|
0
|
1 <tool id="gatk_variant_recalibrator" name="Variant Recalibrator" 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_${var_count}"
|
|
|
14 #end for
|
|
|
15 -p 'java
|
|
|
16 -jar "\$JAVA_JAR_PATH/GenomeAnalysisTK.jar"
|
|
|
17 -T "VariantRecalibrator"
|
|
|
18 --num_threads \${GALAXY_SLOTS:-4}
|
|
|
19 -et "NO_ET" ##ET no phone home
|
|
|
20 ##-log "${output_log}" ##don't use this to log to file, instead directly capture stdout
|
|
|
21 #if $reference_source.reference_source_selector != "history":
|
|
|
22 -R "${reference_source.ref_file.fields.path}"
|
|
|
23 #end if
|
|
|
24 --recal_file "${output_recal}"
|
|
|
25 --tranches_file "${output_tranches}"
|
|
|
26 --rscript_file "${output_rscript}"
|
|
|
27 '
|
|
|
28
|
|
|
29 #set $rod_binding_names = dict()
|
|
|
30 #for $rod_binding in $rod_bind:
|
|
|
31 #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom':
|
|
|
32 #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name
|
|
|
33 #elif str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'comp':
|
|
|
34 #set $rod_bind_name = "comp" + $rod_binding.rod_bind_type.custom_rod_name
|
|
|
35 #else
|
|
|
36 #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector
|
|
|
37 #end if
|
|
|
38 #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1
|
|
|
39 #if $rod_binding.rod_bind_type.rod_training_type.rod_training_type_selector == "not_training_truth_known":
|
|
|
40 -d "--resource:${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[$rod_bind_name]}"
|
|
|
41 #else:
|
|
|
42 -d "--resource:${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]}"
|
|
|
43 #end if
|
|
|
44 #end for
|
|
|
45
|
|
|
46 #include source=$standard_gatk_options#
|
|
|
47
|
|
|
48 ##start analysis specific options
|
|
|
49 -p '
|
|
|
50 #if str( $annotations ) != "None":
|
|
|
51 #for $annotation in str( $annotations.fields.gatk_value ).split( ',' ):
|
|
|
52 --use_annotation "${annotation}"
|
|
|
53 #end for
|
|
|
54 #end if
|
|
|
55 #for $additional_annotation in $additional_annotations:
|
|
|
56 --use_annotation "${additional_annotation.additional_annotation_name}"
|
|
|
57 #end for
|
|
|
58 --mode "${mode}"
|
|
|
59 '
|
|
|
60
|
|
|
61 #if $analysis_param_type.analysis_param_type_selector == "advanced":
|
|
|
62 -p '
|
|
|
63 --maxGaussians "${analysis_param_type.max_gaussians}"
|
|
|
64 --maxIterations "${analysis_param_type.max_iterations}"
|
|
|
65 --numKMeans "${analysis_param_type.num_k_means}"
|
|
|
66 --stdThreshold "${analysis_param_type.std_threshold}"
|
|
|
67 --qualThreshold "${analysis_param_type.qual_threshold}"
|
|
|
68 --shrinkage "${analysis_param_type.shrinkage}"
|
|
|
69 --dirichlet "${analysis_param_type.dirichlet}"
|
|
|
70 --priorCounts "${analysis_param_type.prior_counts}"
|
|
|
71 #if str( $analysis_param_type.bad_variant_selector.bad_variant_selector_type ) == 'percent':
|
|
|
72 --percentBadVariants "${analysis_param_type.bad_variant_selector.percent_bad_variants}"
|
|
|
73 #else:
|
|
|
74 --minNumBadVariants "${analysis_param_type.bad_variant_selector.min_num_bad_variants}"
|
|
|
75 #end if
|
|
|
76 --target_titv "${analysis_param_type.target_titv}"
|
|
|
77 #for $tranche in [ $tranche.strip() for $tranche in str( $analysis_param_type.ts_tranche ).split( ',' ) if $tranche.strip() ]
|
|
|
78 --TStranche "${tranche}"
|
|
|
79 #end for
|
|
|
80 #for $ignore_filter in $analysis_param_type.ignore_filters:
|
|
|
81 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.ignore_filter_type_selector )
|
|
|
82 #if $ignore_filter_name == "custom":
|
|
|
83 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.filter_name )
|
|
|
84 #end if
|
|
|
85 --ignore_filter "${ignore_filter_name}"
|
|
|
86 #end for
|
|
|
87 --ts_filter_level "${analysis_param_type.ts_filter_level}"
|
|
|
88 '
|
|
|
89 #end if
|
|
|
90
|
|
|
91
|
|
|
92 &&
|
|
|
93 mv "${output_rscript}.pdf" "${output_tranches_pdf}"
|
|
|
94
|
|
|
95 </command>
|
|
|
96 <inputs>
|
|
|
97 <conditional name="reference_source">
|
|
|
98 <expand macro="reference_source_selector_param" />
|
|
|
99 <when value="cached">
|
|
|
100 <repeat name="variants" title="Variant" min="1" help="-input,--input &lt;input&gt;">
|
|
|
101 <param name="input_variants" type="data" format="vcf" label="Variant file to recalibrate" />
|
|
|
102 </repeat>
|
|
|
103 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;">
|
|
|
104 <options from_data_table="gatk_picard_indexes">
|
|
|
105 <!-- <filter type="data_meta" key="dbkey" ref="variants[0].input_variants" column="dbkey"/> -->
|
|
|
106 </options>
|
|
|
107 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
|
|
|
108 </param>
|
|
|
109 </when>
|
|
|
110 <when value="history"> <!-- FIX ME!!!! -->
|
|
|
111 <repeat name="variants" title="Variant" min="1" help="-input,--input &lt;input&gt;">
|
|
|
112 <param name="input_variants" type="data" format="vcf" label="Variant file to recalibrate" />
|
|
|
113 </repeat>
|
|
|
114 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" />
|
|
|
115 </when>
|
|
|
116 </conditional>
|
|
|
117
|
|
|
118 <repeat name="rod_bind" title="Binding for reference-ordered data" help="-resource,--resource &lt;resource&gt;">
|
|
|
119 <conditional name="rod_bind_type">
|
|
|
120 <param name="rod_bind_type_selector" type="select" label="Binding Type">
|
|
|
121 <option value="dbsnp" selected="True">dbSNP</option>
|
|
|
122 <option value="variant">Variants</option>
|
|
|
123 <option value="snps">SNPs</option>
|
|
|
124 <option value="indels">INDELs</option>
|
|
|
125 <option value="hapmap">HapMap</option>
|
|
|
126 <option value="omni">OMNI</option>
|
|
|
127 <option value="mask">Mask</option>
|
|
|
128 <option value="custom">Custom</option>
|
|
|
129 <option value="comp">Comp</option>
|
|
|
130 </param>
|
|
|
131 <when value="variant">
|
|
|
132 <param name="input_rod" type="data" format="vcf" label="Variant ROD file" />
|
|
|
133 <conditional name="rod_training_type">
|
|
|
134 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
135 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
136 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
137 </param>
|
|
|
138 <when value="not_training_truth_known">
|
|
|
139 <!-- do nothing here -->
|
|
|
140 </when>
|
|
|
141 <when value="is_training_truth_known">
|
|
|
142 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
143 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
144 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
145 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
146 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
147 </when>
|
|
|
148 </conditional>
|
|
|
149 </when>
|
|
|
150 <when value="comp">
|
|
|
151 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
152 <param name="custom_rod_name" type="text" value="Unnamed" label="ROD Name"/>
|
|
|
153 <conditional name="rod_training_type">
|
|
|
154 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
155 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
156 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
157 </param>
|
|
|
158 <when value="not_training_truth_known">
|
|
|
159 <!-- do nothing here -->
|
|
|
160 </when>
|
|
|
161 <when value="is_training_truth_known">
|
|
|
162 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
163 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
164 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
165 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
166 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
167 </when>
|
|
|
168 </conditional>
|
|
|
169 </when>
|
|
|
170 <when value="mask">
|
|
|
171 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
172 <conditional name="rod_training_type">
|
|
|
173 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
174 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
175 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
176 </param>
|
|
|
177 <when value="not_training_truth_known">
|
|
|
178 <!-- do nothing here -->
|
|
|
179 </when>
|
|
|
180 <when value="is_training_truth_known">
|
|
|
181 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
182 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
183 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
184 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
185 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
186 </when>
|
|
|
187 </conditional>
|
|
|
188 </when>
|
|
|
189 <when value="dbsnp">
|
|
|
190 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
191 <conditional name="rod_training_type">
|
|
|
192 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
193 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
194 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
195 </param>
|
|
|
196 <when value="not_training_truth_known">
|
|
|
197 <!-- do nothing here -->
|
|
|
198 </when>
|
|
|
199 <when value="is_training_truth_known">
|
|
|
200 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
201 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
202 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
203 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
204 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
205 </when>
|
|
|
206 </conditional>
|
|
|
207 </when>
|
|
|
208 <when value="snps">
|
|
|
209 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
210 <conditional name="rod_training_type">
|
|
|
211 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
212 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
213 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
214 </param>
|
|
|
215 <when value="not_training_truth_known">
|
|
|
216 <!-- do nothing here -->
|
|
|
217 </when>
|
|
|
218 <when value="is_training_truth_known">
|
|
|
219 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
220 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
221 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
222 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
223 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
224 </when>
|
|
|
225 </conditional>
|
|
|
226 </when>
|
|
|
227 <when value="hapmap">
|
|
|
228 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
229 <conditional name="rod_training_type">
|
|
|
230 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
231 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
232 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
233 </param>
|
|
|
234 <when value="not_training_truth_known">
|
|
|
235 <!-- do nothing here -->
|
|
|
236 </when>
|
|
|
237 <when value="is_training_truth_known">
|
|
|
238 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
239 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
240 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
241 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
242 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
243 </when>
|
|
|
244 </conditional>
|
|
|
245 </when>
|
|
|
246 <when value="omni">
|
|
|
247 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
248 <conditional name="rod_training_type">
|
|
|
249 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
250 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
251 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
252 </param>
|
|
|
253 <when value="not_training_truth_known">
|
|
|
254 <!-- do nothing here -->
|
|
|
255 </when>
|
|
|
256 <when value="is_training_truth_known">
|
|
|
257 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
258 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
259 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
260 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
261 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
262 </when>
|
|
|
263 </conditional>
|
|
|
264 </when>
|
|
|
265 <when value="indels">
|
|
|
266 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
267 <conditional name="rod_training_type">
|
|
|
268 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
269 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
270 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
271 </param>
|
|
|
272 <when value="not_training_truth_known">
|
|
|
273 <!-- do nothing here -->
|
|
|
274 </when>
|
|
|
275 <when value="is_training_truth_known">
|
|
|
276 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
277 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
278 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
279 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
280 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
281 </when>
|
|
|
282 </conditional>
|
|
|
283 </when>
|
|
|
284 <when value="custom">
|
|
|
285 <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
|
|
|
286 <param name="input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
287 <conditional name="rod_training_type">
|
|
|
288 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
|
|
|
289 <option value="is_training_truth_known">Set training/truth/known sites</option>
|
|
|
290 <option value="not_training_truth_known" selected="True">Don't Set options</option>
|
|
|
291 </param>
|
|
|
292 <when value="not_training_truth_known">
|
|
|
293 <!-- do nothing here -->
|
|
|
294 </when>
|
|
|
295 <when value="is_training_truth_known">
|
|
|
296 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
|
|
|
297 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
|
|
|
298 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
|
|
|
299 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
|
|
|
300 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
|
|
|
301 </when>
|
|
|
302 </conditional>
|
|
|
303 </when>
|
|
|
304 </conditional>
|
|
|
305 </repeat>
|
|
|
306
|
|
|
307 <param name="annotations" type="select" multiple="True" display="checkboxes" label="annotations which should used for calculations" help="-an,--use_annotation &lt;use_annotation&gt;">
|
|
|
308 <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs -->
|
|
|
309 <options from_data_table="gatk_annotations">
|
|
|
310 <filter type="multiple_splitter" column="tools_valid_for" separator=","/>
|
|
|
311 <filter type="static_value" value="VariantRecalibrator" column="tools_valid_for"/>
|
|
|
312 </options>
|
|
|
313 </param>
|
|
|
314
|
|
|
315 <repeat name="additional_annotations" title="Additional annotation" help="-an,--use_annotation &lt;use_annotation&gt;">
|
|
|
316 <param name="additional_annotation_name" type="text" value="" label="Annotation name" />
|
|
|
317 </repeat>
|
|
|
318
|
|
|
319 <param name="mode" type="select" label="Recalibration mode" help="-mode,--mode &lt;mode&gt;">
|
|
|
320 <option value="SNP" selected="True">SNP</option>
|
|
|
321 <option value="INDEL">INDEL</option>
|
|
|
322 <option value="BOTH">BOTH</option>
|
|
|
323 </param>
|
|
|
324
|
|
|
325 <expand macro="gatk_param_type_conditional" />
|
|
|
326
|
|
|
327 <expand macro="analysis_type_conditional">
|
|
|
328 <param name="max_gaussians" type="integer" label="maximum number of Gaussians to try during variational Bayes Algorithm" value="10" help="-mG,--maxGaussians &lt;maxGaussians&gt;"/>
|
|
|
329 <param name="max_iterations" type="integer" label="maximum number of maximum number of VBEM iterations to be performed in variational Bayes Algorithm" value="100" help="-mI,--maxIterations &lt;maxIterations&gt;"/>
|
|
|
330 <param name="num_k_means" type="integer" label="number of k-means iterations to perform in order to initialize the means of the Gaussians in the Gaussian mixture model" value="30" help="-nKM,--numKMeans &lt;numKMeans&gt;"/>
|
|
|
331 <param name="std_threshold" type="float" label="If a variant has annotations more than -std standard deviations away from mean then don't use it for building the Gaussian mixture model." value="8.0" help="-std,--stdThreshold &lt;stdThreshold&gt;"/>
|
|
|
332 <param name="qual_threshold" type="float" label="If a known variant has raw QUAL value less than -qual then don't use it for building the Gaussian mixture model." value="80.0" help="-qual,--qualThreshold &lt;qualThreshold&gt;"/>
|
|
|
333 <param name="shrinkage" type="float" label="shrinkage parameter in variational Bayes algorithm" value="1.0" help="-shrinkage,--shrinkage &lt;shrinkage&gt;"/>
|
|
|
334 <param name="dirichlet" type="float" label="dirichlet parameter in variational Bayes algorithm" value="0.001" help="-dirichlet,--dirichlet &lt;dirichlet&gt;"/>
|
|
|
335 <param name="prior_counts" type="float" label="number of prior counts to use in variational Bayes algorithm" value="20.0" help="-priorCounts,--priorCounts &lt;priorCounts&gt;"/>
|
|
|
336 <conditional name="bad_variant_selector">
|
|
|
337 <param name="bad_variant_selector_type" type="select" label="How to specify bad variants">
|
|
|
338 <option value="percent" selected="True">Percent</option>
|
|
|
339 <option value="min_num">Number</option>
|
|
|
340 </param>
|
|
|
341 <when value="percent">
|
|
|
342 <param name="percent_bad_variants" type="float" label="percentage of the worst scoring variants to use when building the Gaussian mixture model of bad variants. 0.07 means bottom 7 percent." value="0.03" help="-percentBad,--percentBadVariants &lt;percentBadVariants&gt;"/>
|
|
|
343 </when>
|
|
|
344 <when value="min_num">
|
|
|
345 <param name="min_num_bad_variants" type="integer" label="minimum amount of worst scoring variants to use when building the Gaussian mixture model of bad variants. Will override -percentBad arugment if necessary" value="2000" help="-minNumBad,--minNumBadVariants &lt;minNumBadVariants&gt;"/>
|
|
|
346 </when>
|
|
|
347 </conditional>
|
|
|
348 <param name="target_titv" type="float" label="expected novel Ti/Tv ratio to use when calculating FDR tranches and for display on optimization curve output figures. (approx 2.15 for whole genome experiments). ONLY USED FOR PLOTTING PURPOSES!" value="2.15" help="-titv,--target_titv &lt;target_titv&gt;"/>
|
|
|
349 <param name="ts_tranche" type="text" label="levels of novel false discovery rate (FDR, implied by ti/tv) at which to slice the data. (in percent, that is 1.0 for 1 percent)" value="100.0, 99.9, 99.0, 90.0" help="-tranche,--TStranche &lt;TStranche&gt;"/>
|
|
|
350 <repeat name="ignore_filters" title="Ignore Filter" help="-ignoreFilter,--ignore_filter &lt;ignore_filter&gt;">
|
|
|
351 <conditional name="ignore_filter_type">
|
|
|
352 <param name="ignore_filter_type_selector" type="select" label="Filter Type">
|
|
|
353 <option value="HARD_TO_VALIDATE">HARD_TO_VALIDATE</option>
|
|
|
354 <option value="LowQual" >LowQual</option>
|
|
|
355 <option value="custom" selected="True">Other</option>
|
|
|
356 </param>
|
|
|
357 <when value="custom">
|
|
|
358 <param name="filter_name" type="text" value="" label="Filter name"/>
|
|
|
359 </when>
|
|
|
360 <when value="HARD_TO_VALIDATE" />
|
|
|
361 <when value="LowQual" />
|
|
|
362 </conditional>
|
|
|
363 </repeat>
|
|
|
364 <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;"/>
|
|
|
365 </expand>
|
|
|
366 </inputs>
|
|
|
367 <outputs>
|
|
|
368 <data format="gatk_recal" name="output_recal" label="${tool.name} on ${on_string} (Recalibration File)" />
|
|
|
369 <data format="gatk_tranche" name="output_tranches" label="${tool.name} on ${on_string} (Tranches File)" />
|
|
|
370 <data format="txt" name="output_rscript" label="${tool.name} on ${on_string} (RScript File)" />
|
|
|
371 <data format="pdf" name="output_tranches_pdf" label="${tool.name} on ${on_string} (PDF File)" />
|
|
|
372 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
|
|
|
373 </outputs>
|
|
|
374 <tests>
|
|
|
375 <!-- ADD TESTS -->
|
|
|
376 </tests>
|
|
|
377 <help>
|
|
|
378 **What it does**
|
|
|
379
|
|
|
380 Takes variant calls as .vcf files, learns a Gaussian mixture model over the variant annotations and evaluates the variant -- assigning an informative lod score
|
|
|
381
|
|
|
382 For more information on using the VariantRecalibrator module, see this `tool specific page <http://www.broadinstitute.org/gsa/wiki/index.php/Variant_quality_score_recalibration>`_.
|
|
|
383
|
|
|
384 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>`_.
|
|
|
385
|
|
|
386 If you encounter errors, please view the `GATK FAQ <http://www.broadinstitute.org/gsa/wiki/index.php/Frequently_Asked_Questions>`_.
|
|
|
387
|
|
|
388 ------
|
|
|
389
|
|
|
390 **Inputs**
|
|
|
391
|
|
|
392 GenomeAnalysisTK: VariantRecalibrator accepts a variant input file.
|
|
|
393
|
|
|
394
|
|
|
395 **Outputs**
|
|
|
396
|
|
|
397 The output is in VCF format.
|
|
|
398
|
|
|
399
|
|
|
400 Go `here <http://www.broadinstitute.org/gsa/wiki/index.php/Input_files_for_the_GATK>`_ for details on GATK file formats.
|
|
|
401
|
|
|
402 -------
|
|
|
403
|
|
|
404 **Settings**::
|
|
|
405
|
|
|
406
|
|
|
407 tranches_file The output tranches file used by ApplyRecalibration
|
|
|
408 use_annotation The names of the annotations which should used for calculations
|
|
|
409 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)
|
|
|
410 maxGaussians The maximum number of Gaussians to try during variational Bayes algorithm
|
|
|
411 maxIterations The maximum number of VBEM iterations to be performed in variational Bayes algorithm. Procedure will normally end when convergence is detected.
|
|
|
412 numKMeans The number of k-means iterations to perform in order to initialize the means of the Gaussians in the Gaussian mixture model.
|
|
|
413 stdThreshold If a variant has annotations more than -std standard deviations away from mean then don't use it for building the Gaussian mixture model.
|
|
|
414 qualThreshold If a known variant has raw QUAL value less than -qual then don't use it for building the Gaussian mixture model.
|
|
|
415 shrinkage The shrinkage parameter in variational Bayes algorithm.
|
|
|
416 dirichlet The dirichlet parameter in variational Bayes algorithm.
|
|
|
417 priorCounts The number of prior counts to use in variational Bayes algorithm.
|
|
|
418 percentBadVariants What percentage of the worst scoring variants to use when building the Gaussian mixture model of bad variants. 0.07 means bottom 7 percent.
|
|
|
419 minNumBadVariants The minimum amount of worst scoring variants to use when building the Gaussian mixture model of bad variants. Will override -percentBad arugment if necessary.
|
|
|
420 recal_file The output recal file used by ApplyRecalibration
|
|
|
421 target_titv The expected novel Ti/Tv ratio to use when calculating FDR tranches and for display on optimization curve output figures. (approx 2.15 for whole genome experiments). ONLY USED FOR PLOTTING PURPOSES!
|
|
|
422 TStranche The levels of novel false discovery rate (FDR, implied by ti/tv) at which to slice the data. (in percent, that is 1.0 for 1 percent)
|
|
|
423 ignore_filter If specified the optimizer will use variants even if the specified filter name is marked in the input VCF file
|
|
|
424 path_to_Rscript The path to your implementation of Rscript. For Broad users this is maybe /broad/tools/apps/R-2.6.0/bin/Rscript
|
|
|
425 rscript_file The output rscript file generated by the VQSR to aid in visualization of the input data and learned model
|
|
|
426 path_to_resources Path to resources folder holding the Sting R scripts.
|
|
|
427 ts_filter_level The truth sensitivity level at which to start filtering, used here to indicate filtered variants in plots
|
|
|
428
|
|
|
429 @CITATION_SECTION@
|
|
|
430 </help>
|
|
|
431 </tool>
|
