|
0
|
1 <tool id="gatk_variant_apply_recalibration" name="Apply Variant Recalibration" version="0.0.4">
|
|
|
2 <description></description>
|
|
|
3 <requirements>
|
|
|
4 <requirement type="package" version="1.4">gatk</requirement>
|
|
|
5 </requirements>
|
|
|
6 <macros>
|
|
|
7 <import>gatk_macros.xml</import>
|
|
|
8 </macros>
|
|
|
9 <command interpreter="python">gatk_wrapper.py
|
|
|
10 --max_jvm_heap_fraction "1"
|
|
|
11 --stdout "${output_log}"
|
|
|
12 #for $var_count, $variant in enumerate( $reference_source.variants ):
|
|
|
13 -d "--input:input_${var_count},%(file_type)s" "${variant.input_variants}" "${variant.input_variants.ext}" "input_variants_${var_count}"
|
|
|
14 #end for
|
|
|
15 -p 'java
|
|
|
16 -jar "\$JAVA_JAR_PATH/GenomeAnalysisTK.jar"
|
|
|
17 -T "ApplyRecalibration"
|
|
|
18 ##--num_threads 4 ##hard coded, for now
|
|
|
19 -et "NO_ET" ##ET no phone home
|
|
|
20 #if $reference_source.reference_source_selector != "history":
|
|
|
21 -R "${reference_source.ref_file.fields.path}"
|
|
|
22 #end if
|
|
|
23 --recal_file "${reference_source.input_recal}"
|
|
|
24 --tranches_file "${reference_source.input_tranches}"
|
|
|
25 --out "${output_variants}"
|
|
|
26 '
|
|
|
27
|
|
|
28 #include source=$standard_gatk_options#
|
|
|
29
|
|
|
30 ##start analysis specific options
|
|
|
31 -p '
|
|
|
32 --mode "${mode}"
|
|
|
33
|
|
|
34 #for $ignore_filter in $ignore_filters:
|
|
|
35 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.ignore_filter_type_selector )
|
|
|
36 #if $ignore_filter_name == "custom":
|
|
|
37 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.filter_name )
|
|
|
38 #end if
|
|
|
39 --ignore_filter "${ignore_filter_name}"
|
|
|
40 #end for
|
|
|
41 --ts_filter_level "${ts_filter_level}"
|
|
|
42 '
|
|
|
43 </command>
|
|
|
44 <inputs>
|
|
|
45 <conditional name="reference_source">
|
|
|
46 <expand macro="reference_source_selector_param" />
|
|
|
47 <when value="cached">
|
|
|
48 <repeat name="variants" title="Variant" min="1" help="-input,--input &lt;input&gt;">
|
|
|
49 <param name="input_variants" type="data" format="vcf" label="Variant file to annotate"/>
|
|
|
50 </repeat>
|
|
|
51 <param name="input_recal" type="data" format="gatk_recal" label="Variant Recalibration file" help="-recalFile,--recal_file &lt;recal_file&gt;" />
|
|
|
52 <param name="input_tranches" type="data" format="gatk_tranche" label="Variant Tranches file" help="-tranchesFile,--tranches_file &lt;tranches_file&gt;" />
|
|
|
53 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;">
|
|
|
54 <options from_data_table="gatk_picard_indexes">
|
|
|
55 <!-- <filter type="data_meta" key="dbkey" ref="variants[0].input_variants" column="dbkey"/> -->
|
|
|
56 </options>
|
|
|
57 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
|
|
|
58 </param>
|
|
|
59 </when>
|
|
|
60 <when value="history"> <!-- FIX ME!!!! -->
|
|
|
61 <repeat name="variants" title="Variant" min="1" help="-input,--input &lt;input&gt;">
|
|
|
62 <param name="input_variants" type="data" format="vcf" label="Variant file to annotate" />
|
|
|
63 </repeat>
|
|
|
64 <param name="input_recal" type="data" format="gatk_recal" label="Variant Recalibration file" help="-recalFile,--recal_file &lt;recal_file&gt;" />
|
|
|
65 <param name="input_tranches" type="data" format="gatk_tranche" label="Variant Tranches file" help="-tranchesFile,--tranches_file &lt;tranches_file&gt;" />
|
|
|
66 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" />
|
|
|
67 </when>
|
|
|
68 </conditional>
|
|
|
69
|
|
|
70 <expand macro="gatk_param_type_conditional" />
|
|
|
71
|
|
|
72 <param name="mode" type="select" label="Recalibration mode" help="-mode,--mode &lt;mode&gt;">
|
|
|
73 <option value="SNP" selected="True">SNP</option>
|
|
|
74 <option value="INDEL">INDEL</option>
|
|
|
75 <option value="BOTH">BOTH</option>
|
|
|
76 </param>
|
|
|
77 <repeat name="ignore_filters" title="Ignore Filter" help="-ignoreFilter,--ignore_filter &lt;ignore_filter&gt;">
|
|
|
78 <conditional name="ignore_filter_type">
|
|
|
79 <param name="ignore_filter_type_selector" type="select" label="Filter Type">
|
|
|
80 <option value="HARD_TO_VALIDATE">HARD_TO_VALIDATE</option>
|
|
|
81 <option value="LowQual" >LowQual</option>
|
|
|
82 <option value="custom" selected="True">Other</option>
|
|
|
83 </param>
|
|
|
84 <when value="custom">
|
|
|
85 <param name="filter_name" type="text" value="" label="Filter name"/>
|
|
|
86 </when>
|
|
|
87 <when value="HARD_TO_VALIDATE" />
|
|
|
88 <when value="LowQual" />
|
|
|
89 </conditional>
|
|
|
90 </repeat>
|
|
|
91 <param name="ts_filter_level" type="float" label="truth sensitivity level at which to start filtering, used here to indicate filtered variants in plots" value="99.0" help="-ts_filter_level,--ts_filter_level &lt;ts_filter_level&gt;"/>
|
|
|
92 </inputs>
|
|
|
93 <outputs>
|
|
|
94 <data format="vcf" name="output_variants" label="${tool.name} on ${on_string} (Variants File)" />
|
|
|
95 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
|
|
|
96 </outputs>
|
|
|
97 <tests>
|
|
|
98 <!-- ADD TESTS -->
|
|
|
99 </tests>
|
|
|
100 <help>
|
|
|
101 **What it does**
|
|
|
102
|
|
|
103 Applies cuts to the input vcf file (by adding filter lines) to achieve the desired novel FDR levels which were specified during VariantRecalibration
|
|
|
104
|
|
|
105 For more information on using the ApplyRecalibration module, see this `tool specific page <http://www.broadinstitute.org/gsa/wiki/index.php/Variant_quality_score_recalibration>`_.
|
|
|
106
|
|
|
107 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>`_.
|
|
|
108
|
|
|
109 If you encounter errors, please view the `GATK FAQ <http://www.broadinstitute.org/gsa/wiki/index.php/Frequently_Asked_Questions>`_.
|
|
|
110
|
|
|
111 ------
|
|
|
112
|
|
|
113 **Inputs**
|
|
|
114
|
|
|
115 GenomeAnalysisTK: ApplyRecalibration accepts a variant input file, a recalibration file and a tranches file.
|
|
|
116
|
|
|
117
|
|
|
118 **Outputs**
|
|
|
119
|
|
|
120 The output is in VCF format.
|
|
|
121
|
|
|
122
|
|
|
123 Go `here <http://www.broadinstitute.org/gsa/wiki/index.php/Input_files_for_the_GATK>`_ for details on GATK file formats.
|
|
|
124
|
|
|
125 -------
|
|
|
126
|
|
|
127 **Settings**::
|
|
|
128
|
|
|
129
|
|
|
130 recal_file The output recal file used by ApplyRecalibration
|
|
|
131 tranches_file The input tranches file describing where to cut the data
|
|
|
132 out The output filtered, recalibrated VCF file
|
|
|
133 ts_filter_level The truth sensitivity level at which to start filtering
|
|
|
134 ignore_filter If specified the optimizer will use variants even if the specified filter name is marked in the input VCF file
|
|
|
135 mode Recalibration mode to employ: 1.) SNP for recalibrating only SNPs (emitting indels untouched in the output VCF); 2.) INDEL for indels; and 3.) BOTH for recalibrating both SNPs and indels simultaneously. (SNP|INDEL|BOTH)
|
|
|
136
|
|
|
137 @CITATION_SECTION@
|
|
|
138 </help>
|
|
|
139 </tool>
|
