Mercurial > repos > jjohnson > gatk2

comparison variant_recalibrator.xml @ 0:74c05070a3f8 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author jjohnson
date Thu, 01 Nov 2012 13:53:22 -0400
parents
children 7533db8dfb5b
comparison
equal deleted inserted replaced
-1:000000000000 0:74c05070a3f8
1 <tool id="gatk2_variant_recalibrator" name="Variant Recalibrator" version="0.0.4">
2 <description></description>
3 <requirements>
4 <requirement type="package" version="2.2">gatk</requirement>
5 </requirements>
6 <command interpreter="python">gatk2_wrapper.py
7 --max_jvm_heap_fraction "1"
8 --stdout "${output_log}"
9 #for $var_count, $variant in enumerate( $reference_source.variants ):
10 -d "--input:input_${var_count},%(file_type)s" "${variant.input_variants}" "${variant.input_variants.ext}" "input_variants_${var_count}"
11 #end for
12 -p 'java
13 -jar "\$GATK2_PATH/GenomeAnalysisTK.jar"
14 -T "VariantRecalibrator"
15 \$GATK2_SITE_OPTIONS
16 ##--num_threads 4 ##hard coded, for now
17 ##-et "NO_ET" -K "\$GATK2_BASE/gatk2_key_file"##ET no phone home
18 ##-log "${output_log}" ##don't use this to log to file, instead directly capture stdout
19 #if $reference_source.reference_source_selector != "history":
20 -R "${reference_source.ref_file.fields.path}"
21 #end if
22 --recal_file "${output_recal}"
23 --tranches_file "${output_tranches}"
24 --rscript_file "${output_rscript}"
25 '
26
27 #set $rod_binding_names = dict()
28 #for $rod_binding in $rod_bind:
29 #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom':
30 #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name
31 #elif str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'comp':
32 #set $rod_bind_name = "comp" + $rod_binding.rod_bind_type.custom_rod_name
33 #else
34 #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector
35 #end if
36 #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1
37 #if $rod_binding.rod_bind_type.rod_training_type.rod_training_type_selector == "not_training_truth_known":
38 -d "--resource:${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[$rod_bind_name]}"
39 #else:
40 -d "--resource:${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]}"
41 #end if
42 #end for
43
44 ##start standard gatk options
45 #if $gatk_param_type.gatk_param_type_selector == "advanced":
46 #for $pedigree in $gatk_param_type.pedigree:
47 -p '--pedigree "${pedigree.pedigree_file}"'
48 #end for
49 #for $pedigree_string in $gatk_param_type.pedigree_string_repeat:
50 -p '--pedigreeString "${pedigree_string.pedigree_string}"'
51 #end for
52 -p '--pedigreeValidationType "${gatk_param_type.pedigree_validation_type}"'
53 #for $read_filter in $gatk_param_type.read_filter:
54 -p '--read_filter "${read_filter.read_filter_type.read_filter_type_selector}"
55 ###raise Exception( str( dir( $read_filter ) ) )
56 #for $name, $param in $read_filter.read_filter_type.iteritems():
57 #if $name not in [ "__current_case__", "read_filter_type_selector" ]:
58 #if hasattr( $param.input, 'truevalue' ):
59 ${param}
60 #else:
61 --${name} "${param}"
62 #end if
63 #end if
64 #end for
65 '
66 #end for
67 #for $interval_count, $input_intervals in enumerate( $gatk_param_type.input_interval_repeat ):
68 -d "--intervals" "${input_intervals.input_intervals}" "${input_intervals.input_intervals.ext}" "input_intervals_${interval_count}"
69 #end for
70
71 #for $interval_count, $input_intervals in enumerate( $gatk_param_type.input_exclude_interval_repeat ):
72 -d "--excludeIntervals" "${input_intervals.input_exclude_intervals}" "${input_intervals.input_exclude_intervals.ext}" "input_exlude_intervals_${interval_count}"
73 #end for
74
75 -p '--interval_set_rule "${gatk_param_type.interval_set_rule}"'
76
77 -p '--downsampling_type "${gatk_param_type.downsampling_type.downsampling_type_selector}"'
78 #if str( $gatk_param_type.downsampling_type.downsampling_type_selector ) != "NONE":
79 -p '--${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_type_selector} "${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_value}"'
80 #end if
81 -p '
82 --baq "${gatk_param_type.baq}"
83 --baqGapOpenPenalty "${gatk_param_type.baq_gap_open_penalty}"
84 ${gatk_param_type.use_original_qualities}
85 --defaultBaseQualities "${gatk_param_type.default_base_qualities}"
86 --validation_strictness "${gatk_param_type.validation_strictness}"
87 --interval_merging "${gatk_param_type.interval_merging}"
88 ${gatk_param_type.disable_experimental_low_memory_sharding}
89 ${gatk_param_type.non_deterministic_random_seed}
90 '
91 #for $rg_black_list_count, $rg_black_list in enumerate( $gatk_param_type.read_group_black_list_repeat ):
92 #if $rg_black_list.read_group_black_list_type.read_group_black_list_type_selector == "file":
93 -d "--read_group_black_list" "${rg_black_list.read_group_black_list_type.read_group_black_list}" "txt" "input_read_group_black_list_${rg_black_list_count}"
94 #else
95 -p '--read_group_black_list "${rg_black_list.read_group_black_list_type.read_group_black_list}"'
96 #end if
97 #end for
98 #end if
99
100 #if str( $reference_source.reference_source_selector ) == "history":
101 -d "-R" "${reference_source.ref_file}" "${reference_source.ref_file.ext}" "gatk_input"
102 #end if
103 ##end standard gatk options
104
105 ##start analysis specific options
106 -p '
107 #if str( $annotations ) != "None":
108 #for $annotation in str( $annotations.fields.gatk_value ).split( ',' ):
109 --use_annotation "${annotation}"
110 #end for
111 #end if
112 #for $additional_annotation in $additional_annotations:
113 --use_annotation "${additional_annotation.additional_annotation_name}"
114 #end for
115 --mode "${mode}"
116 '
117
118 #if $analysis_param_type.analysis_param_type_selector == "advanced":
119 -p '
120 --maxGaussians "${analysis_param_type.max_gaussians}"
121 --maxIterations "${analysis_param_type.max_iterations}"
122 --numKMeans "${analysis_param_type.num_k_means}"
123 --stdThreshold "${analysis_param_type.std_threshold}"
124 --qualThreshold "${analysis_param_type.qual_threshold}"
125 --shrinkage "${analysis_param_type.shrinkage}"
126 --dirichlet "${analysis_param_type.dirichlet}"
127 --priorCounts "${analysis_param_type.prior_counts}"
128 #if str( $analysis_param_type.bad_variant_selector.bad_variant_selector_type ) == 'percent':
129 --percentBadVariants "${analysis_param_type.bad_variant_selector.percent_bad_variants}"
130 #else:
131 --minNumBadVariants "${analysis_param_type.bad_variant_selector.min_num_bad_variants}"
132 #end if
133 --target_titv "${analysis_param_type.target_titv}"
134 #for $tranche in [ $tranche.strip() for $tranche in str( $analysis_param_type.ts_tranche ).split( ',' ) if $tranche.strip() ]
135 --TStranche "${tranche}"
136 #end for
137 #for $ignore_filter in $analysis_param_type.ignore_filters:
138 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.ignore_filter_type_selector )
139 #if $ignore_filter_name == "custom":
140 #set $ignore_filter_name = str( $ignore_filter.ignore_filter_type.filter_name )
141 #end if
142 --ignore_filter "${ignore_filter_name}"
143 #end for
144 --ts_filter_level "${analysis_param_type.ts_filter_level}"
145 '
146 #end if
147
148
149 &amp;&amp;
150 mv "${output_rscript}.pdf" "${output_tranches_pdf}"
151
152 </command>
153 <inputs>
154 <conditional name="reference_source">
155 <param name="reference_source_selector" type="select" label="Choose the source for the reference list">
156 <option value="cached">Locally cached</option>
157 <option value="history">History</option>
158 </param>
159 <when value="cached">
160 <repeat name="variants" title="Variant" min="1" help="-input,--input &amp;lt;input&amp;gt;">
161 <param name="input_variants" type="data" format="vcf" label="Variant file to recalibrate" />
162 </repeat>
163 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;">
164 <options from_data_table="gatk_picard_indexes">
165 <!-- <filter type="data_meta" key="dbkey" ref="variants[0].input_variants" column="dbkey"/> -->
166 </options>
167 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
168 </param>
169 </when>
170 <when value="history"> <!-- FIX ME!!!! -->
171 <repeat name="variants" title="Variant" min="1" help="-input,--input &amp;lt;input&amp;gt;">
172 <param name="input_variants" type="data" format="vcf" label="Variant file to recalibrate" />
173 </repeat>
174 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;" />
175 </when>
176 </conditional>
177
178 <repeat name="rod_bind" title="Binding for reference-ordered data" help="-resource,--resource &amp;lt;resource&amp;gt;">
179 <conditional name="rod_bind_type">
180 <param name="rod_bind_type_selector" type="select" label="Binding Type">
181 <option value="dbsnp" selected="True">dbSNP</option>
182 <option value="variant">Variants</option>
183 <option value="snps">SNPs</option>
184 <option value="indels">INDELs</option>
185 <option value="hapmap">HapMap</option>
186 <option value="omni">OMNI</option>
187 <option value="mask">Mask</option>
188 <option value="custom">Custom</option>
189 <option value="comp">Comp</option>
190 </param>
191 <when value="variant">
192 <param name="input_rod" type="data" format="vcf" label="Variant ROD file" />
193 <conditional name="rod_training_type">
194 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
195 <option value="is_training_truth_known">Set training/truth/known sites</option>
196 <option value="not_training_truth_known" selected="True">Don't Set options</option>
197 </param>
198 <when value="not_training_truth_known">
199 <!-- do nothing here -->
200 </when>
201 <when value="is_training_truth_known">
202 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
203 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
204 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
205 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
206 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
207 </when>
208 </conditional>
209 </when>
210 <when value="comp">
211 <param name="input_rod" type="data" format="vcf" label="ROD file" />
212 <param name="custom_rod_name" type="text" value="Unnamed" label="ROD Name"/>
213 <conditional name="rod_training_type">
214 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
215 <option value="is_training_truth_known">Set training/truth/known sites</option>
216 <option value="not_training_truth_known" selected="True">Don't Set options</option>
217 </param>
218 <when value="not_training_truth_known">
219 <!-- do nothing here -->
220 </when>
221 <when value="is_training_truth_known">
222 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
223 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
224 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
225 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
226 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
227 </when>
228 </conditional>
229 </when>
230 <when value="mask">
231 <param name="input_rod" type="data" format="vcf" label="ROD file" />
232 <conditional name="rod_training_type">
233 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
234 <option value="is_training_truth_known">Set training/truth/known sites</option>
235 <option value="not_training_truth_known" selected="True">Don't Set options</option>
236 </param>
237 <when value="not_training_truth_known">
238 <!-- do nothing here -->
239 </when>
240 <when value="is_training_truth_known">
241 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
242 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
243 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
244 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
245 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
246 </when>
247 </conditional>
248 </when>
249 <when value="dbsnp">
250 <param name="input_rod" type="data" format="vcf" label="ROD file" />
251 <conditional name="rod_training_type">
252 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
253 <option value="is_training_truth_known">Set training/truth/known sites</option>
254 <option value="not_training_truth_known" selected="True">Don't Set options</option>
255 </param>
256 <when value="not_training_truth_known">
257 <!-- do nothing here -->
258 </when>
259 <when value="is_training_truth_known">
260 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
261 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
262 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
263 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
264 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
265 </when>
266 </conditional>
267 </when>
268 <when value="snps">
269 <param name="input_rod" type="data" format="vcf" label="ROD file" />
270 <conditional name="rod_training_type">
271 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
272 <option value="is_training_truth_known">Set training/truth/known sites</option>
273 <option value="not_training_truth_known" selected="True">Don't Set options</option>
274 </param>
275 <when value="not_training_truth_known">
276 <!-- do nothing here -->
277 </when>
278 <when value="is_training_truth_known">
279 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
280 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
281 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
282 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
283 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
284 </when>
285 </conditional>
286 </when>
287 <when value="hapmap">
288 <param name="input_rod" type="data" format="vcf" label="ROD file" />
289 <conditional name="rod_training_type">
290 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
291 <option value="is_training_truth_known">Set training/truth/known sites</option>
292 <option value="not_training_truth_known" selected="True">Don't Set options</option>
293 </param>
294 <when value="not_training_truth_known">
295 <!-- do nothing here -->
296 </when>
297 <when value="is_training_truth_known">
298 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
299 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
300 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
301 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
302 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
303 </when>
304 </conditional>
305 </when>
306 <when value="omni">
307 <param name="input_rod" type="data" format="vcf" label="ROD file" />
308 <conditional name="rod_training_type">
309 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
310 <option value="is_training_truth_known">Set training/truth/known sites</option>
311 <option value="not_training_truth_known" selected="True">Don't Set options</option>
312 </param>
313 <when value="not_training_truth_known">
314 <!-- do nothing here -->
315 </when>
316 <when value="is_training_truth_known">
317 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
318 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
319 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
320 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
321 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
322 </when>
323 </conditional>
324 </when>
325 <when value="indels">
326 <param name="input_rod" type="data" format="vcf" label="ROD file" />
327 <conditional name="rod_training_type">
328 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
329 <option value="is_training_truth_known">Set training/truth/known sites</option>
330 <option value="not_training_truth_known" selected="True">Don't Set options</option>
331 </param>
332 <when value="not_training_truth_known">
333 <!-- do nothing here -->
334 </when>
335 <when value="is_training_truth_known">
336 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
337 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
338 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
339 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
340 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
341 </when>
342 </conditional>
343 </when>
344 <when value="custom">
345 <param name="custom_rod_name" type="text" value="Unknown" label="ROD Name"/>
346 <param name="input_rod" type="data" format="vcf" label="ROD file" />
347 <conditional name="rod_training_type">
348 <param name="rod_training_type_selector" type="select" label="Use as training/truth/known sites">
349 <option value="is_training_truth_known">Set training/truth/known sites</option>
350 <option value="not_training_truth_known" selected="True">Don't Set options</option>
351 </param>
352 <when value="not_training_truth_known">
353 <!-- do nothing here -->
354 </when>
355 <when value="is_training_truth_known">
356 <param name="known" type="boolean" label="Is Known Site" truevalue="true" falsevalue="false"/>
357 <param name="training" type="boolean" label="Is Training Site" truevalue="true" falsevalue="false"/>
358 <param name="truth" type="boolean" label="Is Truth Site" truevalue="true" falsevalue="false"/>
359 <param name="bad" type="boolean" label="Is Bad Site" truevalue="true" falsevalue="false"/>
360 <param name="prior" type="float" label="prior probability of being true" value="12.0"/>
361 </when>
362 </conditional>
363 </when>
364 </conditional>
365 </repeat>
366
367 <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;">
368 <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs -->
369 <options from_data_table="gatk2_annotations">
370 <filter type="multiple_splitter" column="tools_valid_for" separator=","/>
371 <filter type="static_value" value="VariantRecalibrator" column="tools_valid_for"/>
372 </options>
373 </param>
374
375 <repeat name="additional_annotations" title="Additional annotation" help="-an,--use_annotation &amp;lt;use_annotation&amp;gt;">
376 <param name="additional_annotation_name" type="text" value="" label="Annotation name" />
377 </repeat>
378
379 <param name="mode" type="select" label="Recalibration mode" help="-mode,--mode &amp;lt;mode&amp;gt;">
380 <option value="SNP" selected="True">SNP</option>
381 <option value="INDEL">INDEL</option>
382 <option value="BOTH">BOTH</option>
383 </param>
384
385 <conditional name="gatk_param_type">
386 <param name="gatk_param_type_selector" type="select" label="Basic or Advanced GATK options">
387 <option value="basic" selected="True">Basic</option>
388 <option value="advanced">Advanced</option>
389 </param>
390 <when value="basic">
391 <!-- Do nothing here -->
392 </when>
393 <when value="advanced">
394 <repeat name="pedigree" title="Pedigree file" help="-ped,--pedigree &amp;lt;pedigree&amp;gt;">
395 <param name="pedigree_file" type="data" format="txt" label="Pedigree files for samples"/>
396 </repeat>
397 <repeat name="pedigree_string_repeat" title="Pedigree string" help="-pedString,--pedigreeString &amp;lt;pedigreeString&amp;gt;">
398 <param name="pedigree_string" type="text" value="" label="Pedigree string for samples"/>
399 </repeat>
400 <param name="pedigree_validation_type" type="select" label="How strict should we be in validating the pedigree information" help="-pedValidationType,--pedigreeValidationType &amp;lt;pedigreeValidationType&amp;gt;">
401 <option value="STRICT" selected="True">STRICT</option>
402 <option value="SILENT">SILENT</option>
403 </param>
404 <repeat name="read_filter" title="Read Filter" help="-rf,--read_filter &amp;lt;read_filter&amp;gt;">
405 <conditional name="read_filter_type">
406 <param name="read_filter_type_selector" type="select" label="Read Filter Type">
407 <option value="BadCigar">BadCigar</option>
408 <option value="BadMate">BadMate</option>
409 <option value="DuplicateRead">DuplicateRead</option>
410 <option value="FailsVendorQualityCheck">FailsVendorQualityCheck</option>
411 <option value="MalformedRead">MalformedRead</option>
412 <option value="MappingQuality">MappingQuality</option>
413 <option value="MappingQualityUnavailable">MappingQualityUnavailable</option>
414 <option value="MappingQualityZero">MappingQualityZero</option>
415 <option value="MateSameStrand">MateSameStrand</option>
416 <option value="MaxInsertSize">MaxInsertSize</option>
417 <option value="MaxReadLength" selected="True">MaxReadLength</option>
418 <option value="MissingReadGroup">MissingReadGroup</option>
419 <option value="NoOriginalQualityScores">NoOriginalQualityScores</option>
420 <option value="NotPrimaryAlignment">NotPrimaryAlignment</option>
421 <option value="Platform454">Platform454</option>
422 <option value="Platform">Platform</option>
423 <option value="PlatformUnit">PlatformUnit</option>
424 <option value="ReadGroupBlackList">ReadGroupBlackList</option>
425 <option value="ReadName">ReadName</option>
426 <option value="ReadStrand">ReadStrand</option>
427 <option value="ReassignMappingQuality">ReassignMappingQuality</option>
428 <option value="Sample">Sample</option>
429 <option value="SingleReadGroup">SingleReadGroup</option>
430 <option value="UnmappedRead">UnmappedRead</option>
431 </param>
432 <when value="BadCigar">
433 <!-- no extra options -->
434 </when>
435 <when value="BadMate">
436 <!-- no extra options -->
437 </when>
438 <when value="DuplicateRead">
439 <!-- no extra options -->
440 </when>
441 <when value="FailsVendorQualityCheck">
442 <!-- no extra options -->
443 </when>
444 <when value="MalformedRead">
445 <!-- no extra options -->
446 </when>
447 <when value="MappingQuality">
448 <param name="min_mapping_quality_score" type="integer" value="10" label="Minimum read mapping quality required to consider a read for calling"/>
449 </when>
450 <when value="MappingQualityUnavailable">
451 <!-- no extra options -->
452 </when>
453 <when value="MappingQualityZero">
454 <!-- no extra options -->
455 </when>
456 <when value="MateSameStrand">
457 <!-- no extra options -->
458 </when>
459 <when value="MaxInsertSize">
460 <param name="maxInsertSize" type="integer" value="1000000" label="Discard reads with insert size greater than the specified value"/>
461 </when>
462 <when value="MaxReadLength">
463 <param name="maxReadLength" type="integer" value="76" label="Max Read Length"/>
464 </when>
465 <when value="MissingReadGroup">
466 <!-- no extra options -->
467 </when>
468 <when value="NoOriginalQualityScores">
469 <!-- no extra options -->
470 </when>
471 <when value="NotPrimaryAlignment">
472 <!-- no extra options -->
473 </when>
474 <when value="Platform454">
475 <!-- no extra options -->
476 </when>
477 <when value="Platform">
478 <param name="PLFilterName" type="text" value="" label="Discard reads with RG:PL attribute containing this string"/>
479 </when>
480 <when value="PlatformUnit">
481 <!-- no extra options -->
482 </when>
483 <when value="ReadGroupBlackList">
484 <!-- no extra options -->
485 </when>
486 <when value="ReadName">
487 <param name="readName" type="text" value="" label="Filter out all reads except those with this read name"/>
488 </when>
489 <when value="ReadStrand">
490 <param name="filterPositive" type="boolean" truevalue="--filterPositive" falsevalue="" label="Discard reads on the forward strand"/>
491 </when>
492 <when value="ReassignMappingQuality">
493 <param name="default_mapping_quality" type="integer" value="60" label="Default read mapping quality to assign to all reads"/>
494 </when>
495 <when value="Sample">
496 <param name="sample_to_keep" type="text" value="" label="The name of the sample(s) to keep, filtering out all others"/>
497 </when>
498 <when value="SingleReadGroup">
499 <param name="read_group_to_keep" type="integer" value="76" label="The name of the read group to keep, filtering out all others"/>
500 </when>
501 <when value="UnmappedRead">
502 <!-- no extra options -->
503 </when>
504 </conditional>
505 </repeat>
506 <repeat name="input_interval_repeat" title="Operate on Genomic intervals" help="-L,--intervals &amp;lt;intervals&amp;gt;">
507 <param name="input_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" />
508 </repeat>
509 <repeat name="input_exclude_interval_repeat" title="Exclude Genomic intervals" help="-XL,--excludeIntervals &amp;lt;excludeIntervals&amp;gt;">
510 <param name="input_exclude_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" />
511 </repeat>
512
513 <param name="interval_set_rule" type="select" label="Interval set rule" help="-isr,--interval_set_rule &amp;lt;interval_set_rule&amp;gt;">
514 <option value="UNION" selected="True">UNION</option>
515 <option value="INTERSECTION">INTERSECTION</option>
516 </param>
517
518 <conditional name="downsampling_type">
519 <param name="downsampling_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="-dt,--downsampling_type &amp;lt;downsampling_type&amp;gt;">
520 <option value="NONE" selected="True">NONE</option>
521 <option value="ALL_READS">ALL_READS</option>
522 <option value="BY_SAMPLE">BY_SAMPLE</option>
523 </param>
524 <when value="NONE">
525 <!-- no more options here -->
526 </when>
527 <when value="ALL_READS">
528 <conditional name="downsample_to_type">
529 <param name="downsample_to_type_selector" type="select" label="Downsample method">
530 <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
531 <option value="downsample_to_coverage">Downsample by Coverage</option>
532 </param>
533 <when value="downsample_to_fraction">
534 <param name="downsample_to_value" type="float" label="Fraction [0.0-1.0] of reads to downsample to" value="1" min="0" max="1" help="-dfrac,--downsample_to_fraction &amp;lt;downsample_to_fraction&amp;gt;"/>
535 </when>
536 <when value="downsample_to_coverage">
537 <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0" help="-dcov,--downsample_to_coverage &amp;lt;downsample_to_coverage&amp;gt;"/>
538 </when>
539 </conditional>
540 </when>
541 <when value="BY_SAMPLE">
542 <conditional name="downsample_to_type">
543 <param name="downsample_to_type_selector" type="select" label="Downsample method">
544 <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
545 <option value="downsample_to_coverage">Downsample by Coverage</option>
546 </param>
547 <when value="downsample_to_fraction">
548 <param name="downsample_to_value" type="float" label="Fraction [0.0-1.0] of reads to downsample to" value="1" min="0" max="1" help="-dfrac,--downsample_to_fraction &amp;lt;downsample_to_fraction&amp;gt;"/>
549 </when>
550 <when value="downsample_to_coverage">
551 <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0" help="-dcov,--downsample_to_coverage &amp;lt;downsample_to_coverage&amp;gt;"/>
552 </when>
553 </conditional>
554 </when>
555 </conditional>
556 <param name="baq" type="select" label="Type of BAQ calculation to apply in the engine" help="-baq,--baq &amp;lt;baq&amp;gt;">
557 <option value="OFF" selected="True">OFF</option>
558 <option value="CALCULATE_AS_NECESSARY">CALCULATE_AS_NECESSARY</option>
559 <option value="RECALCULATE">RECALCULATE</option>
560 </param>
561 <param name="baq_gap_open_penalty" type="float" label="BAQ gap open penalty (Phred Scaled)" value="40" help="Default value is 40. 30 is perhaps better for whole genome call sets. -baqGOP,--baqGapOpenPenalty &amp;lt;baqGapOpenPenalty&amp;gt;" />
562 <param name="use_original_qualities" type="boolean" truevalue="--useOriginalQualities" falsevalue="" label="Use the original base quality scores from the OQ tag" help="-OQ,--useOriginalQualities" />
563 <param name="default_base_qualities" type="integer" label="Value to be used for all base quality scores, when some are missing" value="-1" help="-DBQ,--defaultBaseQualities &amp;lt;defaultBaseQualities&amp;gt;"/>
564 <param name="validation_strictness" type="select" label="How strict should we be with validation" help="-S,--validation_strictness &amp;lt;validation_strictness&amp;gt;">
565 <option value="STRICT" selected="True">STRICT</option>
566 <option value="LENIENT">LENIENT</option>
567 <option value="SILENT">SILENT</option>
568 <!-- <option value="DEFAULT_STRINGENCY">DEFAULT_STRINGENCY</option> listed in docs, but not valid value...-->
569 </param>
570 <param name="interval_merging" type="select" label="Interval merging rule" help="-im,--interval_merging &amp;lt;interval_merging&amp;gt;">
571 <option value="ALL" selected="True">ALL</option>
572 <option value="OVERLAPPING_ONLY">OVERLAPPING_ONLY</option>
573 </param>
574
575 <repeat name="read_group_black_list_repeat" title="Read group black list" help="-rgbl,--read_group_black_list &amp;lt;read_group_black_list&amp;gt;">
576 <conditional name="read_group_black_list_type">
577 <param name="read_group_black_list_type_selector" type="select" label="Type of reads read group black list">
578 <option value="file" selected="True">Filters in file</option>
579 <option value="text">Specify filters as a string</option>
580 </param>
581 <when value="file">
582 <param name="read_group_black_list" type="data" format="txt" label="Read group black list file" />
583 </when>
584 <when value="text">
585 <param name="read_group_black_list" type="text" value="tag:string" label="Read group black list tag:string" />
586 </when>
587 </conditional>
588 </repeat>
589
590 <param name="disable_experimental_low_memory_sharding" type="boolean" truevalue="--disable_experimental_low_memory_sharding" falsevalue="" label="Disable experimental low-memory sharding functionality." checked="False" help="--disable_experimental_low_memory_sharding"/>
591 <param name="non_deterministic_random_seed" type="boolean" truevalue="--nonDeterministicRandomSeed" falsevalue="" label="Makes the GATK behave non deterministically, that is, the random numbers generated will be different in every run" checked="False" help="-ndrs,--nonDeterministicRandomSeed"/>
592
593 </when>
594 </conditional>
595
596 <conditional name="analysis_param_type">
597 <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options">
598 <option value="basic" selected="True">Basic</option>
599 <option value="advanced">Advanced</option>
600 </param>
601 <when value="basic">
602 <!-- Do nothing here -->
603 </when>
604 <when value="advanced">
605 <param name="max_gaussians" type="integer" label="maximum number of Gaussians to try during variational Bayes Algorithm" value="10" help="-mG,--maxGaussians &amp;lt;maxGaussians&amp;gt;"/>
606 <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 &amp;lt;maxIterations&amp;gt;"/>
607 <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 &amp;lt;numKMeans&amp;gt;"/>
608 <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 &amp;lt;stdThreshold&amp;gt;"/>
609 <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 &amp;lt;qualThreshold&amp;gt;"/>
610 <param name="shrinkage" type="float" label="shrinkage parameter in variational Bayes algorithm" value="1.0" help="-shrinkage,--shrinkage &amp;lt;shrinkage&amp;gt;"/>
611 <param name="dirichlet" type="float" label="dirichlet parameter in variational Bayes algorithm" value="0.001" help="-dirichlet,--dirichlet &amp;lt;dirichlet&amp;gt;"/>
612 <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;"/>
613 <conditional name="bad_variant_selector">
614 <param name="bad_variant_selector_type" type="select" label="How to specify bad variants">
615 <option value="percent" selected="True">Percent</option>
616 <option value="min_num">Number</option>
617 </param>
618 <when value="percent">
619 <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 &amp;lt;percentBadVariants&amp;gt;"/>
620 </when>
621 <when value="min_num">
622 <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 &amp;lt;minNumBadVariants&amp;gt;"/>
623 </when>
624 </conditional>
625 <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;"/>
626 <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;"/>
627 <repeat name="ignore_filters" title="Ignore Filter" help="-ignoreFilter,--ignore_filter &amp;lt;ignore_filter&amp;gt;">
628 <conditional name="ignore_filter_type">
629 <param name="ignore_filter_type_selector" type="select" label="Filter Type">
630 <option value="HARD_TO_VALIDATE">HARD_TO_VALIDATE</option>
631 <option value="LowQual" >LowQual</option>
632 <option value="custom" selected="True">Other</option>
633 </param>
634 <when value="custom">
635 <param name="filter_name" type="text" value="" label="Filter name"/>
636 </when>
637 <when value="HARD_TO_VALIDATE" />
638 <when value="LowQual" />
639 </conditional>
640 </repeat>
641 <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 &amp;lt;ts_filter_level&amp;gt;"/>
642 </when>
643 </conditional>
644 </inputs>
645 <outputs>
646 <data format="gatk_recal" name="output_recal" label="${tool.name} on ${on_string} (Recalibration File)" />
647 <data format="gatk_tranche" name="output_tranches" label="${tool.name} on ${on_string} (Tranches File)" />
648 <data format="txt" name="output_rscript" label="${tool.name} on ${on_string} (RScript File)" />
649 <data format="pdf" name="output_tranches_pdf" label="${tool.name} on ${on_string} (PDF File)" />
650 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
651 </outputs>
652 <tests>
653 <!-- ADD TESTS -->
654 </tests>
655 <help>
656 **What it does**
657
658 Takes variant calls as .vcf files, learns a Gaussian mixture model over the variant annotations and evaluates the variant -- assigning an informative lod score
659
660 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;`_.
661
662 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;`_.
663
664 If you encounter errors, please view the `GATK FAQ &lt;http://www.broadinstitute.org/gatk/guide/topic?name=faqs&gt;`_.
665
666 ------
667
668 **Inputs**
669
670 GenomeAnalysisTK: VariantRecalibrator accepts a variant input file.
671
672
673 **Outputs**
674
675 The output is in VCF format.
676
677
678 Go `here &lt;http://www.broadinstitute.org/gatk/guide/topic?name=intro&gt;`_ for details on GATK file formats.
679
680 -------
681
682 **Settings**::
683
684
685 tranches_file The output tranches file used by ApplyRecalibration
686 use_annotation The names of the annotations which should used for calculations
687 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)
688 maxGaussians The maximum number of Gaussians to try during variational Bayes algorithm
689 maxIterations The maximum number of VBEM iterations to be performed in variational Bayes algorithm. Procedure will normally end when convergence is detected.
690 numKMeans The number of k-means iterations to perform in order to initialize the means of the Gaussians in the Gaussian mixture model.
691 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.
692 qualThreshold If a known variant has raw QUAL value less than -qual then don't use it for building the Gaussian mixture model.
693 shrinkage The shrinkage parameter in variational Bayes algorithm.
694 dirichlet The dirichlet parameter in variational Bayes algorithm.
695 priorCounts The number of prior counts to use in variational Bayes algorithm.
696 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.
697 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.
698 recal_file The output recal file used by ApplyRecalibration
699 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!
700 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)
701 ignore_filter If specified the optimizer will use variants even if the specified filter name is marked in the input VCF file
702 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
703 rscript_file The output rscript file generated by the VQSR to aid in visualization of the input data and learned model
704 path_to_resources Path to resources folder holding the Sting R scripts.
705 ts_filter_level The truth sensitivity level at which to start filtering, used here to indicate filtered variants in plots
706
707 ------
708
709 **Citation**
710
711 For the underlying tool, please cite `DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, Philippakis AA, del Angel G, Rivas MA, Hanna M, McKenna A, Fennell TJ, Kernytsky AM, Sivachenko AY, Cibulskis K, Gabriel SB, Altshuler D, Daly MJ. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011 May;43(5):491-8. &lt;http://www.ncbi.nlm.nih.gov/pubmed/21478889&gt;`_
712
713 Please also site `McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K, Altshuler D, Gabriel S, Daly M, DePristo MA (2010). The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20:1297-303. Epub 2010 Jul 19. &lt;http://www.ncbi.nlm.nih.gov/pubmed/20644199&gt;`_
714
715 If you use this tool in Galaxy, please cite `Blankenberg D, Von Kuster G, Coraor N, Ananda G, Lazarus R, Mangan M, Nekrutenko A, Taylor J. Galaxy: a web-based genome analysis tool for experimentalists. Curr Protoc Mol Biol. 2010 Jan;Chapter 19:Unit 19.10.1-21. &lt;http://www.ncbi.nlm.nih.gov/pubmed/20069535&gt;`_
716
717 </help>
718 </tool>