Mercurial > repos > lz_hust > gatktools

comparison variant_recalibrator.xml @ 15:01ff8dd37d4d draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author lz_hust
date Sat, 01 Jun 2019 07:20:41 -0400
parents
children
comparison
equal deleted inserted replaced
14:68426930d59c 15:01ff8dd37d4d
1 <tool id="gatk2_variant_recalibrator" name="Variant Recalibrator" version="@VERSION@.1">
2 <description></description>
3 <macros>
4 <import>gatk2_macros.xml</import>
5 </macros>
6 <expand macro="requirements">
7 <requirement type="package" version="0.9.3">ggplot2</requirement>
8 </expand>
9 <expand macro="version_command" />
10 <command interpreter="python">
11 gatk2_wrapper.py
12 --stdout "${output_log}"
13 #for $var_count, $variant in enumerate( $reference_source.input_variants ):
14 -d "--input:input_${var_count},%(file_type)s" "${variant}" "${variant.ext}" "input_variants_${var_count}"
15 #end for
16 -p '
17 @JAR_PATH@
18 -T "VariantRecalibrator"
19 \$GATK2_SITE_OPTIONS
20
21 @THREADS@
22
23 #if $reference_source.reference_source_selector != "history":
24 -R "${reference_source.ref_file.fields.path}"
25 #end if
26 --recal_file "${output_recal}"
27 --tranches_file "${output_tranches}"
28 --rscript_file "${output_rscript}"
29 '
30
31 #set $rod_binding_names = dict()
32 #for $rod_binding in $rod_bind:
33 #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom':
34 #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name
35 #elif str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'comp':
36 #set $rod_bind_name = "comp" + $rod_binding.rod_bind_type.custom_rod_name
37 #else
38 #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector
39 #end if
40 #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1
41 #if $rod_binding.rod_bind_type.rod_training_type.rod_training_type_selector == "not_training_truth_known":
42 -d "--resource:${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[$rod_bind_name]}"
43 #else:
44 -d "--resource:${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]}"
45 #end if
46 #end for
47
48 #include source=$standard_gatk_options#
49
50 ##start analysis specific options
51 -p '
52 #if str( $annotations ) != "None":
53 #for $annotation in str( $annotations.fields.gatk_value ).split( ',' ):
54 --use_annotation "${annotation}"
55 #end for
56 #end if
57 #for $additional_annotation in $additional_annotations:
58 --use_annotation "${additional_annotation.additional_annotation_name}"
59 #end for
60 --mode "${mode}"
61 '
62
63 #if $analysis_param_type.analysis_param_type_selector == "advanced":
64 -p '
65 --maxGaussians "${analysis_param_type.max_gaussians}"
66 --maxIterations "${analysis_param_type.max_iterations}"
67 --numKMeans "${analysis_param_type.num_k_means}"
68 --stdThreshold "${analysis_param_type.std_threshold}"
69 --shrinkage "${analysis_param_type.shrinkage}"
70 --dirichlet "${analysis_param_type.dirichlet}"
71 --priorCounts "${analysis_param_type.prior_counts}"
72
73 --minNumBadVariants "${analysis_param_type.min_num_bad_variants}"
74
75 --target_titv "${analysis_param_type.target_titv}"
76 #for $tranche in [ $tranche.strip() for $tranche in str( $analysis_param_type.ts_tranche ).split( ',' ) if $tranche.strip() ]
77 --TStranche "${tranche}"
78 #end for
79 #for $ignore_filter in $analysis_param_type.ignore_filters:
80 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.ignore_filter_type_selector )
81 #if $ignore_filter_name == "custom":
82 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.filter_name )
83 #end if
84 --ignore_filter "${ignore_filter_name}"
85 #end for
86 '
87 #end if
88
89 &amp;&amp;
90 mv "${output_rscript}.pdf" "${output_tranches_pdf}"
91
92 </command>
93 <inputs>
94 <conditional name="reference_source">
95 <expand macro="reference_source_selector_param" />
96 <when value="cached">
97 <expand macro="input_variants" />
98 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;">
99 <options from_data_table="gatk2_picard_indexes">
100 <!-- <filter type="data_meta" key="dbkey" ref="variants[0].input_variants" column="dbkey"/> -->
101 </options>
102 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
103 </param>
104 </when>
105 <when value="history"> <!-- FIX ME!!!! -->
106 <expand macro="input_variants" />
107 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;" />
108 </when>
109 </conditional>
110
111 <repeat name="rod_bind" title="Binding for reference-ordered data" help="-resource,--resource &amp;lt;resource&amp;gt;" min="2">
112 <conditional name="rod_bind_type">
113 <param name="rod_bind_type_selector" type="select" label="Binding Type">
114 <option value="dbsnp" selected="True">dbSNP</option>
115 <option value="variant">Variants</option>
116 <option value="snps">SNPs</option>
117 <option value="indels">INDELs</option>
118 <option value="hapmap">HapMap</option>
119 <option value="omni">OMNI</option>
120 <option value="mask">Mask</option>
121 <option value="custom">Custom</option>
122 <option value="comp">Comp</option>
123 </param>
124 <when value="variant">
125 <param name="input_rod" type="data" format="vcf" label="Variant ROD file" />
126 <conditional name="rod_training_type">
127 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
128 <option value="is_training_truth_known">Set training/truth/known sites</option>
129 <option value="not_training_truth_known" selected="True">Don't Set options</option>
130 </param>
131 <when value="not_training_truth_known">
132 <!-- do nothing here -->
133 </when>
134 <when value="is_training_truth_known">
135 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
136 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
137 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
138 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
139 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
140 </when>
141 </conditional>
142 </when>
143 <when value="comp">
144 <param name="input_rod" type="data" format="vcf" label="ROD file" />
145 <param name="custom_rod_name" type="text" value="Unnamed" label="ROD Name"/>
146 <conditional name="rod_training_type">
147 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
148 <option value="is_training_truth_known">Set training/truth/known sites</option>
149 <option value="not_training_truth_known" selected="True">Don't Set options</option>
150 </param>
151 <when value="not_training_truth_known">
152 <!-- do nothing here -->
153 </when>
154 <when value="is_training_truth_known">
155 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
156 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
157 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
158 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
159 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
160 </when>
161 </conditional>
162 </when>
163 <when value="mask">
164 <param name="input_rod" type="data" format="vcf" label="ROD file" />
165 <conditional name="rod_training_type">
166 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
167 <option value="is_training_truth_known">Set training/truth/known sites</option>
168 <option value="not_training_truth_known" selected="True">Don't Set options</option>
169 </param>
170 <when value="not_training_truth_known">
171 <!-- do nothing here -->
172 </when>
173 <when value="is_training_truth_known">
174 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
175 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
176 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
177 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
178 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
179 </when>
180 </conditional>
181 </when>
182 <when value="dbsnp">
183 <param name="input_rod" type="data" format="vcf" label="ROD file" />
184 <conditional name="rod_training_type">
185 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
186 <option value="is_training_truth_known">Set training/truth/known sites</option>
187 <option value="not_training_truth_known" selected="True">Don't Set options</option>
188 </param>
189 <when value="not_training_truth_known">
190 <!-- do nothing here -->
191 </when>
192 <when value="is_training_truth_known">
193 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
194 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
195 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
196 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
197 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
198 </when>
199 </conditional>
200 </when>
201 <when value="snps">
202 <param name="input_rod" type="data" format="vcf" label="ROD file" />
203 <conditional name="rod_training_type">
204 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
205 <option value="is_training_truth_known">Set training/truth/known sites</option>
206 <option value="not_training_truth_known" selected="True">Don't Set options</option>
207 </param>
208 <when value="not_training_truth_known">
209 <!-- do nothing here -->
210 </when>
211 <when value="is_training_truth_known">
212 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
213 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
214 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
215 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
216 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
217 </when>
218 </conditional>
219 </when>
220 <when value="hapmap">
221 <param name="input_rod" type="data" format="vcf" label="ROD file" />
222 <conditional name="rod_training_type">
223 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
224 <option value="is_training_truth_known">Set training/truth/known sites</option>
225 <option value="not_training_truth_known" selected="True">Don't Set options</option>
226 </param>
227 <when value="not_training_truth_known">
228 <!-- do nothing here -->
229 </when>
230 <when value="is_training_truth_known">
231 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
232 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
233 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
234 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
235 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
236 </when>
237 </conditional>
238 </when>
239 <when value="omni">
240 <param name="input_rod" type="data" format="vcf" label="ROD file" />
241 <conditional name="rod_training_type">
242 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
243 <option value="is_training_truth_known">Set training/truth/known sites</option>
244 <option value="not_training_truth_known" selected="True">Don't Set options</option>
245 </param>
246 <when value="not_training_truth_known">
247 <!-- do nothing here -->
248 </when>
249 <when value="is_training_truth_known">
250 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
251 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
252 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
253 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
254 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
255 </when>
256 </conditional>
257 </when>
258 <when value="indels">
259 <param name="input_rod" type="data" format="vcf" label="ROD file" />
260 <conditional name="rod_training_type">
261 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
262 <option value="is_training_truth_known">Set training/truth/known sites</option>
263 <option value="not_training_truth_known" selected="True">Don't Set options</option>
264 </param>
265 <when value="not_training_truth_known">
266 <!-- do nothing here -->
267 </when>
268 <when value="is_training_truth_known">
269 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
270 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
271 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
272 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
273 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
274 </when>
275 </conditional>
276 </when>
277 <when value="custom">
278 <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
279 <param name="input_rod" type="data" format="vcf" label="ROD file" />
280 <conditional name="rod_training_type">
281 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
282 <option value="is_training_truth_known">Set training/truth/known sites</option>
283 <option value="not_training_truth_known" selected="True">Don't Set options</option>
284 </param>
285 <when value="not_training_truth_known">
286 <!-- do nothing here -->
287 </when>
288 <when value="is_training_truth_known">
289 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
290 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
291 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
292 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
293 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
294 </when>
295 </conditional>
296 </when>
297 </conditional>
298 </repeat>
299
300 <param name="annotations" type="select" multiple="True" display="checkboxes" label="annotations which should used for calculations" help="-an,--use_annotation &amp;lt;use_annotation&amp;gt;">
301 <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs -->
302 <options from_data_table="gatk2_annotations">
303 <filter type="multiple_splitter" column="tools_valid_for" separator=","/>
304 <filter type="static_value" value="VariantRecalibrator" column="tools_valid_for"/>
305 </options>
306 </param>
307
308 <repeat name="additional_annotations" title="Additional annotation" help="-an,--use_annotation &amp;lt;use_annotation&amp;gt;">
309 <param name="additional_annotation_name" type="text" value="" label="Annotation name" />
310 </repeat>
311
312 <param name="mode" type="select" label="Recalibration mode" help="-mode,--mode &amp;lt;mode&amp;gt;">
313 <option value="SNP" selected="True">SNP</option>
314 <option value="INDEL">INDEL</option>
315 <option value="BOTH">BOTH</option>
316 </param>
317
318 <expand macro="gatk_param_type_conditional" />
319
320 <expand macro="analysis_type_conditional">
321 <param name="max_gaussians" type="integer" label="maximum number of Gaussians to try during variational Bayes Algorithm" value="8" help="-mG,--maxGaussians &amp;lt;maxGaussians&amp;gt;"/>
322 <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 &amp;lt;maxIterations&amp;gt;"/>
323 <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="100" help="-nKM,--numKMeans &amp;lt;numKMeans&amp;gt;"/>
324 <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="10.0" help="-std,--stdThreshold &amp;lt;stdThreshold&amp;gt;"/>
325 <param name="shrinkage" type="float" label="shrinkage parameter in variational Bayes algorithm" value="1.0" help="-shrinkage,--shrinkage &amp;lt;shrinkage&amp;gt;"/>
326 <param name="dirichlet" type="float" label="dirichlet parameter in variational Bayes algorithm" value="0.001" help="-dirichlet,--dirichlet &amp;lt;dirichlet&amp;gt;"/>
327 <param name="prior_counts" type="float" label="number of prior counts to use in variational Bayes algorithm" value="20.0" help="-priorCounts,--priorCounts &amp;lt;priorCounts&amp;gt;"/>
328 <!--<param name="trustAllPolymorphic" type="boolean" label="trustAllPolymorphic" truevalue="-/-trustAllPolymorphic=true" falsevalue="-/-trustAllPolymorphic=false"
329 help="Trust that all the input training sets' unfiltered records contain only polymorphic sites to drastically speed up the computation. -trustAllPolymorphic" />-->
330 <param name="min_num_bad_variants" type="integer" label="Minimum number of worst scoring variants to use when building the Gaussian mixture model of bad variants" value="1000" help="--minNumBadVariants &amp;lt;minNumBadVariants&amp;gt;"/>
331 <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 &amp;lt;target_titv&amp;gt;"/>
332 <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 &amp;lt;TStranche&amp;gt;"/>
333 <repeat name="ignore_filters" title="Ignore Filter" help="-ignoreFilter,--ignore_filter &amp;lt;ignore_filter&amp;gt;">
334 <conditional name="ignore_filter_type">
335 <param name="ignore_filter_type_selector" type="select" label="Filter Type">
336 <option value="HARD_TO_VALIDATE">HARD_TO_VALIDATE</option>
337 <option value="LowQual" >LowQual</option>
338 <option value="custom" selected="True">Other</option>
339 </param>
340 <when value="custom">
341 <param name="filter_name" type="text" value="" label="Filter name"/>
342 </when>
343 <when value="HARD_TO_VALIDATE" />
344 <when value="LowQual" />
345 </conditional>
346 </repeat>
347 </expand>
348 </inputs>
349 <outputs>
350 <data format="gatk_recal" name="output_recal" label="${tool.name} on ${on_string} (Recalibration File)" />
351 <data format="gatk_tranche" name="output_tranches" label="${tool.name} on ${on_string} (Tranches File)" />
352 <data format="txt" name="output_rscript" label="${tool.name} on ${on_string} (RScript File)" />
353 <data format="pdf" name="output_tranches_pdf" label="${tool.name} on ${on_string} (PDF File)" />
354 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
355 </outputs>
356 <tests>
357 <!-- ADD TESTS -->
358 </tests>
359 <help>
360 **What it does**
361
362 Takes variant calls as .vcf files, learns a Gaussian mixture model over the variant annotations and evaluates the variant -- assigning an informative lod score
363
364 For more information on using the VariantRecalibrator module, see this `tool specific page &lt;http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_variantrecalibration_VariantRecalibrator.html&gt;`_.
365
366 To learn about best practices for variant detection using GATK, see this `overview &lt;http://www.broadinstitute.org/gatk/guide/topic?name=best-practices&gt;`_.
367
368 If you encounter errors, please view the `GATK FAQ &lt;http://www.broadinstitute.org/gatk/guide/topic?name=faqs&gt;`_.
369
370 ------
371
372 **Inputs**
373
374 GenomeAnalysisTK: VariantRecalibrator accepts a variant input file.
375
376
377 **Outputs**
378
379 The output is in VCF format.
380
381
382 Go `here &lt;http://www.broadinstitute.org/gatk/guide/topic?name=intro&gt;`_ for details on GATK file formats.
383
384 -------
385
386 **Settings**::
387
388
389 tranches_file The output tranches file used by ApplyRecalibration
390 use_annotation The names of the annotations which should used for calculations
391 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)
392 maxGaussians The maximum number of Gaussians to try during variational Bayes algorithm
393 maxIterations The maximum number of VBEM iterations to be performed in variational Bayes algorithm. Procedure will normally end when convergence is detected.
394 numKMeans The number of k-means iterations to perform in order to initialize the means of the Gaussians in the Gaussian mixture model.
395 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.
396 shrinkage The shrinkage parameter in variational Bayes algorithm.
397 dirichlet The dirichlet parameter in variational Bayes algorithm.
398 priorCounts The number of prior counts to use in variational Bayes algorithm.
399 minNumBadVariants The minimum amount of worst scoring variants to use when building the Gaussian mixture model of bad variants.
400 recal_file The output recal file used by ApplyRecalibration
401 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!
402 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)
403 ignore_filter If specified the optimizer will use variants even if the specified filter name is marked in the input VCF file
404 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
405 rscript_file The output rscript file generated by the VQSR to aid in visualization of the input data and learned model
406 path_to_resources Path to resources folder holding the Sting R scripts.
407
408 @CITATION_SECTION@
409 </help>
410 <expand macro="citations" />
411 </tool>