene_summary_library}
|
|
&& mv ${output_per_locus_coverage}.library_gene_statistics ${output_gene_statistics_library}
|
|
#end if
|
|
#if str( $analysis_param_type.analysis_param_type_selector ) == "basic" or str( $analysis_param_type.omit_depth_output_at_each_base ) == "":
|
|
&& mv ${output_per_locus_coverage}.library_cumulative_coverage_counts ${output_cumulative_coverage_counts_library}
|
|
&& mv ${output_per_locus_coverage}.library_cumulative_coverage_proportions ${output_cumulative_coverage_proportions_library}
|
|
#end if
|
|
#end if
|
|
|
|
|
|
</command>
|
|
<inputs>
|
|
<conditional name="reference_source">
|
|
<expand macro="reference_source_selector_param" />
|
|
<when value="cached">
|
|
<repeat name="input_bams" title="BAM file" min="1" help="-I,--input_file &lt;input_file&gt;">
|
|
<param name="input_bam" type="data" format="bam" label="BAM file">
|
|
<validator type="unspecified_build" />
|
|
<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 -->
|
|
</param>
|
|
</repeat>
|
|
<param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;">
|
|
<options from_data_table="gatk_picard_indexes">
|
|
<!-- <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/> does not yet work in a repeat...-->
|
|
</options>
|
|
<validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
|
|
</param>
|
|
</when>
|
|
<when value="history"> <!-- FIX ME!!!! -->
|
|
<repeat name="input_bams" title="BAM file" min="1" help="-I,--input_file &lt;input_file&gt;">
|
|
<param name="input_bam" type="data" format="bam" label="BAM file" />
|
|
</repeat>
|
|
<param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" />
|
|
</when>
|
|
</conditional>
|
|
|
|
<param name="input_calculate_coverage_over_genes" type="data" format="data" label="RefSeq Rod" optional="True" help="-geneList,--calculateCoverageOverGenes &lt;calculateCoverageOverGenes&gt;" />
|
|
|
|
<param name="partition_type" type="select" label="Partition type for depth of coverage" multiple="True" display="checkboxes" help="-pt,--partitionType &lt;partitionType&gt;">
|
|
<option value="sample" selected="True">sample</option>
|
|
<option value="readgroup">readgroup</option>
|
|
<option value="library">library</option>
|
|
</param>
|
|
|
|
<repeat name="summary_coverage_threshold_group" title="Summary coverage threshold" help="-ct,--summaryCoverageThreshold &lt;summaryCoverageThreshold&gt;">
|
|
<param name="summary_coverage_threshold" type="integer" value="15" label="for summary file outputs, report the % of bases covered to >= this number" />
|
|
</repeat>
|
|
|
|
<param name="output_format" type="select" label="Output format" help="--outputFormat &lt;outputFormat&gt;" >
|
|
<option value="csv">csv</option>
|
|
<option value="table">table</option>
|
|
<option value="rtable" selected="True">rtable</option>
|
|
</param>
|
|
|
|
<expand macro="gatk_param_type_conditional" />
|
|
|
|
<expand macro="analysis_type_conditional">
|
|
<param name="ignore_deletion_sites" type="boolean" truevalue="--ignoreDeletionSites" falsevalue="" checked="False" label="Ignore sites consisting only of deletions" help="--ignoreDeletionSites" />
|
|
<param name="include_deletions" type="boolean" truevalue="--includeDeletions" falsevalue="" checked="False" label="Include information on deletions" help="-dels,--includeDeletions" />
|
|
<param name="max_base_quality" type="integer" value="127" label="Maximum quality of bases to count towards depth" help="--maxBaseQuality &lt;maxBaseQuality&gt;" />
|
|
<param name="min_base_quality" type="integer" value="-1" label="Minimum quality of bases to count towards depth" help="-mbq,--minBaseQuality &lt;minBaseQuality&gt;" />
|
|
<param name="max_mapping_quality" type="integer" value="2147483647" label="Maximum mapping quality of reads to count towards depth." help="--maxMappingQuality &lt;maxMappingQuality&gt;" />
|
|
<param name="min_mapping_quality" type="integer" value="127" label="Minimum mapping quality of reads to count towards depth" help="-mmq,--minMappingQuality &lt;minMappingQuality&gt;" />
|
|
<param name="n_bins" type="integer" value="499" label="Number of bins to use for granular binning" help="--nBins &lt;nBins&gt;" />
|
|
<param name="omit_depth_output_at_each_base" type="boolean" truevalue="--omitDepthOutputAtEachBase" falsevalue="" checked="False" label="Omit the output of the depth of coverage at each base" help="-omitBaseOutput,--omitDepthOutputAtEachBase" />
|
|
<param name="omit_interval_statistics" type="boolean" truevalue="--omitIntervalStatistics" falsevalue="" checked="False" label="Omit the per-interval statistics section" help="-omitIntervals,--omitIntervalStatistics" />
|
|
<param name="omit_locus_table" type="boolean" truevalue="--omitLocusTable" falsevalue="" checked="False" label="Do not calculate the per-sample per-depth counts of loci" help="-omitLocusTable,--omitLocusTable" />
|
|
<param name="omit_per_sample_stats" type="boolean" truevalue="--omitPerSampleStats" falsevalue="" checked="False" label="Omit the summary files per-sample." help="-omitSampleSummary,--omitPerSampleStats" />
|
|
<param name="print_base_counts" type="boolean" truevalue="--printBaseCounts" falsevalue="" checked="False" label="Add base counts to per-locus output" help="-baseCounts,--printBaseCounts" />
|
|
<param name="print_bin_endpoints_and_exit" type="boolean" truevalue="--printBinEndpointsAndExit" falsevalue="" checked="False" label="Print the bin values and exits immediately" help="--printBinEndpointsAndExit" />
|
|
<param name="start" type="integer" value="1" label="Starting (left endpoint) for granular binning" help="--start &lt;start&gt;" />
|
|
<param name="stop" type="integer" value="500" label="Ending (right endpoint) for granular binning" help="--stop &lt;stop&gt;" />
|
|
</expand>
|
|
</inputs>
|
|
<outputs>
|
|
<data format="tabular" name="output_per_locus_coverage" label="${tool.name} on ${on_string} (per locus coverage)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_summary_sample" label="${tool.name} on ${on_string} (output summary sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'sample' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_statistics_sample" label="${tool.name} on ${on_string} (output statistics sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'sample' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_interval_summary_sample" label="${tool.name} on ${on_string} (output interval summary sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'sample' in partition_type or not partition_type</filter>
|
|
<filter>gatk_param_type['gatk_param_type_selector'] == "advanced" and len( gatk_param_type['input_interval_repeat'] )</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_interval_statistics'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_interval_statistics_sample" label="${tool.name} on ${on_string} (output interval statistics sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'sample' in partition_type or not partition_type</filter>
|
|
<filter>gatk_param_type['gatk_param_type_selector'] == "advanced" and len( gatk_param_type['input_interval_repeat'] )</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_interval_statistics'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_gene_summary_sample" label="${tool.name} on ${on_string} (output gene summary sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>input_calculate_coverage_over_genes is not None and 'sample' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_gene_statistics_sample" label="${tool.name} on ${on_string} (output gene statistics sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>input_calculate_coverage_over_genes is not None and 'sample' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_cumulative_coverage_counts_sample" label="${tool.name} on ${on_string} (output cumulative coverage counts sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_depth_output_at_each_base'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'sample' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_cumulative_coverage_proportions_sample" label="${tool.name} on ${on_string} (output cumulative coverage proportions sample)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_depth_output_at_each_base'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'sample' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
|
|
<data format="tabular" name="output_summary_readgroup" label="${tool.name} on ${on_string} (output summary readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'readgroup' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_statistics_readgroup" label="${tool.name} on ${on_string} (output statistics readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'readgroup' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_interval_summary_readgroup" label="${tool.name} on ${on_string} (output interval summary readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'readgroup' in partition_type</filter>
|
|
<filter>gatk_param_type['gatk_param_type_selector'] == "advanced" and len( gatk_param_type['input_interval_repeat'] )</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_interval_statistics'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_interval_statistics_readgroup" label="${tool.name} on ${on_string} (output interval statistics readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'readgroup' in partition_type</filter>
|
|
<filter>gatk_param_type['gatk_param_type_selector'] == "advanced" and len( gatk_param_type['input_interval_repeat'] )</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_interval_statistics'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_gene_summary_readgroup" label="${tool.name} on ${on_string} (output gene summary readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>input_calculate_coverage_over_genes is not None and 'readgroup' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_gene_statistics_readgroup" label="${tool.name} on ${on_string} (output gene statistics readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>input_calculate_coverage_over_genes is not None and 'readgroup' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_cumulative_coverage_counts_readgroup" label="${tool.name} on ${on_string} (output cumulative coverage counts readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_depth_output_at_each_base'] == False</filter>
|
|
<filter>'readgroup' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_cumulative_coverage_proportions_readgroup" label="${tool.name} on ${on_string} (output cumulative coverage proportions readgroup)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_depth_output_at_each_base'] == False</filter>
|
|
<filter>'readgroup' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
|
|
<data format="tabular" name="output_summary_library" label="${tool.name} on ${on_string} (output summary library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'library' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_statistics_library" label="${tool.name} on ${on_string} (output statistics library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'library' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_interval_summary_library" label="${tool.name} on ${on_string} (output interval summary library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'library' in partition_type</filter>
|
|
<filter>gatk_param_type['gatk_param_type_selector'] == "advanced" and len( gatk_param_type['input_interval_repeat'] )</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_interval_statistics'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_interval_statistics_library" label="${tool.name} on ${on_string} (output interval statistics library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'library' in partition_type</filter>
|
|
<filter>gatk_param_type['gatk_param_type_selector'] == "advanced" and len( gatk_param_type['input_interval_repeat'] )</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_interval_statistics'] == False</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_gene_summary_library" label="${tool.name} on ${on_string} (output gene summary library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>input_calculate_coverage_over_genes is not None and 'library' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_gene_statistics_library" label="${tool.name} on ${on_string} (output gene statistics library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>input_calculate_coverage_over_genes is not None and 'library' in partition_type or not partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_cumulative_coverage_counts_library" label="${tool.name} on ${on_string} (output cumulative coverage counts library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_depth_output_at_each_base'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'library' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
<data format="tabular" name="output_cumulative_coverage_proportions_library" label="${tool.name} on ${on_string} (output cumulative coverage proportions library)" >
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_depth_output_at_each_base'] == False</filter>
|
|
<filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter>
|
|
<filter>'library' in partition_type</filter>
|
|
<actions>
|
|
<conditional name="output_format">
|
|
<when value="rtable">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
<when value="csv">
|
|
<action type="format">
|
|
<option type="from_param" name="output_format" />
|
|
</action>
|
|
</when>
|
|
</conditional>
|
|
</actions>
|
|
</data>
|
|
|
|
<data format="tabular" name="output_log" label="${tool.name} on ${on_string} (log)" />
|
|
</outputs>
|
|
<trackster_conf/>
|
|
<tests>
|
|
<test>
|
|
<param name="reference_source_selector" value="history" />
|
|
<param name="ref_file" value="phiX.fasta" ftype="fasta" />
|
|
<param name="input_bam" value="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.bam" ftype="bam" />
|
|
<param name="input_calculate_coverage_over_genes" />
|
|
<param name="partition_type" value="sample" />
|
|
<param name="summary_coverage_threshold_group" value="0" />
|
|
<param name="output_format" value="rtable" />
|
|
<param name="gatk_param_type_selector" value="basic" />
|
|
<param name="analysis_param_type_selector" value="basic" />
|
|
<output name="output_per_locus_coverage" file="gatk/gatk_depth_of_coverage/gatk_depth_of_coverage_out_1_output_per_locus_coverage.tabular" />
|
|
<output name="output_summary_sample" file="gatk/gatk_depth_of_coverage/gatk_depth_of_coverage_out_1_output_summary_sample.tabular" />
|
|
<output name="output_statistics_sample" file="gatk/gatk_depth_of_coverage/gatk_depth_of_coverage_out_1_output_statistics_sample.tabular" />
|
|
<output name="output_cumulative_coverage_counts_sample" file="gatk/gatk_depth_of_coverage/gatk_depth_of_coverage_out_1_output_cumulative_coverage_counts_sample.tabular" />
|
|
<output name="output_cumulative_coverage_proportions_sample" file="gatk/gatk_depth_of_coverage/gatk_depth_of_coverage_out_1_output_output_cumulative_coverage_proportions_sample.tabular" />
|
|
<output name="output_log" file="gatk/gatk_depth_of_coverage/gatk_depth_of_coverage_out_1.log.contains" compare="contains" />
|
|
</test>
|
|
</tests>
|
|
<help>
|
|
**What it does**
|
|
|
|
DepthOfCoverage processes a set of bam files to determine coverage at different levels of partitioning and aggregation. Coverage can be analyzed per locus, per interval, per gene, or in total; can be partitioned by sample, by read group, by technology, by center, or by library; and can be summarized by mean, median, quartiles, and/or percentage of bases covered to or beyond a threshold. Additionally, reads and bases can be filtered by mapping or base quality score.
|
|
|
|
For more information on the GATK Depth of Coverage, see this `tool specific page <http://www.broadinstitute.org/gsa/wiki/index.php/Depth_of_Coverage>`_.
|
|
|
|
To learn about best practices for variant detection using GATK, see this `overview <http://www.broadinstitute.org/gsa/wiki/index.php/Best_Practice_Variant_Detection_with_the_GATK_v3>`_.
|
|
|
|
If you encounter errors, please view the `GATK FAQ <http://www.broadinstitute.org/gsa/wiki/index.php/Frequently_Asked_Questions>`_.
|
|
|
|
------
|
|
|
|
**Inputs**
|
|
|
|
GenomeAnalysisTK: DepthOfCoverage accepts aligned BAM input files.
|
|
|
|
|
|
**Outputs**
|
|
|
|
The output is in various table formats.
|
|
|
|
|
|
Go `here <http://www.broadinstitute.org/gsa/wiki/index.php/Input_files_for_the_GATK>`_ for details on GATK file formats.
|
|
|
|
-------
|
|
|
|
**Settings**::
|
|
|
|
calculateCoverageOverGenes File NA Calculate the coverage statistics over this list of genes. Currently accepts RefSeq.
|
|
ignoreDeletionSites boolean false Ignore sites consisting only of deletions
|
|
includeDeletions boolean false Include information on deletions
|
|
maxBaseQuality byte 127 Maximum quality of bases to count towards depth. Defaults to 127 (Byte.MAX_VALUE).
|
|
maxMappingQuality int 2147483647 Maximum mapping quality of reads to count towards depth. Defaults to 2^31-1 (Integer.MAX_VALUE).
|
|
minBaseQuality byte -1 Minimum quality of bases to count towards depth. Defaults to -1.
|
|
minMappingQuality int -1 Minimum mapping quality of reads to count towards depth. Defaults to -1.
|
|
nBins int 499 Number of bins to use for granular binning
|
|
omitDepthOutputAtEachBase boolean false Will omit the output of the depth of coverage at each base, which should result in speedup
|
|
omitIntervalStatistics boolean false Will omit the per-interval statistics section, which should result in speedup
|
|
omitLocusTable boolean false Will not calculate the per-sample per-depth counts of loci, which should result in speedup
|
|
omitPerSampleStats boolean false Omits the summary files per-sample. These statistics are still calculated, so this argument will not improve runtime.
|
|
outputFormat String rtable the format of the output file (e.g. csv, table, rtable); defaults to r-readable table
|
|
partitionType Set[Partition] [sample] Partition type for depth of coverage. Defaults to sample. Can be any combination of sample, readgroup, library.
|
|
printBaseCounts boolean false Will add base counts to per-locus output.
|
|
printBinEndpointsAndExit boolean false Prints the bin values and exits immediately. Use to calibrate what bins you want before running on data.
|
|
start int 1 Starting (left endpoint) for granular binning
|
|
stop int 500 Ending (right endpoint) for granular binning
|
|
summaryCoverageThreshold int[] [15] for summary file outputs, report the % of bases coverd to >= this number. Defaults to 15; can take multiple arguments.
|
|
|
|
@CITATION_SECTION@
|
|
</help>
|
|
</tool>
|