Mercurial > repos > jjohnson > gatk2

comparison haplotype_caller.xml @ 11:5faf7ace8aee draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add HaplotypeCaller
author Jim Johnson <jj@umn.edu>
date Thu, 08 Nov 2012 10:38:18 -0600
parents
children b40671bc73d5
comparison
equal deleted inserted replaced
10:a14e79e7ac75 11:5faf7ace8aee
1 <tool id="gatk2_haplotype_caller" name="Haplotype Caller" version="0.0.4">
2 <description>Call SNPs and indels simultaneously via local de-novo assembly of haplotypes in an active region</description>
3 <requirements>
4 <requirement type="package" version="2.2">gatk</requirement>
5 <requirement type="package" version="0.1.18">samtools</requirement>
6 </requirements>
7 <command interpreter="python">gatk2_wrapper.py
8 --max_jvm_heap_fraction "1"
9 --stdout "${output_log}"
10 -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
11 #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
12 -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
13 #end if
14 -p 'java
15 -jar "\$GATK2_PATH/GenomeAnalysisTK.jar"
16 -T "HaplotypeCaller"
17 -o "${output_vcf}"
18 ## \$GATK2_SITE_OPTIONS
19 ##-et "NO_ET" -K "/data/galaxy/appList/GenomeAnalysisTK-2.0-36-gf5c1c1a/gatk2_key_file" ##ET no phone home
20 ##--num_threads 4 ##not supported yet
21 ##-log "${output_log}" ##don't use this to log to file, instead directly capture stdout
22 #if $reference_source.reference_source_selector != "history":
23 -R "${reference_source.ref_file.fields.path}"
24 #end if
25 #if str($input_recal) != 'None':
26 --BQSR "${input_recal}"
27 #end if
28 --disable_bam_indexing
29 '
30 ##start standard gatk options
31 #if $gatk_param_type.gatk_param_type_selector == "advanced":
32 #for $pedigree in $gatk_param_type.pedigree:
33 -p '--pedigree "${pedigree.pedigree_file}"'
34 #end for
35 #for $pedigree_string in $gatk_param_type.pedigree_string_repeat:
36 -p '--pedigreeString "${pedigree_string.pedigree_string}"'
37 #end for
38 -p '--pedigreeValidationType "${gatk_param_type.pedigree_validation_type}"'
39 #for $read_filter in $gatk_param_type.read_filter:
40 -p '--read_filter "${read_filter.read_filter_type.read_filter_type_selector}"
41 ###raise Exception( str( dir( $read_filter ) ) )
42 #for $name, $param in $read_filter.read_filter_type.iteritems():
43 #if $name not in [ "__current_case__", "read_filter_type_selector" ]:
44 #if hasattr( $param.input, 'truevalue' ):
45 ${param}
46 #else:
47 --${name} "${param}"
48 #end if
49 #end if
50 #end for
51 '
52 #end for
53 #for $interval_count, $input_intervals in enumerate( $gatk_param_type.input_interval_repeat ):
54 -d "--intervals" "${input_intervals.input_intervals}" "${input_intervals.input_intervals.ext}" "input_intervals_${interval_count}"
55 #end for
56
57 #for $interval_count, $input_intervals in enumerate( $gatk_param_type.input_exclude_interval_repeat ):
58 -d "--excludeIntervals" "${input_intervals.input_exclude_intervals}" "${input_intervals.input_exclude_intervals.ext}" "input_exlude_intervals_${interval_count}"
59 #end for
60
61 -p '--interval_set_rule "${gatk_param_type.interval_set_rule}"'
62
63 -p '--downsampling_type "${gatk_param_type.downsampling_type.downsampling_type_selector}"'
64 #if str( $gatk_param_type.downsampling_type.downsampling_type_selector ) != "NONE":
65 -p '--${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_type_selector} "${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_value}"'
66 #end if
67 -p '
68 --baq "${gatk_param_type.baq}"
69 --baqGapOpenPenalty "${gatk_param_type.baq_gap_open_penalty}"
70 ${gatk_param_type.use_original_qualities}
71 --defaultBaseQualities "${gatk_param_type.default_base_qualities}"
72 --validation_strictness "${gatk_param_type.validation_strictness}"
73 --interval_merging "${gatk_param_type.interval_merging}"
74 ${gatk_param_type.disable_experimental_low_memory_sharding}
75 ${gatk_param_type.non_deterministic_random_seed}
76 '
77 #for $rg_black_list_count, $rg_black_list in enumerate( $gatk_param_type.read_group_black_list_repeat ):
78 #if $rg_black_list.read_group_black_list_type.read_group_black_list_type_selector == "file":
79 -d "--read_group_black_list" "${rg_black_list.read_group_black_list_type.read_group_black_list}" "txt" "input_read_group_black_list_${rg_black_list_count}"
80 #else
81 -p '--read_group_black_list "${rg_black_list.read_group_black_list_type.read_group_black_list}"'
82 #end if
83 #end for
84 #end if
85
86 #if str( $reference_source.reference_source_selector ) == "history":
87 -d "-R" "${reference_source.ref_file}" "${reference_source.ref_file.ext}" "gatk_input"
88 #end if
89 ##end standard gatk options
90
91 ##start analysis specific options
92 #if $analysis_param_type.analysis_param_type_selector == "advanced":
93 -p '
94 ## files
95 #if str($analysis_param_type.activeRegionIn) != 'None':
96 --activeRegionIn "$analysis_param_type.activeRegionIn"
97 #end if
98 #if str($analysis_param_type.alleles) != 'None':
99 --alleles "$analysis_param_type.alleles"
100 #end if
101 #if str($analysis_param_type.comp) != 'None':
102 --comp "$analysis_param_type.comp"
103 #end if
104 #if str($analysis_param_type.dbsnp) != 'None':
105 --dbsnp "$analysis_param_type.dbsnp"
106 #end if
107 ## text
108 #if len($analysis_param_type.annotation.__str__) > 0:
109 --annotation $analysis_param_type.annotation
110 #end if
111 #if len($analysis_param_type.excludeAnnotation.__str__) > 0:
112 --excludeAnnotation $analysis_param_type.excludeAnnotation
113 #end if
114 #if len($analysis_param_type.group.__str__) > 0:
115 --group $analysis_param_type.group
116 #end if
117 ## value setings
118 #if $analysis_param_type.contamination_fraction_to_filter.__str__.strip() != '':
119 --contamination_fraction_to_filter $analysis_param_type.contamination_fraction_to_filter
120 #end if
121 #if $analysis_param_type.downsampleRegion.__str__.strip() != '':
122 --downsampleRegion $analysis_param_type.downsampleRegion
123 #end if
124 #if $analysis_param_type.heterozygosity.__str__.strip() != '':
125 --heterozygosity $analysis_param_type.heterozygosity
126 #end if
127 #if $analysis_param_type.minPruning.__str__.strip() != '':
128 --minPruning $analysis_param_type.minPruning
129 #end if
130 #if $analysis_param_type.standard_min_confidence_threshold_for_calling.__str__.strip() != '':
131 --standard_min_confidence_threshold_for_calling $analysis_param_type.standard_min_confidence_threshold_for_calling
132 #end if
133 #if $analysis_param_type.standard_min_confidence_threshold_for_emitting.__str__.strip() != '':
134 --standard_min_confidence_threshold_for_emitting $analysis_param_type.standard_min_confidence_threshold_for_emitting
135 #end if
136 #if $analysis_param_type.gcpHMM.__str__.strip() != '':
137 --gcpHMM $analysis_param_type.gcpHMM
138 #end if
139 #if $analysis_param_type.max_alternate_alleles.__str__.strip() != '':
140 --max_alternate_alleles $analysis_param_type.max_alternate_alleles
141 #end if
142 ## mode selections
143 #if $analysis_param_type.genotyping_mode.__str__ != "None" and len($analysis_param_type.genotyping_mode.__str__) > 0:
144 --genotyping_mode $analysis_param_type.genotyping_mode
145 #end if
146 #if $analysis_param_type.output_mode.__str__ != "None" and len($analysis_param_type.output_mode.__str__) > 0:
147 --output_mode $analysis_param_type.output_mode
148 #end if
149 #if $analysis_param_type.pair_hmm_implementation.__str__ != "None" and len($analysis_param_type.pair_hmm_implementation.__str__) > 0:
150 --pair_hmm_implementation $analysis_param_type.pair_hmm_implementation
151 #end if
152 #if $analysis_param_type.p_nonref_model.__str__ != "None" and len($analysis_param_type.p_nonref_model.__str__) > 0:
153 --p_nonref_model $analysis_param_type.p_nonref_model
154 #end if
155 ## optional outputs
156 #if $analysis_param_type.activeRegionOut:
157 --activeRegionOut $active_region_out
158 #end if
159 #if $analysis_param_type.graphOutput:
160 --graphOutput $graph_out
161 #end if
162 ## flags
163 $analysis_param_type.useAllelesTrigger
164 $analysis_param_type.fullHaplotype
165 $analysis_param_type.genotypeFullActiveRegion
166 $analysis_param_type.debug
167 '
168 #end if
169 </command>
170 <inputs>
171 <param name="input_recal" type="data" format="csv" optional="true" label="Covariates table recalibration file" help="-BQSR,--BQSR &amp;lt;recal_file&amp;gt;" >
172 <help>The input covariates table file which enables on-the-fly base quality score recalibration.
173 Enables on-the-fly recalibrate of base qualities. The covariates tables are produced by the BaseQualityScoreRecalibrator tool.
174 Please be aware that one should only run recalibration with the covariates file created on the same input bam(s).
175 </help>
176 </param>
177 <conditional name="reference_source">
178 <param name="reference_source_selector" type="select" label="Choose the source for the reference list">
179 <option value="cached">Locally cached</option>
180 <option value="history">History</option>
181 </param>
182 <when value="cached">
183 <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &amp;lt;input_file&amp;gt;">
184 <validator type="unspecified_build" />
185 <validator type="dataset_metadata_in_data_table" table_name="gatk2_picard_indexes" metadata_name="dbkey" metadata_column="dbkey" message="Sequences are not currently available for the specified build." /> <!-- fixme!!! this needs to be a select -->
186 </param>
187 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;" >
188 <options from_data_table="gatk2_picard_indexes">
189 <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/>
190 </options>
191 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
192 </param>
193 </when>
194 <when value="history">
195 <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &amp;lt;input_file&amp;gt;" />
196 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;">
197 <options>
198 <filter type="data_meta" key="dbkey" ref="input_bam" />
199 </options>
200 </param>
201 </when>
202 </conditional>
203
204 <conditional name="gatk_param_type">
205 <param name="gatk_param_type_selector" type="select" label="Basic or Advanced GATK options">
206 <option value="basic" selected="True">Basic</option>
207 <option value="advanced">Advanced</option>
208 </param>
209 <when value="basic">
210 <!-- Do nothing here -->
211 </when>
212 <when value="advanced">
213 <repeat name="pedigree" title="Pedigree file" help="-ped,--pedigree &amp;lt;pedigree&amp;gt;">
214 <param name="pedigree_file" type="data" format="txt" label="Pedigree files for samples"/>
215 </repeat>
216 <repeat name="pedigree_string_repeat" title="Pedigree string" help="-pedString,--pedigreeString &amp;lt;pedigreeString&amp;gt;">
217 <param name="pedigree_string" type="text" value="" label="Pedigree string for samples"/>
218 </repeat>
219 <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;">
220 <option value="STRICT" selected="True">STRICT</option>
221 <option value="SILENT">SILENT</option>
222 </param>
223 <repeat name="read_filter" title="Read Filter" help="-rf,--read_filter &amp;lt;read_filter&amp;gt;">
224 <conditional name="read_filter_type">
225 <param name="read_filter_type_selector" type="select" label="Read Filter Type">
226 <option value="BadCigar">BadCigar</option>
227 <option value="BadMate">BadMate</option>
228 <option value="DuplicateRead">DuplicateRead</option>
229 <option value="FailsVendorQualityCheck">FailsVendorQualityCheck</option>
230 <option value="MalformedRead">MalformedRead</option>
231 <option value="MappingQuality">MappingQuality</option>
232 <option value="MappingQualityUnavailable">MappingQualityUnavailable</option>
233 <option value="MappingQualityZero">MappingQualityZero</option>
234 <option value="MateSameStrand">MateSameStrand</option>
235 <option value="MaxInsertSize">MaxInsertSize</option>
236 <option value="MaxReadLength" selected="True">MaxReadLength</option>
237 <option value="MissingReadGroup">MissingReadGroup</option>
238 <option value="NoOriginalQualityScores">NoOriginalQualityScores</option>
239 <option value="NotPrimaryAlignment">NotPrimaryAlignment</option>
240 <option value="Platform454">Platform454</option>
241 <option value="Platform">Platform</option>
242 <option value="PlatformUnit">PlatformUnit</option>
243 <option value="ReadGroupBlackList">ReadGroupBlackList</option>
244 <option value="ReadName">ReadName</option>
245 <option value="ReadStrand">ReadStrand</option>
246 <option value="ReassignMappingQuality">ReassignMappingQuality</option>
247 <option value="Sample">Sample</option>
248 <option value="SingleReadGroup">SingleReadGroup</option>
249 <option value="UnmappedRead">UnmappedRead</option>
250 </param>
251 <when value="BadCigar">
252 <!-- no extra options -->
253 </when>
254 <when value="BadMate">
255 <!-- no extra options -->
256 </when>
257 <when value="DuplicateRead">
258 <!-- no extra options -->
259 </when>
260 <when value="FailsVendorQualityCheck">
261 <!-- no extra options -->
262 </when>
263 <when value="MalformedRead">
264 <!-- no extra options -->
265 </when>
266 <when value="MappingQuality">
267 <param name="min_mapping_quality_score" type="integer" value="10" label="Minimum read mapping quality required to consider a read for calling"/>
268 </when>
269 <when value="MappingQualityUnavailable">
270 <!-- no extra options -->
271 </when>
272 <when value="MappingQualityZero">
273 <!-- no extra options -->
274 </when>
275 <when value="MateSameStrand">
276 <!-- no extra options -->
277 </when>
278 <when value="MaxInsertSize">
279 <param name="maxInsertSize" type="integer" value="1000000" label="Discard reads with insert size greater than the specified value"/>
280 </when>
281 <when value="MaxReadLength">
282 <param name="maxReadLength" type="integer" value="76" label="Max Read Length"/>
283 </when>
284 <when value="MissingReadGroup">
285 <!-- no extra options -->
286 </when>
287 <when value="NoOriginalQualityScores">
288 <!-- no extra options -->
289 </when>
290 <when value="NotPrimaryAlignment">
291 <!-- no extra options -->
292 </when>
293 <when value="Platform454">
294 <!-- no extra options -->
295 </when>
296 <when value="Platform">
297 <param name="PLFilterName" type="text" value="" label="Discard reads with RG:PL attribute containing this string"/>
298 </when>
299 <when value="PlatformUnit">
300 <!-- no extra options -->
301 </when>
302 <when value="ReadGroupBlackList">
303 <!-- no extra options -->
304 </when>
305 <when value="ReadName">
306 <param name="readName" type="text" value="" label="Filter out all reads except those with this read name"/>
307 </when>
308 <when value="ReadStrand">
309 <param name="filterPositive" type="boolean" truevalue="--filterPositive" falsevalue="" label="Discard reads on the forward strand"/>
310 </when>
311 <when value="ReassignMappingQuality">
312 <param name="default_mapping_quality" type="integer" value="60" label="Default read mapping quality to assign to all reads"/>
313 </when>
314 <when value="Sample">
315 <param name="sample_to_keep" type="text" value="" label="The name of the sample(s) to keep, filtering out all others"/>
316 </when>
317 <when value="SingleReadGroup">
318 <param name="read_group_to_keep" type="integer" value="76" label="The name of the read group to keep, filtering out all others"/>
319 </when>
320 <when value="UnmappedRead">
321 <!-- no extra options -->
322 </when>
323 </conditional>
324 </repeat>
325 <repeat name="input_interval_repeat" title="Operate on Genomic intervals" help="-L,--intervals &amp;lt;intervals&amp;gt;">
326 <param name="input_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" />
327 </repeat>
328 <repeat name="input_exclude_interval_repeat" title="Exclude Genomic intervals" help="-XL,--excludeIntervals &amp;lt;excludeIntervals&amp;gt;">
329 <param name="input_exclude_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" />
330 </repeat>
331
332 <param name="interval_set_rule" type="select" label="Interval set rule" help="-isr,--interval_set_rule &amp;lt;interval_set_rule&amp;gt;">
333 <option value="UNION" selected="True">UNION</option>
334 <option value="INTERSECTION">INTERSECTION</option>
335 </param>
336
337 <conditional name="downsampling_type">
338 <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;">
339 <option value="NONE" selected="True">NONE</option>
340 <option value="ALL_READS">ALL_READS</option>
341 <option value="BY_SAMPLE">BY_SAMPLE</option>
342 </param>
343 <when value="NONE">
344 <!-- no more options here -->
345 </when>
346 <when value="ALL_READS">
347 <conditional name="downsample_to_type">
348 <param name="downsample_to_type_selector" type="select" label="Downsample method">
349 <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
350 <option value="downsample_to_coverage">Downsample by Coverage</option>
351 </param>
352 <when value="downsample_to_fraction">
353 <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;"/>
354 </when>
355 <when value="downsample_to_coverage">
356 <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;"/>
357 </when>
358 </conditional>
359 </when>
360 <when value="BY_SAMPLE">
361 <conditional name="downsample_to_type">
362 <param name="downsample_to_type_selector" type="select" label="Downsample method">
363 <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
364 <option value="downsample_to_coverage">Downsample by Coverage</option>
365 </param>
366 <when value="downsample_to_fraction">
367 <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;"/>
368 </when>
369 <when value="downsample_to_coverage">
370 <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;"/>
371 </when>
372 </conditional>
373 </when>
374 </conditional>
375 <param name="baq" type="select" label="Type of BAQ calculation to apply in the engine" help="-baq,--baq &amp;lt;baq&amp;gt;">
376 <option value="OFF" selected="True">OFF</option>
377 <option value="CALCULATE_AS_NECESSARY">CALCULATE_AS_NECESSARY</option>
378 <option value="RECALCULATE">RECALCULATE</option>
379 </param>
380 <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;" />
381 <param name="use_original_qualities" type="boolean" truevalue="--useOriginalQualities" falsevalue="" label="Use the original base quality scores from the OQ tag" help="-OQ,--useOriginalQualities" />
382 <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;"/>
383 <param name="validation_strictness" type="select" label="How strict should we be with validation" help="-S,--validation_strictness &amp;lt;validation_strictness&amp;gt;">
384 <option value="STRICT" selected="True">STRICT</option>
385 <option value="LENIENT">LENIENT</option>
386 <option value="SILENT">SILENT</option>
387 <!-- <option value="DEFAULT_STRINGENCY">DEFAULT_STRINGENCY</option> listed in docs, but not valid value...-->
388 </param>
389 <param name="interval_merging" type="select" label="Interval merging rule" help="-im,--interval_merging &amp;lt;interval_merging&amp;gt;">
390 <option value="ALL" selected="True">ALL</option>
391 <option value="OVERLAPPING_ONLY">OVERLAPPING_ONLY</option>
392 </param>
393
394 <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;">
395 <conditional name="read_group_black_list_type">
396 <param name="read_group_black_list_type_selector" type="select" label="Type of reads read group black list">
397 <option value="file" selected="True">Filters in file</option>
398 <option value="text">Specify filters as a string</option>
399 </param>
400 <when value="file">
401 <param name="read_group_black_list" type="data" format="txt" label="Read group black list file" />
402 </when>
403 <when value="text">
404 <param name="read_group_black_list" type="text" value="tag:string" label="Read group black list tag:string" />
405 </when>
406 </conditional>
407 </repeat>
408
409 <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"/>
410 <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"/>
411
412 </when>
413 </conditional>
414
415 <conditional name="analysis_param_type">
416 <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options">
417 <option value="basic" selected="True">Basic</option>
418 <option value="advanced">Advanced</option>
419 </param>
420 <when value="basic">
421 <!-- Do nothing here -->
422 </when>
423 <when value="advanced">
424
425 <conditional name="default_read_group_type">
426 <param name="default_read_group_type_selector" type="select" label="Set default Read Group" help="--default_read_group">
427 <option value="default" selected="True">Don't Set</option>
428 <option value="set">Set</option>
429 </param>
430 <when value="default">
431 <!-- do nothing here -->
432 </when>
433 <when value="set">
434 <param name="default_read_group" type="text" value="Unknown" label="If a read has no read group then default to the provided String"/>
435 </when>
436 </conditional>
437 <param name="default_platform" type="select" label="Set default Platform" help="--default_platform">
438 <option value="default" selected="True">Don't Set</option>
439 <option value="illumina">illumina</option>
440 <option value="454">454</option>
441 <option value="solid">solid</option>
442 </param>
443 <conditional name="force_read_group_type">
444 <param name="force_read_group_type_selector" type="select" label="Force Read Group" help="--force_read_group">
445 <option value="default" selected="True">Don't Force</option>
446 <option value="set">Force</option>
447 </param>
448 <when value="default">
449 <!-- do nothing here -->
450 </when>
451 <when value="set">
452 <param name="force_read_group" type="text" value="Unknown" label="If provided, the read group ID of EVERY read will be forced to be the provided String."/>
453 </when>
454 </conditional>
455 <param name="force_platform" type="select" label="Force Platform" help="--force_platform">
456 <option value="default" selected="True">Don't Force</option>
457 <option value="illumina">illumina</option>
458 <option value="454">454</option>
459 <option value="solid">solid</option>
460 </param>
461 <param name="exception_if_no_tile" type="boolean" checked="False" truevalue="--exception_if_no_tile" falsevalue="" label="Throw an exception when no tile can be found" help="--exception_if_no_tile"/>
462 <conditional name="solid_options_type">
463 <param name="solid_options_type_selector" type="select" label="Set SOLiD specific options">
464 <option value="default" selected="True">Don't Set</option>
465 <option value="set">Set</option>
466 </param>
467 <when value="default">
468 <!-- do nothing here -->
469 </when>
470 <when value="set">
471 <param name="solid_recal_mode" type="select" label="How should we recalibrate solid bases in which the reference was inserted" help="-sMode,--solid_recal_mode &amp;lt;solid_recal_mode&amp;gt;">
472 <option value="default" selected="True">Don't set</option>
473 <option value="DO_NOTHING">DO_NOTHING</option>
474 <option value="SET_Q_ZERO">SET_Q_ZERO</option>
475 <option value="SET_Q_ZERO_BASE_N">SET_Q_ZERO_BASE_N</option>
476 <option value="REMOVE_REF_BIAS">REMOVE_REF_BIAS</option>
477 </param>
478 <param name="solid_nocall_strategy" type="select" label="Behavior of the recalibrator when it encounters no calls" help="-solid_nocall_strategy,--solid_nocall_strategy &amp;lt;solid_nocall_strategy&amp;gt;">
479 <option value="default" selected="True">Don't set</option>
480 <option value="THROW_EXCEPTION">THROW_EXCEPTION</option>
481 <option value="LEAVE_READ_UNRECALIBRATED">LEAVE_READ_UNRECALIBRATED</option>
482 <option value="PURGE_READ">PURGE_READ</option>
483 </param>
484 </when>
485 </conditional>
486 <param name="simplify_bam" type="boolean" checked="False" truevalue="-simplifyBAM" falsevalue="" label="Simplify BAM" help="-simplifyBAM,--simplifyBAM"/>
487 <param name="window_size_nqs" type="integer" value="5" label="Window size used by MinimumNQSCovariate" help="--window_size_nqs"/>
488 <param name="homopolymer_nback" type="integer" value="7" label="Number of previous bases to look at in HomopolymerCovariate" help="-nback,--homopolymer_nback &amp;lt;homopolymer_nback&amp;gt;" />
489 <param name="preserve_qscores_less_than" type="integer" value="5" label="Bases with quality scores less than this threshold won't be recalibrated" help="-pQ,--preserve_qscores_less_than &amp;lt;preserve_qscores_less_than&amp;gt;"/>
490 <param name="smoothing" type="integer" value="1" label="smoothing" help="-sm,--smoothing &amp;lt;smoothing&amp;gt;"/>
491 <param name="max_quality_score" type="integer" value="50" label="Max quality score" help="-maxQ,--max_quality_score &amp;lt;max_quality_score&amp;gt;"/>
492 <param name="do_not_write_original_quals" type="boolean" checked="False" truevalue="--doNotWriteOriginalQuals" falsevalue="" label="Do Not Write Original Quality tag" help="-noOQs,--doNotWriteOriginalQuals"/>
493 </when>
494 </conditional>
495
496 <conditional name="analysis_param_type">
497 <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options">
498 <option value="basic" selected="True">Basic</option>
499 <option value="advanced">Advanced</option>
500 </param>
501 <when value="basic">
502 <!-- Do nothing here -->
503 </when>
504 <when value="advanced">
505
506 <param name="activeRegionIn" type="data" format="bed,gatk_interval,picard_interval_list,vcf" optional="true" label="activeRegionIn" help="--activeRegionIn / -AR Use this interval list file as the active regions to process"/>
507 <param name="activeRegionOut" type="boolean" checked="False" truevalue="" falsevalue="" label="activeRegionOut" help="--activeRegionOut / -ARO Output the active region to an interval list file"/>
508 <param name="alleles" type="data" format="vcf" optional="true" label="alleles" help="--alleles / -alleles The set of alleles at which to genotype when --genotyping_mode is GENOTYPE_GIVEN_ALLELES"/>
509 <param name="annotation" type="text" value="" optional="true" label="annotation" help="--annotation / -A One or more specific annotations to apply to variant calls default: ClippingRankSumTest"/>
510 <param name="comp" type="data" format="vcf" optional="true" label="comp" help="--comp / -comp comparison VCF file"/>
511 <param name="contamination_fraction_to_filter" type="float" value="0.05" optional="true" label="contamination_fraction_to_filter" help="--contamination_fraction_to_filter / -contamination Fraction of contamination in sequencing data (for all samples) to aggressively remove">
512 <validator type="in_range" message="value between 0.00 and 1.00" min="0" max="1"/>
513 </param>
514 <param name="dbsnp" type="data" format="vcf" optional="true" label="dbsnp" help="--dbsnp / -D dbSNP file"/>
515 <param name="debug" type="boolean" checked="False" truevalue="-debug" falsevalue="" label="debug" help="--debug / -debug If specified, print out very verbose debug information about each triggering active region"/>
516 <param name="downsampleRegion" type="integer" value="1000" optional="true" label="downsampleRegion" help="--downsampleRegion / -dr coverage, per-sample, to downsample each active region to"/>
517 <param name="excludeAnnotation" type="text" optional="true" label="excludeAnnotation" help="--excludeAnnotation / -XA One or more specific annotations to exclude"/>
518 <param name="genotyping_mode" type="select" optional="true" label="genotyping_mode" help="--genotyping_mode / -gt_mode Specifies how to determine the alternate alleles to use for genotyping">
519 <option value="DISCOVERY" selected="True">DISCOVERY</option>
520 <option value="GENOTYPE_GIVEN_ALLELES">GENOTYPE_GIVEN_ALLELES</option>
521 </param>
522 <param name="graphOutput" type="boolean" checked="False" truevalue="" falsevalue="" label="graphOutput" help="--graphOutput / -graph File to which debug assembly graph information should be written"/>
523 <param name="group" type="text" optional="true" label="group" help="--group / -G One or more classes/groups of annotations to apply to variant calls"/>
524 <param name="heterozygosity" type="float" value="0.0010" optional="true" label="heterozygosity" help="--heterozygosity / -hets Heterozygosity value used to compute prior likelihoods for any locus"/>
525 <param name="minPruning" type="integer" value="1" optional="true" label="minPruning" help="--minPruning / -minPruning The minimum allowed pruning factor in assembly graph. Paths with <= X supporting kmers are pruned from the graph">
526 <validator type="in_range" message="value between 0 and 127" min="0" max="127"/>
527 </param>
528 <param name="output_mode" type="select" optional="true" label="output_mode" help="--output_mode / -out_mode Specifies which type of calls we should output">
529 <option value="EMIT_VARIANTS_ONLY" selected="True">EMIT_VARIANTS_ONLY</option>
530 <option value="EMIT_ALL_CONFIDENT_SITES">EMIT_ALL_CONFIDENT_SITES</option>
531 <option value="EMIT_ALL_SITES">EMIT_ALL_SITES</option>
532 </param>
533 <param name="pair_hmm_implementation" type="select" optional="true" label="pair_hmm_implementation" help="--pair_hmm_implementation / -pairHMM The PairHMM implementation to use for genotype likelihood calculations">
534 <option value="EXACT">EXACT</option>
535 <option value="ORIGINAL">ORIGINAL</option>
536 <option value="CACHING">CACHING</option>
537 <option value="LOGLESS_CACHING" selected="True">LOGLESS_CACHING</option>
538 </param>
539 <param name="standard_min_confidence_threshold_for_calling" type="float" value="30.0" optional="true" label="standard_min_confidence_threshold_for_calling" help="--standard_min_confidence_threshold_for_calling / -stand_call_conf The minimum phred-scaled confidence threshold at which variants should be called"/>
540 <param name="standard_min_confidence_threshold_for_emitting" type="float" value="30.0" optional="true" label="standard_min_confidence_threshold_for_emitting" help="--standard_min_confidence_threshold_for_emitting / -stand_emit_conf The minimum phred-scaled confidence threshold at which variants should be emitted (and filtered with LowQual if less than the calling threshold)"/>
541 <param name="useAllelesTrigger" type="boolean" checked="False" truevalue="-allelesTrigger" falsevalue="" label="useAllelesTrigger" help="--useAllelesTrigger / -allelesTrigger If specified, use additional trigger on variants found in an external alleles file"/>
542 <param name="fullHaplotype" type="boolean" checked="False" truevalue="-fullHaplotype" falsevalue="" label="fullHaplotype" help="--fullHaplotype / -fullHaplotype If specified, output the full haplotype sequence instead of converting to individual variants w.r.t. the reference"/>
543 <param name="gcpHMM" type="integer" value="10" optional="true" label="gcpHMM" help="--gcpHMM / -gcpHMM Flat gap continuation penalty for use in the Pair HMM"/>
544 <param name="genotypeFullActiveRegion" type="boolean" checked="False" truevalue="-genotypeFullActiveRegion" falsevalue="" label="genotypeFullActiveRegion" help="--genotypeFullActiveRegion / -genotypeFullActiveRegion If specified, alternate alleles are considered to be the full active region for the purposes of genotyping"/>
545 <param name="max_alternate_alleles" type="integer" value="6" optional="true" label="max_alternate_alleles" help="--max_alternate_alleles / -maxAltAlleles Maximum number of alternate alleles to genotype"/>
546 <param name="p_nonref_model" type="select" optional="true" label="p_nonref_model" help="--p_nonref_model / -pnrm Non-reference probability calculation model to employ">
547 <option value="EXACT_INDEPENDENT" selected="True">EXACT_INDEPENDENT experimental implementation - for testing only</option>
548 <option value="EXACT_REFERENCE">EXACT_REFERENCE reference implementation of multi-allelic EXACT model. Extremely slow for many alternate alleles</option>
549 <option value="EXACT_ORIGINAL">EXACT_ORIGINAL original biallelic exact model, for testing only</option>
550 <option value="EXACT_GENERAL_PLOIDY">implementation that supports any sample ploidy</option>
551 </param>
552
553 </when>
554 <plotypes are evaluated using an affine gap penalty Pair HMM./conditional>
555 </inputs>
556 <outputs>
557 <data format="vcf" name="output_vcf" label="${tool.name} on ${on_string} (VCF)" />
558 <data format="vcf" name="graph_out" label="${tool.name} on ${on_string} graph" >
559 <filter>analysis_param_type['graphOutput'] == True</filter>
560 </data>
561 <data format="vcf" name="active_region_out" label="${tool.name} on ${on_string} activeRegion" >
562 <filter>analysis_param_type['activeRegionOut'] == True</filter>
563 </data>
564 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
565 </outputs>
566 <tests>
567 <test>
568 <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" />
569 <param name="reference_source_selector" value="history" />
570 <param name="ref_file" value="phiX.fasta" ftype="fasta" />
571 <param name="input_bam" value="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" />
572 <param name="gatk_param_type_selector" value="basic" />
573 <param name="analysis_param_type_selector" value="basic" />
574 <output name="output_bam" file="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.bam" ftype="bam" lines_diff="4" />
575 <output name="output_log" file="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.log.contains" compare="contains" />
576 </test>
577 </tests>
578 <help>
579 **What it does**
580
581 HaplotypeCaller
582 Call SNPs and indels simultaneously via local de-novo assembly of haplotypes in an active region.
583 Haplotypes are evaluated using an affine gap penalty Pair HMM.
584
585 For more information on using read based compression in the GATK, see this `tool specific page &lt;http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_haplotypecaller_HaplotypeCaller.html&gt;`_.
586
587 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;`_.
588
589 If you encounter errors, please view the `GATK FAQ &lt;http://www.broadinstitute.org/gatk/guide/topic?name=faqs&gt;`_.
590
591 ------
592
593 **Inputs**
594
595 GenomeAnalysisTK: PrintReads accepts an aligned BAM files.
596
597
598 **Outputs**
599
600 The output is a VCF file with raw, unrecalibrated SNP and indel calls.
601
602
603 Go `here &lt;http://www.broadinstitute.org/gatk/guide/topic?name=intro&gt;`_ for details on GATK file formats.
604
605 -------
606
607 **Settings**::
608
609 default_read_group If a read has no read group then default to the provided String.
610 default_platform If a read has no platform then default to the provided String. Valid options are illumina, 454, and solid.
611 force_read_group If provided, the read group ID of EVERY read will be forced to be the provided String. This is useful to collapse all data into a single read group.
612 force_platform If provided, the platform of EVERY read will be forced to be the provided String. Valid options are illumina, 454, and solid.
613 window_size_nqs The window size used by MinimumNQSCovariate for its calculation
614 homopolymer_nback The number of previous bases to look at in HomopolymerCovariate
615 exception_if_no_tile If provided, TileCovariate will throw an exception when no tile can be found. The default behavior is to use tile = -1
616 solid_recal_mode How should we recalibrate solid bases in whichthe reference was inserted? Options = DO_NOTHING, SET_Q_ZERO, SET_Q_ZERO_BASE_N, or REMOVE_REF_BIAS (DO_NOTHING|SET_Q_ZERO|SET_Q_ZERO_BASE_N|REMOVE_REF_BIAS)
617 solid_nocall_strategy Defines the behavior of the recalibrator when it encounters no calls in the color space. Options = THROW_EXCEPTION, LEAVE_READ_UNRECALIBRATED, or PURGE_READ (THROW_EXCEPTION|LEAVE_READ_UNRECALIBRATED|PURGE_READ)
618 recal_file Filename for the input covariates table recalibration .csv file
619 out The output BAM file
620 bam_compression Compression level to use for writing BAM files
621 disable_bam_indexing Turn off on-the-fly creation of indices for output BAM files.
622 simplifyBAM If provided, output BAM files will be simplified to include just key reads for downstream variation discovery analyses (removing duplicates, PF-, non-primary reads), as well stripping all extended tags from the kept reads except the read group identifier
623 preserve_qscores_less_than Bases with quality scores less than this threshold won't be recalibrated, default=5. In general it's unsafe to change qualities scores below &lt; 5, since base callers use these values to indicate random or bad bases
624 smoothing Number of imaginary counts to add to each bin bin order to smooth out bins with few data points, default=1
625 max_quality_score The integer value at which to cap the quality scores, default=50
626 doNotWriteOriginalQuals If true, we will not write the original quality (OQ) tag for each read
627
628 ------
629
630 **Citation**
631
632 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;`_
633
634 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;`_
635
636 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;`_
637
638 </help>
639 </tool>