Mercurial > repos > iuc > gatk2
changeset 11:0d369d08ad6e draft
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/gatk2 commit 344140b8df53b8b7024618bb04594607a045c03a
| author | iuc |
|---|---|
| date | Mon, 04 May 2015 22:47:06 -0400 |
| parents | 5db8d6815cf3 |
| children | 86007c982f00 |
| files | base_recalibrator.xml depth_of_coverage.xml haplotype_caller.xml indel_realigner.xml print_reads.xml realigner_target_creator.xml reduce_reads.xml tool_dependencies.xml unified_genotyper.xml variant_annotator.xml variant_apply_recalibration.xml variant_combine.xml variant_eval.xml variant_filtration.xml variant_recalibrator.xml variant_select.xml variant_validate.xml |
| diffstat | 17 files changed, 226 insertions(+), 247 deletions(-) [+] |
line wrap: on
line diff
--- a/base_recalibrator.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/base_recalibrator.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_base_recalibrator" name="Base Recalibrator" version="@VERSION@.0"> <description>calculates covariates used to recalibrate base quality scores of reads</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -37,7 +37,7 @@ #end for #end if ' - + #set $snp_dataset_provided = False #set $rod_binding_names = dict() #for $rod_binding in $rod_bind: @@ -52,9 +52,9 @@ #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1 -d "--knownSites:${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[$rod_bind_name]}" #end for - + #include source=$standard_gatk_options# - + ##start analysis specific options #if $analysis_param_type.analysis_param_type_selector == "advanced": -p ' @@ -73,10 +73,10 @@ ${analysis_param_type.exception_if_no_tile} #if str( $analysis_param_type.solid_options_type.solid_options_type_selector ) == "set": #if str( $analysis_param_type.solid_options_type.solid_recal_mode ) != "default": - --solid_recal_mode "${analysis_param_type.solid_options_type.solid_recal_mode}" + --solid_recal_mode "${analysis_param_type.solid_options_type.solid_recal_mode}" #end if #if str( $analysis_param_type.solid_options_type.solid_nocall_strategy ) != "default": - --solid_nocall_strategy "${analysis_param_type.solid_options_type.solid_nocall_strategy}" + --solid_nocall_strategy "${analysis_param_type.solid_options_type.solid_nocall_strategy}" #end if #end if --window_size_nqs "${analysis_param_type.window_size_nqs}" @@ -121,13 +121,13 @@ <option value="RepeatUnitCovariate" /> <option value="RepeatUnitAndLengthCovariate" /> <!-- - Note: ReadGroupCovariate and QualityScoreCovariate are required covariates and will + 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="ReadGroupCovariate" /> --> </param> - + <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-knownSites,--knownSites &lt;knownSites&gt;)"> <conditional name="rod_bind_type"> <param name="rod_bind_type_selector" type="select" label="Variant Type"> @@ -157,7 +157,7 @@ </repeat> <expand macro="gatk_param_type_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> @@ -248,7 +248,7 @@ <param name="covariates" value="ReadGroupCovariate,HomopolymerCovariate,MinimumNQSCovariate,PositionCovariate" /> <param name="gatk_param_type_selector" value="basic" /> <param name="analysis_param_type_selector" value="basic" /> - <output name="output_recal" file="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" /> + <output name="output_recal" file="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" /> <output name="output_log" file="gatk/gatk_count_covariates/gatk_count_covariates_out_1.log.contains" compare="contains" /> </test> </tests> @@ -257,7 +257,7 @@ "This calculation is critically dependent on being able to skip over known variant sites. Please provide a dbSNP ROD or a VCF file containing known sites of genetic variation." However, if you do not provide this file, the '--run_without_dbsnp_potentially_ruining_quality' flag will be automatically used, and the command will be allowed to run. - + **What it does** This walker is designed to work as the first pass in a two-pass processing step. It does a by-locus traversal operating only at sites that are not in dbSNP. We assume that all reference mismatches we see are therefore errors and indicative of poor base quality. This walker generates tables based on various user-specified covariates (such as read group, reported quality score, cycle, and dinucleotide) Since there is a large amount of data one can then calculate an empirical probability of error given the particular covariates seen at this site, where p(error) = num mismatches / num observations The output file is a CSV list of (the several covariate values, num observations, num mismatches, empirical quality score) The first non-comment line of the output file gives the name of the covariates that were used for this calculation. Note: ReadGroupCovariate and QualityScoreCovariate are required covariates and will be added for the user regardless of whether or not they were specified Note: This walker is designed to be used in conjunction with TableRecalibrationWalker.
--- a/depth_of_coverage.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/depth_of_coverage.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_depth_of_coverage" name="Depth of Coverage" version="@VERSION@.0"> <description>on BAM files</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python">gatk2_wrapper.py ##--max_jvm_heap_fraction "1" --stdout "${output_log}" @@ -33,13 +33,13 @@ #end for #end if --out "${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": @@ -85,7 +85,7 @@ && 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} @@ -104,7 +104,7 @@ && 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} @@ -123,7 +123,6 @@ && mv ${output_per_locus_coverage}.library_cumulative_coverage_proportions ${output_cumulative_coverage_proportions_library} #end if #end if - </command> <inputs> @@ -138,7 +137,7 @@ </repeat> <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;"> <options from_data_table="gatk2_picard_indexes"> - <!-- <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/> does not yet work in a repeat...--> + <!-- <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> @@ -150,27 +149,27 @@ <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" /> @@ -359,7 +358,7 @@ </conditional> </actions> </data> - + <data format="tabular" name="output_summary_readgroup" label="${tool.name} on ${on_string} (output summary readgroup)" > <filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter> <filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter> @@ -512,7 +511,7 @@ </conditional> </actions> </data> - + <data format="tabular" name="output_summary_library" label="${tool.name} on ${on_string} (output summary library)" > <filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['omit_per_sample_stats'] == False</filter> <filter>analysis_param_type['analysis_param_type_selector'] == "basic" or analysis_param_type['print_bin_endpoints_and_exit'] == False</filter> @@ -665,7 +664,7 @@ </conditional> </actions> </data> - + <data format="tabular" name="output_log" label="${tool.name} on ${on_string} (log)" /> </outputs> <trackster_conf/> @@ -691,7 +690,7 @@ <help> **What it does** -DepthOfCoverage processes a set of bam files to determine coverage at different levels of partitioning and aggregation. Coverage can be analyzed per locus, per interval, per gene, or in total; can be partitioned by sample, by read group, by technology, by center, or by library; and can be summarized by mean, median, quartiles, and/or percentage of bases covered to or beyond a threshold. Additionally, reads and bases can be filtered by mapping or base quality score. +DepthOfCoverage processes a set of bam files to determine coverage at different levels of partitioning and aggregation. Coverage can be analyzed per locus, per interval, per gene, or in total; can be partitioned by sample, by read group, by technology, by center, or by library; and can be summarized by mean, median, quartiles, and/or percentage of bases covered to or beyond a threshold. Additionally, reads and bases can be filtered by mapping or base quality score. For more information on the GATK Depth of Coverage, see this `tool specific page <http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_coverage_DepthOfCoverage.html>`_.
--- a/haplotype_caller.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/haplotype_caller.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_haplotype_caller" name="Haplotype Caller" version="@VERSION@.1"> <description>Call SNPs and indels simultaneously via local de-novo assembly of haplotypes in an active region</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -33,7 +33,7 @@ @DBSNP_OPTIONS@ $allow_n_cigar_reads #include source=$standard_gatk_options# - + ##start analysis specific options #if $analysis_param_type.analysis_param_type_selector == "advanced": -p ' @@ -55,7 +55,7 @@ #if str($analysis_param_type.comp) != 'None': --comp "$analysis_param_type.comp" #end if - ## + ## #if str( $analysis_param_type.annotation ) != "None": #for $annotation in str( $analysis_param_type.annotation.fields.gatk_value ).split( ','): --annotation "${annotation}" @@ -216,8 +216,6 @@ <param name="input_alleles_rod" type="data" format="vcf" label="Alleles ROD file" help="-alleles,--alleles &lt;alleles&gt;" /> </when> </conditional> - - <param name="graphOutput" type="boolean" checked="False" truevalue="" falsevalue="" label="graphOutput" help="--graphOutput / -graph File to which debug assembly graph information should be written"/> <param name="heterozygosity" type="float" value="0.0010" optional="true" label="heterozygosity" help="--heterozygosity / -hets Heterozygosity value used to compute prior likelihoods for any locus"/> <param name="minPruning" type="integer" value="1" optional="true" label="minPruning" help="--minPruning / -minPruning The minimum allowed pruning factor in assembly graph. Paths with >= X supporting kmers are pruned from the graph"> @@ -257,7 +255,7 @@ </outputs> <tests> <test> - <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" /> + <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" /> <param name="reference_source_selector" value="history" /> <param name="ref_file" value="phiX.fasta" ftype="fasta" /> <param name="input_bam" value="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" />
--- a/indel_realigner.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/indel_realigner.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_indel_realigner" name="Indel Realigner" version="@VERSION@.1"> <description>- perform local realignment</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -28,7 +28,7 @@ -LOD "${lod_threshold}" ${knowns_only} ' - + #set $rod_binding_names = dict() #for $rod_binding in $rod_bind: #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom': @@ -62,7 +62,7 @@ #end if </command> <inputs> - + <conditional name="reference_source"> <expand macro="reference_source_selector_param" /> <when value="cached"> @@ -115,9 +115,9 @@ <expand macro="allow_n_cigar_reads" /> <expand macro="gatk_param_type_conditional" /> - + <expand macro="analysis_type_conditional"> - + <param name="entropy_threshold" type="float" value="0.15" label="percentage of mismatching base quality scores at a position to be considered having high entropy" help="-entropy,--entropyThreshold &lt;entropyThreshold&gt;" /> <param name="simplify_bam" type="boolean" checked="False" truevalue="-simplifyBAM" falsevalue="" label="Simplify BAM" help="-simplifyBAM,--simplifyBAM"/> <param name="consensus_determination_model" type="select" label="Consensus Determination Model" help="-model,--consensusDeterminationModel &lt;consensusDeterminationModel&gt;"> @@ -130,7 +130,7 @@ <param name="max_consensuses" type="integer" value="30" label="Max alternate consensuses to try" help="-maxConsensuses,--maxConsensuses &lt;maxConsensuses&gt;" /> <param name="max_reads_for_consensuses" type="integer" value="120" label="Max reads (chosen randomly) used for finding the potential alternate consensuses" help="-greedy,--maxReadsForConsensuses &lt;maxReadsForConsensuses&gt;" /> <param name="max_reads_for_realignment" type="integer" value="20000" label="Max reads allowed at an interval for realignment" help="-maxReads,--maxReadsForRealignment &lt;maxReadsForRealignment&gt;" /> - <param name="no_original_alignment_tags" type="boolean" checked="False" truevalue="--noOriginalAlignmentTags" falsevalue="" label="Don't output the original cigar or alignment start tags for each realigned read in the output bam" help="-noTags,--noOriginalAlignmentTags"/> + <param name="no_original_alignment_tags" type="boolean" checked="False" truevalue="--noOriginalAlignmentTags" falsevalue="" label="Don't output the original cigar or alignment start tags for each realigned read in the output bam" help="-noTags,--noOriginalAlignmentTags"/> </expand> </inputs> <outputs> @@ -158,7 +158,7 @@ <param name="max_reads_for_consensuses" value="120" /> <param name="max_reads_for_realignment" value="20000" /> <param name="no_original_alignment_tags" /> - <output name="output_bam" file="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" lines_diff="2" /> + <output name="output_bam" file="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" lines_diff="2" /> <output name="output_log" file="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.log.contains" compare="contains" /> </test> </tests>
--- a/print_reads.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/print_reads.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_print_reads" name="Print Reads" version="@VERSION@.0"> <description>on BAM files</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -21,7 +21,6 @@ ## according to http://www.broadinstitute.org/gatk/guide/article?id=1975 --num_cpu_threads_per_data_thread \${GALAXY_SLOTS:-6} - #if $reference_source.reference_source_selector != "history": -R "${reference_source.ref_file.fields.path}" #end if @@ -30,9 +29,9 @@ #end if --disable_bam_indexing ' - + #include source=$standard_gatk_options# - + ##start analysis specific options #if $analysis_param_type.analysis_param_type_selector == "advanced": -p ' @@ -51,10 +50,10 @@ ${analysis_param_type.exception_if_no_tile} #if str( $analysis_param_type.solid_options_type.solid_options_type_selector ) == "set": #if str( $analysis_param_type.solid_options_type.solid_recal_mode ) != "default": - --solid_recal_mode "${analysis_param_type.solid_options_type.solid_recal_mode}" + --solid_recal_mode "${analysis_param_type.solid_options_type.solid_recal_mode}" #end if #if str( $analysis_param_type.solid_options_type.solid_nocall_strategy ) != "default": - --solid_nocall_strategy "${analysis_param_type.solid_options_type.solid_nocall_strategy}" + --solid_nocall_strategy "${analysis_param_type.solid_options_type.solid_nocall_strategy}" #end if #end if ${analysis_param_type.simplify_bam} @@ -68,7 +67,7 @@ #end if </command> <inputs> - <param name="input_recal" type="data" format="gatk_report" optional="true" label="Covariates table recalibration file" + <param name="input_recal" type="data" format="gatk_report" optional="true" label="Covariates table recalibration file" help="The input covariates table file which enables on-the-fly base quality score recalibration (intended for use with BaseRecalibrator files) (-BQSR,--BQSR)" /> <conditional name="reference_source"> <expand macro="reference_source_selector_param" /> @@ -93,10 +92,9 @@ </param> </when> </conditional> - + <expand macro="gatk_param_type_conditional" /> - - + <conditional name="analysis_param_type"> <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options"> <option value="basic" selected="True">Basic</option> @@ -183,7 +181,7 @@ </outputs> <tests> <test> - <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" /> + <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" /> <param name="reference_source_selector" value="history" /> <param name="ref_file" value="phiX.fasta" ftype="fasta" /> <param name="input_bam" value="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" />
--- a/realigner_target_creator.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/realigner_target_creator.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_realigner_target_creator" name="Realigner Target Creator" version="@VERSION@.1"> <description>for use in local realignment</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -75,12 +75,12 @@ </param> </when> </conditional> - + <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-known,--known &lt;known&gt;)"> <conditional name="rod_bind_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> @@ -104,11 +104,11 @@ <expand macro="gatk_param_type_conditional" /> <expand macro="analysis_type_conditional"> - <param name="windowSize" type="integer" value="10" label="Window size for calculating entropy or SNP clusters (windowSize)" + <param name="windowSize" type="integer" value="10" label="Window size for calculating entropy or SNP clusters (windowSize)" help="-window,--windowSize &lt;windowSize&gt;" /> - <param name="mismatchFraction" type="float" value="0.15" label="Fraction of base qualities needing to mismatch for a position to have high entropy (mismatchFraction)" + <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 <= 0 or > 1 (-mismatch,--mismatchFraction &lt;mismatchFraction&gt;)"/> - <param name="minReadsAtLocus" type="integer" value="4" label="Minimum reads at a locus to enable using the entropy calculation (minReadsAtLocus)" + <param name="minReadsAtLocus" type="integer" value="4" label="Minimum reads at a locus to enable using the entropy calculation (minReadsAtLocus)" help="-minReads,--minReadsAtLocus &lt;minReadsAtLocus&gt;" /> <param name="maxIntervalSize" type="integer" value="500" label="Maximum interval size" help="-maxInterval,--maxIntervalSize &lt;maxIntervalSize&gt;" /> </expand> @@ -130,7 +130,7 @@ <param name="mismatchFraction" value="0.15" /> <param name="minReadsAtLocus" value="4" /> <param name="maxIntervalSize" value="500" /> - <output name="output_interval" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.gatk_interval" /> + <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>
--- a/reduce_reads.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/reduce_reads.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_reduce_reads" name="Reduce Reads" version="@VERSION@.0"> <description>in BAM files</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -31,7 +31,7 @@ --disable_bam_indexing ' #include source=$standard_gatk_options# - + ##start analysis specific options #if $analysis_param_type.analysis_param_type_selector == "advanced": -p ' @@ -91,9 +91,9 @@ </param> </when> </conditional> - + <expand macro="gatk_param_type_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> @@ -137,7 +137,7 @@ </outputs> <tests> <test> - <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" /> + <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" /> <param name="reference_source_selector" value="history" /> <param name="ref_file" value="phiX.fasta" ftype="fasta" /> <param name="input_bam" value="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" />
--- a/tool_dependencies.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/tool_dependencies.xml Mon May 04 22:47:06 2015 -0400 @@ -3,11 +3,11 @@ <set_environment version="1.0"> <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">
--- a/unified_genotyper.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/unified_genotyper.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_unified_genotyper" name="Unified Genotyper" version="@VERSION@.1"> <description>SNP and indel caller</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -97,7 +97,7 @@ </repeat> <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;"> <options from_data_table="gatk2_picard_indexes"> - <!-- <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/> does not yet work in a repeat...--> + <!-- <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> @@ -111,23 +111,23 @@ </when> </conditional> <expand macro="dbsnp_param" /> - + <param name="genotype_likelihoods_model" type="select" label="Genotype likelihoods calculation model to employ" help="-glm,--genotype_likelihoods_model &lt;genotype_likelihoods_model&gt;"> <option value="BOTH" selected="True">BOTH</option> <option value="SNP">SNP</option> <option value="INDEL">INDEL</option> </param> - + <param name="standard_min_confidence_threshold_for_calling" type="float" value="30.0" label="The minimum phred-scaled confidence threshold at which variants not at 'trigger' track sites should be called" help="-stand_call_conf,--standard_min_confidence_threshold_for_calling &lt;standard_min_confidence_threshold_for_calling&gt;" /> <param name="standard_min_confidence_threshold_for_emitting" type="float" value="30.0" label="The minimum phred-scaled confidence threshold at which variants not at 'trigger' track sites should be emitted (and filtered if less than the calling threshold)" help="-stand_emit_conf,--standard_min_confidence_threshold_for_emitting &lt;standard_min_confidence_threshold_for_emitting&gt;" /> <expand macro="allow_n_cigar_reads" /> <expand macro="gatk_param_type_conditional" /> - + <expand macro="analysis_type_conditional"> - <param name="heterozygosity" type="float" value="1e-3" label="Heterozygosity value used to compute prior likelihoods for any locus" + <param name="heterozygosity" type="float" value="1e-3" label="Heterozygosity value used to compute prior likelihoods for any locus" help="-hets,--heterozygosity &lt;heterozygosity&gt;" /> - <param name="pcr_error_rate" type="float" value="1e-4" label="The PCR error rate to be used for computing fragment-based likelihoods" + <param name="pcr_error_rate" type="float" value="1e-4" label="The PCR error rate to be used for computing fragment-based likelihoods" help="-pcr_error,--pcr_error_rate &lt;pcr_error_rate&gt;" /> <conditional name="genotyping_mode_type"> <param name="genotyping_mode" type="select" label="How to determine the alternate allele to use for genotyping" help="-gt_mode,--genotyping_mode &lt;genotyping_mode&gt;"> @@ -194,7 +194,7 @@ <!-- <option value="none">none</option> --> </param> <!-- <param name="family_string" type="text" value="" label="Family String"/> --> - <param name="exclude_annotations" type="select" multiple="True" display="checkboxes" label="Annotations to exclude" + <param name="exclude_annotations" type="select" multiple="True" display="checkboxes" label="Annotations to exclude" help="-XA,--excludeAnnotation &lt;excludeAnnotation&gt;" > <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs --> <options from_data_table="gatk2_annotations"> @@ -202,7 +202,7 @@ <filter type="static_value" value="UnifiedGenotyper" column="tools_valid_for"/> </options> </param> - <param name="sample_ploidy" type="integer" value="2" + <param name="sample_ploidy" type="integer" value="2" label="Ploidy (number of chromosomes) per sample. For pooled data, set to (Number of samples in each pool * Sample Ploidy)" help="-ploidy,--sample_ploidy" /> </expand> </inputs> @@ -240,8 +240,8 @@ <param name="doContextDependentGapPenalties" /> <!-- <param name="annotation" value="" /> <param name="group" value="" /> --> - <output name="output_vcf" file="gatk/gatk_unified_genotyper/gatk_unified_genotyper_out_1.vcf" lines_diff="4" /> - <output name="output_metrics" file="gatk/gatk_unified_genotyper/gatk_unified_genotyper_out_1.metrics" /> + <output name="output_vcf" file="gatk/gatk_unified_genotyper/gatk_unified_genotyper_out_1.vcf" lines_diff="4" /> + <output name="output_metrics" file="gatk/gatk_unified_genotyper/gatk_unified_genotyper_out_1.metrics" /> <output name="output_log" file="gatk/gatk_unified_genotyper/gatk_unified_genotyper_out_1.log.contains" compare="contains" /> </test> </tests>
--- a/variant_annotator.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_annotator.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_annotator" name="Variant Annotator" version="@VERSION@.0"> <description></description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -47,30 +47,30 @@ #if $reference_source.input_variant_bti: -d "--intervals" "${reference_source.input_variant}" "${reference_source.input_variant.ext}" "input_variant_bti" #end if - + #for $rod_binding in $comp_rod_bind: -d "--comp:${rod_binding.comp_rod_name},%(file_type)s" "${rod_binding.comp_input_rod}" "${rod_binding.comp_input_rod.ext}" "input_comp_${rod_binding.comp_rod_name}" #end for - + @DBSNP_OPTIONS@ - + #for $rod_binding in $resource_rod_bind: -d "--resource:${rod_binding.resource_rod_name},%(file_type)s" "${rod_binding.resource_input_rod}" "${rod_binding.resource_input_rod.ext}" "input_resource_${rod_binding.resource_rod_name}" #end for - + #if str( $snpEff_rod_bind_type.snpEff_rod_bind_type_selector ) == 'set_snpEff': -p '--annotation "SnpEff"' -d "--snpEffFile:${snpEff_rod_bind_type.snpEff_rod_name},%(file_type)s" "${snpEff_rod_bind_type.snpEff_input_rod}" "${snpEff_rod_bind_type.snpEff_input_rod.ext}" "input_snpEff_${snpEff_rod_bind_type.snpEff_rod_name}" #else: -p '--excludeAnnotation "SnpEff"' #end if - + #for $expression in $expressions: -p '--expression "${expression.expression}"' #end for - + #include source=$standard_gatk_options# - + -p ' #if str( $annotation_group ) != "None": #for $group in str( $annotation_group ).split( ',' ): @@ -125,22 +125,22 @@ </param> </when> </conditional> - + <repeat name="additional_annotations" title="Additional annotation" help="-A,--annotation &lt;annotation&gt;"> <param name="additional_annotation_name" type="text" value="" label="Annotation name" /> </repeat> - + <repeat name="comp_rod_bind" title="Binding for reference-ordered comparison data" help="-comp,--comp &lt;comp&gt;"> <param name="comp_input_rod" type="data" format="vcf" label="ROD file" /> <param name="comp_rod_name" type="text" value="Unnamed" label="ROD Name"/> </repeat> <expand macro="dbsnp_param" /> - + <repeat name="resource_rod_bind" title="Binding for reference-ordered resource data" help="-resource,--resource &lt;resource&gt;"> <param name="resource_input_rod" type="data" format="vcf" label="ROD file" /> <param name="resource_rod_name" type="text" value="Unnamed" label="ROD Name"/> </repeat> - + <conditional name="snpEff_rod_bind_type"> <param name="snpEff_rod_bind_type_selector" type="select" label="Provide a snpEff reference-ordered data file (VCF)" help="-snpEffFile,--snpEffFile &lt;snpEffFile&gt;"> <option value="set_snpEff">Set snpEff</option> @@ -154,13 +154,13 @@ <param name="snpEff_rod_name" type="hidden" value="snpEff" label="ROD Name"/> </when> </conditional> - + <repeat name="expressions" title="Expression" help="-E,--expression &lt;expression&gt;"> <param name="expression" type="text" value="" label="Expression"/> </repeat> - + <expand macro="gatk_param_type_conditional" /> - + <param name="annotation_group" type="select" multiple="True" display="checkboxes" label="annotation interfaces/groups to apply to variant calls" help="-G,--group &lt;group&gt;"> <option value="RodRequiringAnnotation">RodRequiringAnnotation</option> <option value="Standard">Standard</option> @@ -177,7 +177,7 @@ <filter type="static_value" value="VariantAnnotator" column="tools_valid_for"/> </options> </param> - + </inputs> <outputs> <data format="vcf" name="output_vcf" label="${tool.name} on ${on_string} (Variant File)" /> @@ -198,7 +198,7 @@ <param name="dbsnp_rod_name" value="dbsnp" /> <param name="snpEff_rod_bind_type_selector" value="exclude_snpEff" /> <param name="gatk_param_type_selector" value="basic" /> - <output name="output_vcf" file="gatk/gatk_variant_annotator/gatk_variant_annotator_out_1.vcf" lines_diff="4" /> + <output name="output_vcf" file="gatk/gatk_variant_annotator/gatk_variant_annotator_out_1.vcf" lines_diff="4" /> <output name="output_log" file="gatk/gatk_variant_annotator/gatk_variant_annotator_out_1.log.contains" compare="contains" /> <param name="comp_rod_bind" value="0" /> <param name="resource_rod_bind" value="0" />
--- a/variant_apply_recalibration.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_apply_recalibration.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_apply_recalibration" name="Apply Variant Recalibration" version="@VERSION@.0"> <description></description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -25,13 +25,13 @@ --tranches_file "${reference_source.input_tranches}" --out "${output_variants}" ' - + #include source=$standard_gatk_options# - + ##start analysis specific options -p ' --mode "${mode}" - + #for $ignore_filter in $ignore_filters: #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.ignore_filter_type_selector ) #if $ignore_filter_name == "custom": @@ -67,9 +67,9 @@ <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> </when> </conditional> - + <expand macro="gatk_param_type_conditional" /> - + <param name="mode" type="select" label="Recalibration mode" help="-mode,--mode &lt;mode&gt;"> <option value="SNP" selected="True">SNP</option> <option value="INDEL">INDEL</option>
--- a/variant_combine.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_combine.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_combine" name="Combine Variants" version="@VERSION@.0"> <description></description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -30,10 +30,9 @@ --genotypemergeoption "${genotype_merge_option}" --rod_priority_list "${ ','.join( $priority_order ) }" ' - + #include source=$standard_gatk_options# - - + ##start analysis specific options #if $analysis_param_type.analysis_param_type_selector == "advanced": -p ' @@ -42,17 +41,17 @@ ${analysis_param_type.filtered_are_uncalled} ${analysis_param_type.minimal_vcf} ${analysis_param_type.assume_identical_samples} - + #if str( $analysis_param_type.set_key ): --setKey "${analysis_param_type.set_key}" #end if - + --minimumN "${analysis_param_type.minimum_n}" ' #end if </command> <inputs> - + <conditional name="reference_source"> <expand macro="reference_source_selector_param" /> <when value="cached"> @@ -79,34 +78,31 @@ <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> </when> </conditional> - + <param name="genotype_merge_option" type="select" label="How should we merge genotype records across records for samples shared across the ROD files" help="-genotypeMergeOptions,--genotypemergeoption &lt;genotypemergeoption&gt;" > <option value="UNIQUIFY" /> <option value="PRIORITIZE" selected="true"/> <option value="UNSORTED" /> <option value="REQUIRE_UNIQUE" /> </param> - + <expand macro="gatk_param_type_conditional" /> - - + <expand macro="analysis_type_conditional"> <param name="filtered_records_merge_type" type="select" label="How should we deal with records seen at the same site in the VCF, but with different FILTER fields?" help="-filteredRecordsMergeType,--filteredrecordsmergetype &lt;filteredrecordsmergetype&gt;" > <option value="KEEP_IF_ANY_UNFILTERED" selected="true"/> <option value="KEEP_IF_ALL_UNFILTERED" /> </param> - + <param name="print_complex_merges" checked="false" type="boolean" truevalue="--printComplexMerges" falsevalue="" label="Print out interesting sites requiring complex compatibility merging" help="-printComplexMerges,--printComplexMerges" /> <param name="filtered_are_uncalled" checked="false" type="boolean" truevalue="--filteredAreUncalled" falsevalue="" label="If true, then filtered VCFs are treated as uncalled, so that filtered set annotation don't appear in the combined VCF" help="-filteredAreUncalled,--filteredAreUncalled" /> <param name="minimal_vcf" checked="false" type="boolean" truevalue="--minimalVCF" falsevalue="" label="If true, then the output VCF will contain no INFO or genotype INFO field" help="-minimalVCF,--minimalVCF" /> - + <param name="set_key" type="text" value="" label="Key, by default set, in the INFO key=value tag emitted describing which set the combined VCF record came from." help="-setKey,--setKey &lt;setKey&gt;"/> <param name="assume_identical_samples" checked="false" type="boolean" truevalue="--assumeIdenticalSamples" falsevalue="" label="If true, assume input VCFs have identical sample sets and disjoint calls so that one can simply perform a merge sort to combine the VCFs into one, drastically reducing the runtime." help="-assumeIdenticalSamples,--assumeIdenticalSamples" /> <param name="minimum_n" type="integer" value="1" label="Combine variants and output site only if variant is present in at least N input files." help="-minN,--minimumN &lt;minimumN&gt;"/> - </expand> - - + </inputs> <outputs> <data format="vcf" name="output_variants" label="${tool.name} on ${on_string} (variants)" />
--- a/variant_eval.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_eval.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_eval" name="Eval Variants" version="@VERSION@.0"> <description></description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> #from binascii import hexlify @@ -25,24 +25,23 @@ -R "${reference_source.ref_file.fields.path}" #end if ' - + #for $rod_binding in $comp_rod_bind: -d "--comp:${rod_binding.comp_rod_name},%(file_type)s" "${rod_binding.comp_input_rod}" "${rod_binding.comp_input_rod.ext}" "input_comp_${rod_binding.comp_rod_name}" #if str( $rod_binding.comp_known_names ): -p '--known_names "${rod_binding.comp_rod_name}"' #end if #end for - + #if $dbsnp_rod_bind_type.dbsnp_rod_bind_type_selector == 'set_dbsnp' -d "--dbsnp:${dbsnp_rod_bind_type.dbsnp_rod_name},%(file_type)s" "${dbsnp_rod_bind_type.dbsnp_input_rod}" "${dbsnp_rod_bind_type.dbsnp_input_rod.ext}" "input_dbsnp_${dbsnp_rod_bind_type.dbsnp_rod_name}" #if $dbsnp_rod_bind_type.dbsnp_known_names -p '--known_names "${dbsnp_rod_bind_type.dbsnp_rod_name}"' #end if #end if - + #include source=$standard_gatk_options# - - + ##start analysis specific options #if $analysis_param_type.analysis_param_type_selector == "advanced": #for $stratification in $analysis_param_type.stratifications: @@ -50,39 +49,39 @@ -o '${ hexlify( $select_string ) }' #end for -p ' - + #for $sample in $analysis_param_type.samples: --sample "${sample.sample}" #end for - + #if str( $analysis_param_type.stratification_modules ) != "None": #for $stratification_module in str( $analysis_param_type.stratification_modules).split( ',' ): --stratificationModule "${stratification_module}" #end for #end if - + ${analysis_param_type.do_not_use_all_standard_stratifications} - + #for $variant_type in $analysis_param_type.only_variants_of_type: --onlyVariantsOfType "${variant_type.variant_type}" #end for - + #if str( $analysis_param_type.eval_modules ) != "None": #for $eval_module in str( $analysis_param_type.eval_modules).split( ',' ): --evalModule "${eval_module}" #end for #end if - + ${analysis_param_type.do_not_use_all_standard_modules} - + #if str( $analysis_param_type.num_samples ) != "0": --numSamples "${analysis_param_type.num_samples}" #end if - + --minPhaseQuality "${analysis_param_type.min_phase_quality}" - + --mendelianViolationQualThreshold "${analysis_param_type.mendelian_violation_qual_threshold}" - + #if str( $analysis_param_type.ancestral_alignments ) != "None": --ancestralAlignments "${analysis_param_type.ancestral_alignments}" #end if @@ -90,14 +89,14 @@ #if str( $analysis_param_type.known_cnvs ) != "None": -d "--knownCNVs" "${analysis_param_type.known_cnvs}" "${analysis_param_type.known_cnvs.ext}" "input_known_cnvs" #end if - + #if str( $analysis_param_type.strat_intervals ) != "None": -d "--stratIntervals" "${analysis_param_type.strat_intervals}" "${analysis_param_type.strat_intervals.ext}" "input_strat_intervals" #end if #end if </command> <inputs> - + <conditional name="reference_source"> <expand macro="reference_source_selector_param" /> <when value="cached"> @@ -118,7 +117,7 @@ <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> </when> </conditional> - + <repeat name="comp_rod_bind" title="Comparison Reference-Ordered Data (ROD) file" help="-comp,--comp &lt;comp&gt;"> <param name="comp_input_rod" type="data" format="vcf" label="Comparison ROD file" /> <param name="comp_rod_name" type="text" value="" label="Comparison ROD name"> @@ -140,10 +139,9 @@ <param name="dbsnp_known_names" type="boolean" label="Use dbSNP ROD file as known_names" help="-knownName,--known_names &lt;known_names&gt;" /> </when> </conditional> - + <expand macro="gatk_param_type_conditional" /> - - + <expand macro="analysis_type_conditional"> <repeat name="stratifications" title="Stratification"> <param name="select_exps" value="" type="text" label="Stratification Expression" help="-select,--select_exps &lt;select_exps&gt;"> @@ -156,11 +154,11 @@ </param> <param name="select_name" value="" type="text" label="Name" help="-selectName,--select_names &lt;select_names&gt;"/> </repeat> - + <repeat name="samples" title="Sample" help="-sn,--sample &lt;sample&gt;"> <param name="sample" value="" type="text" label="Derive eval and comp contexts using only these sample genotypes, when genotypes are available in the original context"/> </repeat> - + <param name="stratification_modules" type="select" multiple="True" display="checkboxes" label="Stratification modules to apply to the eval track(s)" help="-ST,--stratificationModule &lt;stratificationModule&gt;" > <option value="AlleleCount" /> <option value="AlleleFrequency" /> @@ -182,11 +180,11 @@ <option value="VariantType" /> </param> <param name="do_not_use_all_standard_stratifications" checked="false" type="boolean" truevalue="--doNotUseAllStandardStratifications" falsevalue="" label="Do not use the standard stratification modules by default" help="-noST,--doNotUseAllStandardStratifications" /> - + <repeat name="only_variants_of_type" title="only Variants Of Type" help="--onlyVariantsOfType"> <param name="variant_type" type="text" value="" label="only variants of these types will be considered during the evaluation"/> </repeat> - + <param name="eval_modules" type="select" multiple="True" display="checkboxes" label="Eval modules to apply to the eval track(s)" help="-EV,--evalModule &lt;evalModule&gt;" > <option value="CompOverlap" /> <option value="CountVariants" /> @@ -201,17 +199,15 @@ <option value="VariantSummary" /> </param> <param name="do_not_use_all_standard_modules" checked="false" type="boolean" truevalue="--doNotUseAllStandardModules" falsevalue="" label="Do not use the standard eval modules by default" help="-noEV,--doNotUseAllStandardModules" /> - + <param name="num_samples" type="integer" label="Number of samples (used if no samples are available in the VCF file" value="0" help="-ns,--numSamples &lt;numSamples&gt;"/> <param name="min_phase_quality" type="float" label="Minimum phasing quality " value="10.0" help="-mpq,--minPhaseQuality &lt;minPhaseQuality&gt;"/> <param name="mendelian_violation_qual_threshold" type="integer" label="Minimum genotype QUAL score for each trio member required to accept a site as a violation" value="50" help="-mvq,--mendelianViolationQualThreshold &lt;mendelianViolationQualThreshold&gt;"/> <param name="ancestral_alignments" type="data" format="fasta" optional="True" label="Fasta file with ancestral alleles" help="-aa,--ancestralAlignments &lt;ancestralAlignments&gt;" /> <param name="known_cnvs" type="data" format="bed,gatk_interval,picard_interval_list" optional="True" label="File containing tribble-readable features describing a known list of copy number variants" help="-knownCNVs,--knownCNVs &lt;knownCNVs&gt;" /> <param name="strat_intervals" type="data" format="bed,gatk_interval,picard_interval_list" optional="True" label="File containing tribble-readable features for the IntervalStratificiation" help="-stratIntervals,--stratIntervals &lt;stratIntervals&gt;" /> - </expand> - - + </inputs> <outputs> <data format="gatk_report" name="output_report" label="${tool.name} on ${on_string} (report)" /> @@ -270,7 +266,7 @@ known_names Name of ROD bindings containing variant sites that should be treated as known when splitting eval rods into known and novel subsets stratificationModule One or more specific stratification modules to apply to the eval track(s) (in addition to the standard stratifications, unless -noS is specified) doNotUseAllStandardStratifications Do not use the standard stratification modules by default (instead, only those that are specified with the -S option) - onlyVariantsOfType If provided, only variants of these types will be considered during the evaluation, in + onlyVariantsOfType If provided, only variants of these types will be considered during the evaluation, in evalModule One or more specific eval modules to apply to the eval track(s) (in addition to the standard modules, unless -noE is specified) doNotUseAllStandardModules Do not use the standard modules by default (instead, only those that are specified with the -E option) numSamples Number of samples (used if no samples are available in the VCF file
--- a/variant_filtration.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_filtration.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_filtration" name="Variant Filtration" version="@VERSION@.0"> <description>on VCF files</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> #from binascii import hexlify @@ -26,7 +26,7 @@ #set $variant_filter = "--%sExpression '%s' --%sName '%s'" % ( str( $variant_filter.is_genotype_filter ), str( $variant_filter.filter_expression ), str( $variant_filter.is_genotype_filter ), str( $variant_filter.filter_name ) ) -o '${ hexlify( $variant_filter ) }' #end for - + #if str( $mask_rod_bind_type.mask_rod_bind_type_selector ) == 'set_mask': -d "--mask:${mask_rod_bind_type.mask_rod_name},%(file_type)s" "${mask_rod_bind_type.input_mask_rod}" "${mask_rod_bind_type.input_mask_rod.ext}" "input_mask_${mask_rod_bind_type.mask_rod_name}" -p ' @@ -34,9 +34,9 @@ --maskName "${mask_rod_bind_type.mask_rod_name}" ' #end if - + #include source=$standard_gatk_options# - + ##start analysis specific options #if $cluster_snp_type.cluster_snp_type_selector == "cluster_snp": -p ' @@ -63,8 +63,7 @@ <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> </when> </conditional> - - + <repeat name="variant_filters" title="Variant Filters"> <param name="filter_expression" value="AB < 0.2 || MQ0 > 50" type="text" label="Filter expression" help="JEXL formatted expressions (-filter,--filterExpression &lt;filterExpression&gt;)"> <sanitizer> @@ -77,9 +76,7 @@ <param name="filter_name" value="custom_filter" type="text" label="Filter name" help="-filterName,--filterName &lt;filterName&gt;"/> <param name="is_genotype_filter" type="boolean" truevalue="genotypeFilter" falsevalue="filter" label="Use filter at the individual sample level" help="Use -G_filter,--genotypeFilterExpression &lt;genotypeFilterExpression&gt; and -G_filterName,--genotypeFilterName &lt;genotypeFilterName&gt; for filter type" /> </repeat> - - <conditional name="mask_rod_bind_type"> <param name="mask_rod_bind_type_selector" type="select" label="Provide a Mask reference-ordered data file"> <option value="set_mask" selected="True">Set mask</option> @@ -93,11 +90,10 @@ <param name="mask_rod_name" type="text" value="Mask" label="Mask Name" help="-maskName,--maskName &lt;maskName&gt;"/> <param name="mask_extension" type="integer" value="0" label="Mask Extension" help="-maskExtend,--maskExtension &lt;maskExtension&gt;"/> </when> - </conditional> - - + </conditional> + <expand macro="gatk_param_type_conditional" /> - + <conditional name="cluster_snp_type"> <param name="cluster_snp_type_selector" type="select" label="Cluster SNPs"> <option value="cluster_snp">Cluster SNPs</option> @@ -111,9 +107,9 @@ <param name="cluster_window_size" type="integer" value="0" label="The window size (in bases) in which to evaluate clustered SNPs" help="-window,--clusterWindowSize &lt;clusterWindowSize&gt;"/> </when> </conditional> - + <param name="missing_values_in_expressions_should_evaluate_as_failing" type="boolean" truevalue="--missingValuesInExpressionsShouldEvaluateAsFailing" falsevalue="" label="Should missing values be considered failing the expression" help="--missingValuesInExpressionsShouldEvaluateAsFailing" /> - + </inputs> <outputs> <data format="vcf" name="output_vcf" label="${tool.name} on ${on_string} (Variant File)" /> @@ -134,7 +130,7 @@ <param name="gatk_param_type_selector" value="basic" /> <param name="cluster_snp_type_selector" value="do_not_cluster_snp" /> <param name="missing_values_in_expressions_should_evaluate_as_failing" /> - <output name="output_vcf" file="gatk/gatk_variant_annotator/gatk_variant_annotator_out_1.vcf" lines_diff="4" /> + <output name="output_vcf" file="gatk/gatk_variant_annotator/gatk_variant_annotator_out_1.vcf" lines_diff="4" /> <output name="output_log" file="gatk/gatk_variant_filtration/gatk_variant_filtration_out_1.log.contains" compare="contains" /> </test> </tests>
--- a/variant_recalibrator.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_recalibrator.xml Mon May 04 22:47:06 2015 -0400 @@ -1,12 +1,12 @@ <tool id="gatk2_variant_recalibrator" name="Variant Recalibrator" version="@VERSION@.0"> <description></description> + <macros> + <import>gatk2_macros.xml</import> + </macros> <expand macro="requirements"> <requirement type="package" version="0.9.3">ggplot2</requirement> </expand> <expand macro="version_command" /> - <macros> - <import>gatk2_macros.xml</import> - </macros> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -27,14 +27,14 @@ --tranches_file "${output_tranches}" --rscript_file "${output_rscript}" ' - + #set $rod_binding_names = dict() #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 #elif str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'comp': #set $rod_bind_name = "comp" + $rod_binding.rod_bind_type.custom_rod_name - #else + #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( $rod_bind_name, -1 ) + 1 @@ -44,9 +44,9 @@ -d "--resource:${rod_bind_name},%(file_type)s,known=${rod_binding.rod_bind_type.rod_training_type.known},training=${rod_binding.rod_bind_type.rod_training_type.training},truth=${rod_binding.rod_bind_type.rod_training_type.truth},bad=${rod_binding.rod_bind_type.rod_training_type.bad},prior=${rod_binding.rod_bind_type.rod_training_type.prior}" "${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]}" #end if #end for - + #include source=$standard_gatk_options# - + ##start analysis specific options -p ' #if str( $annotations ) != "None": @@ -59,7 +59,7 @@ #end for --mode "${mode}" ' - + #if $analysis_param_type.analysis_param_type_selector == "advanced": -p ' --maxGaussians "${analysis_param_type.max_gaussians}" @@ -85,11 +85,10 @@ #end for ' #end if - - + && mv "${output_rscript}.pdf" "${output_tranches_pdf}" - + </command> <inputs> <conditional name="reference_source"> @@ -112,7 +111,7 @@ <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> </when> </conditional> - + <repeat name="rod_bind" title="Binding for reference-ordered data" help="-resource,--resource &lt;resource&gt;" min="2"> <conditional name="rod_bind_type"> <param name="rod_bind_type_selector" type="select" label="Binding Type"> @@ -183,7 +182,7 @@ <param name="prior" type="float" label="prior probability of being true" value="12.0"/> </when> </conditional> - </when> + </when> <when value="dbsnp"> <param name="input_rod" type="data" format="vcf" label="ROD file" /> <conditional name="rod_training_type"> @@ -301,7 +300,7 @@ </when> </conditional> </repeat> - + <param name="annotations" type="select" multiple="True" display="checkboxes" label="annotations which should used for calculations" help="-an,--use_annotation &lt;use_annotation&gt;"> <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs --> <options from_data_table="gatk2_annotations"> @@ -309,19 +308,19 @@ <filter type="static_value" value="VariantRecalibrator" column="tools_valid_for"/> </options> </param> - + <repeat name="additional_annotations" title="Additional annotation" help="-an,--use_annotation &lt;use_annotation&gt;"> <param name="additional_annotation_name" type="text" value="" label="Annotation name" /> </repeat> - + <param name="mode" type="select" label="Recalibration mode" help="-mode,--mode &lt;mode&gt;"> <option value="SNP" selected="True">SNP</option> <option value="INDEL">INDEL</option> <option value="BOTH">BOTH</option> </param> - + <expand macro="gatk_param_type_conditional" /> - + <expand macro="analysis_type_conditional"> <param name="max_gaussians" type="integer" label="maximum number of Gaussians to try during variational Bayes Algorithm" value="8" help="-mG,--maxGaussians &lt;maxGaussians&gt;"/> <param name="max_iterations" type="integer" label="maximum number of maximum number of VBEM iterations to be performed in variational Bayes Algorithm" value="150" help="-mI,--maxIterations &lt;maxIterations&gt;"/>
--- a/variant_select.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_select.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_select" name="Select Variants" version="@VERSION@.0"> <description>from VCF files</description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> #from binascii import hexlify @@ -18,7 +18,7 @@ @THREADS@ -o "${output_vcf}" - + #if $reference_source.reference_source_selector != "history": -R "${reference_source.ref_file.fields.path}" #end if @@ -30,74 +30,72 @@ #if $input_discordance: --discordance "${input_discordance}" #end if - + #for $exclude_sample_name in $exclude_sample_name_repeat: --exclude_sample_name "${exclude_sample_name.exclude_sample_name}" #end for - + ${exclude_filtered} - + #for $sample_name in $sample_name_repeat: --sample_name "${sample_name.sample_name}" #end for ' - + #for $select_expressions in $select_expressions_repeat: #set $select_expression = "--select_expressions '%s'" % ( str( $select_expressions.select_expressions ) ) -o '${ hexlify( $select_expression ) }' #end for - + ##start tool specific options #if str( $analysis_param_type.analysis_param_type_selector ) == 'advanced': -p ' #for $exclude_sample_file in $analysis_param_type.exclude_sample_file_repeat: --exclude_sample_file "${exclude_sample_file.exclude_sample_file}" #end for - + #for $sample_file in $analysis_param_type.sample_file_repeat: --sample_file "${ample_file.sample_file}" #end for - + #if $analysis_param_type.input_keep_ids: --keepIDs "${analysis_param_type.input_keep_ids}" #end if - + ${analysis_param_type.keep_original_AC} - + ${analysis_param_type.mendelian_violation} - + --mendelianViolationQualThreshold "${analysis_param_type.mendelian_violation_qual_threshold}" - + --remove_fraction_genotypes "${analysis_param_type.remove_fraction_genotypes}" - + --restrictAllelesTo "${analysis_param_type.restrict_alleles_to}" - + #if str( $analysis_param_type.select_random_type.select_random_type_selector ) == 'select_random_fraction': --select_random_fraction "${analysis_param_type.select_random_type.select_random_fraction}" #elif str( $analysis_param_type.select_random_type.select_random_type_selector ) == 'select_random_number': --select_random_number "${analysis_param_type.select_random_type.select_random_number}" #end if - + #if $analysis_param_type.select_type_to_include: #for $type_to_include in str( $analysis_param_type.select_type_to_include ).split( ',' ): --selectTypeToInclude "${type_to_include}" #end for #end if - + ${analysis_param_type.exclude_non_variants} ' - + #for $sample_expressions in $analysis_param_type.sample_expressions_repeat: #set $sample_expression = "--sample_expressions '%s'" % ( str( $sample_expressions.sample_expressions ) ) -o '${ hexlify( $sample_expression ) }' #end for - + #end if ##end tool specific options - + #include source=$standard_gatk_options# - - </command> <inputs> <conditional name="reference_source"> @@ -116,7 +114,7 @@ <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> </when> </conditional> - + <repeat name="select_expressions_repeat" title="Criteria to use when selecting the data" help="-select,--select_expressions &lt;select_expressions&gt;"> <param name="select_expressions" type="text" label="JEXL expression"> <sanitizer> @@ -127,49 +125,48 @@ </sanitizer> </param> </repeat> - + <param name="input_concordance" type="data" format="vcf" label="Output variants that were also called in this comparison track" optional="True" help="-conc,--concordance &lt;concordance&gt;"/> <param name="input_discordance" type="data" format="vcf" label="Output variants that were not called in this comparison track" optional="True" help="-disc,--discordance &lt;discordance&gt;"/> - + <repeat name="sample_name_repeat" title="Include Samples by name" help="-sn,--sample_name &lt;sample_name&gt;"> <param name="sample_name" type="text" label="Include genotypes from this sample"/> </repeat> - + <repeat name="exclude_sample_name_repeat" title="Exclude Samples by name" help="-xl_sn,--exclude_sample_name &lt;exclude_sample_name&gt;"> <param name="exclude_sample_name" type="text" label="Exclude genotypes from this sample"/> </repeat> - + <param name="exclude_filtered" type="boolean" truevalue="--excludeFiltered" falsevalue="" label="Don't include filtered loci in the analysis" help="-ef,--excludeFiltered" /> - + <expand macro="gatk_param_type_conditional" /> - - + <expand macro="analysis_type_conditional"> - + <repeat name="exclude_sample_file_repeat" title="Exclude Samples by file" help="-xl_sf,--exclude_sample_file &lt;exclude_sample_file&gt;"> <param name="exclude_sample_file" type="data" format="txt" label="File containing a list of samples (one per line) to exclude"/> </repeat> - + <repeat name="sample_file_repeat" title="Samples by file" help="-sf,--sample_file &lt;sample_file&gt;"> <param name="sample_file" type="data" format="txt" label="File containing a list of samples (one per line) to include" /> </repeat> - + <param name="input_keep_ids" type="data" format="text" label="Only emit sites whose ID is found in this file" optional="True" help="-IDs,--keepIDs &lt;keepIDs&gt;"/> - + <param name="keep_original_AC" type="boolean" truevalue="--keepOriginalAC" falsevalue="" label="Don't update the AC, AF, or AN values in the INFO field after selecting" help="-keepOriginalAC,--keepOriginalAC" /> - + <param name="mendelian_violation" type="boolean" truevalue="--mendelianViolation" falsevalue="" label="output mendelian violation sites only" help="-mv,--mendelianViolation" /> - + <param name="mendelian_violation_qual_threshold" type="float" label="Minimum genotype QUAL score for each trio member required to accept a site as a mendelian violation" value="0" help="-mvq,--mendelianViolationQualThreshold &lt;mendelianViolationQualThreshold&gt;" /> - + <param name="remove_fraction_genotypes" type="float" label="Selects a fraction (a number between 0 and 1) of the total genotypes at random from the variant track and sets them to nocall" value="0" min="0" max="1" help="-fractionGenotypes,--remove_fraction_genotypes &lt;remove_fraction_genotypes&gt;" /> - + <param name="restrict_alleles_to" type="select" label="Select only variants of a particular allelicity" help="-restrictAllelesTo,--restrictAllelesTo &lt;restrictAllelesTo&gt;"> <option value="ALL" selected="True">ALL</option> <option value="MULTIALLELIC">MULTIALLELIC</option> <option value="BIALLELIC">BIALLELIC</option> </param> - + <repeat name="sample_expressions_repeat" title="Regular expression to select many samples from the ROD tracks provided" help="-se,--sample_expressions &lt;sample_expressions&gt;"> <param name="sample_expressions" type="text" label="Regular expression"> <sanitizer> @@ -180,7 +177,7 @@ </sanitizer> </param> </repeat> - + <conditional name="select_random_type"> <param name="select_random_type_selector" type="select" label="Select a random subset of variants"> <option value="select_all" selected="True">Use all variants</option> @@ -197,9 +194,9 @@ <param name="select_random_number" type="integer" value="0" label="Count" help="-number,--select_random_number &lt;select_random_number&gt;" /> </when> </conditional> - + <param name="exclude_non_variants" type="boolean" truevalue="--excludeNonVariants" falsevalue="" label="Don't include loci found to be non-variant after the subsetting procedure" help="-env,--excludeNonVariants" /> - + <param name="select_type_to_include" type="select" label="Select only a certain type of variants from the input file" multiple="True" display="checkboxes" help="-selectType,--selectTypeToInclude &lt;selectTypeToInclude&gt;"> <option value="INDEL">INDEL</option> <option value="SNP">SNP</option> @@ -209,7 +206,7 @@ <option value="NO_VARIATION">NO_VARIATION</option> </param> </expand> - + </inputs> <outputs> <data format="vcf" name="output_vcf" label="${tool.name} on ${on_string} (Variant File)" /> @@ -228,14 +225,14 @@ <param name="sample_name_repeat" value="0" /> <param name="gatk_param_type_selector" value="basic" /> <param name="analysis_param_type_selector" value="basic" /> - <output name="output_vcf" file="gatk/gatk_variant_select/gatk_variant_select_out_1.vcf" lines_diff="4" /> + <output name="output_vcf" file="gatk/gatk_variant_select/gatk_variant_select_out_1.vcf" lines_diff="4" /> <output name="output_log" file="gatk/gatk_variant_select/gatk_variant_select_out_1.log.contains" compare="contains" /> </test> </tests> <help> **What it does** -Often, a VCF containing many samples and/or variants will need to be subset in order to facilitate certain analyses (e.g. comparing and contrasting cases vs. controls; extracting variant or non-variant loci that meet certain requirements, displaying just a few samples in a browser like IGV, etc.). SelectVariants can be used for this purpose. Given a single VCF file, one or more samples can be extracted from the file (based on a complete sample name or a pattern match). Variants can be further selected by specifying criteria for inclusion, i.e. "DP > 1000" (depth of coverage greater than 1000x), "AF < 0.25" (sites with allele frequency less than 0.25). These JEXL expressions are documented in the `Using JEXL expressions section <http://gatkforums.broadinstitute.org/discussion/1255/what-are-jexl-expressions-and-how-can-i-use-them-with-the-gatk>`_. One can optionally include concordance or discordance tracks for use in selecting overlapping variants. +Often, a VCF containing many samples and/or variants will need to be subset in order to facilitate certain analyses (e.g. comparing and contrasting cases vs. controls; extracting variant or non-variant loci that meet certain requirements, displaying just a few samples in a browser like IGV, etc.). SelectVariants can be used for this purpose. Given a single VCF file, one or more samples can be extracted from the file (based on a complete sample name or a pattern match). Variants can be further selected by specifying criteria for inclusion, i.e. "DP > 1000" (depth of coverage greater than 1000x), "AF < 0.25" (sites with allele frequency less than 0.25). These JEXL expressions are documented in the `Using JEXL expressions section <http://gatkforums.broadinstitute.org/discussion/1255/what-are-jexl-expressions-and-how-can-i-use-them-with-the-gatk>`_. One can optionally include concordance or discordance tracks for use in selecting overlapping variants. For more information on using the SelectVariants module, see this `tool specific page <http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_variantutils_SelectVariants.html>`_.
--- a/variant_validate.xml Tue Apr 28 22:56:25 2015 -0400 +++ b/variant_validate.xml Mon May 04 22:47:06 2015 -0400 @@ -1,10 +1,10 @@ <tool id="gatk2_variant_validate" name="Validate Variants" version="@VERSION@.0"> <description></description> - <expand macro="requirements" /> - <expand macro="version_command" /> <macros> <import>gatk2_macros.xml</import> </macros> + <expand macro="requirements" /> + <expand macro="version_command" /> <command interpreter="python"> gatk2_wrapper.py --stdout "${output_log}" @@ -12,7 +12,7 @@ -p ' @JAR_PATH@ -T "ValidateVariants" - + \$GATK2_SITE_OPTIONS #if $reference_source.reference_source_selector != "history": @@ -22,11 +22,11 @@ ${do_not_validate_filtered_records} ' @DBSNP_OPTIONS@ - + #include source=$standard_gatk_options# </command> <inputs> - + <conditional name="reference_source"> <expand macro="reference_source_selector_param" /> <when value="cached"> @@ -44,12 +44,12 @@ </when> </conditional> <expand macro="dbsnp_param" /> - + <param name="warn_on_errors" type="boolean" checked="False" truevalue="-warnOnErrors" falsevalue="" label="instead of terminating the run at the first error, print warning messages for each error seen." help="-warnOnErrors,--warnOnErrors"/> <param name="do_not_validate_filtered_records" type="boolean" checked="False" truevalue="-doNotValidateFilteredRecords" falsevalue="" label="do not try to validate records that are FILTERed." help="-doNotValidateFilteredRecords,--doNotValidateFilteredRecords"/> - + <expand macro="gatk_param_type_conditional" /> - + </inputs> <outputs> <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />