|
4
|
1 <tool id="gatk2_haplotype_caller" name="Haplotype Caller" version="@VERSION@.0">
|
|
0
|
2 <description>Call SNPs and indels simultaneously via local de-novo assembly of haplotypes in an active region</description>
|
|
|
3 <expand macro="requirements" />
|
|
|
4 <macros>
|
|
|
5 <import>gatk2_macros.xml</import>
|
|
|
6 </macros>
|
|
|
7 <command interpreter="python">
|
|
|
8 gatk2_wrapper.py
|
|
|
9 --stdout "${output_log}"
|
|
2
|
10 #for $i, $input_bam in enumerate( $reference_source.input_bams ):
|
|
|
11 -d "-I" "${input_bam.input_bam}" "${input_bam.input_bam.ext}" "gatk_input_${i}"
|
|
|
12 #if str( $input_bam.input_bam.metadata.bam_index ) != "None":
|
|
|
13 -d "" "${input_bam.input_bam.metadata.bam_index}" "bam_index" "gatk_input_${i}" ##hardcode galaxy ext type as bam_index
|
|
|
14 #end if
|
|
|
15 #end for
|
|
0
|
16 -p '
|
|
|
17 @JAR_PATH@
|
|
|
18 -T "HaplotypeCaller"
|
|
|
19 -o "${output_vcf}"
|
|
|
20
|
|
|
21 \$GATK2_SITE_OPTIONS
|
|
|
22
|
|
2
|
23 --num_cpu_threads_per_data_thread \${GALAXY_SLOTS:-4}
|
|
0
|
24
|
|
|
25 #if $reference_source.reference_source_selector != "history":
|
|
|
26 -R "${reference_source.ref_file.fields.path}"
|
|
|
27 #end if
|
|
|
28 #if str($input_recal) != 'None':
|
|
|
29 --BQSR "${input_recal}"
|
|
|
30 #end if
|
|
|
31 '
|
|
2
|
32 @DBSNP_OPTIONS@
|
|
0
|
33 #include source=$standard_gatk_options#
|
|
|
34
|
|
|
35 ##start analysis specific options
|
|
|
36 #if $analysis_param_type.analysis_param_type_selector == "advanced":
|
|
|
37 -p '
|
|
|
38 #if $analysis_param_type.heterozygosity.__str__.strip() != '':
|
|
|
39 --heterozygosity $analysis_param_type.heterozygosity
|
|
|
40 #end if
|
|
|
41 --genotyping_mode "${analysis_param_type.genotyping_mode_type.genotyping_mode}"
|
|
|
42 #if str( $analysis_param_type.genotyping_mode_type.genotyping_mode ) == 'GENOTYPE_GIVEN_ALLELES':
|
|
|
43 --alleles "${analysis_param_type.genotyping_mode_type.input_alleles_rod}"
|
|
|
44 #end if
|
|
2
|
45 #if not $analysis_param_type.emitRefConfidence is None:
|
|
|
46 --emitRefConfidence $analysis_param_type.emitRefConfidence
|
|
0
|
47 #end if
|
|
|
48
|
|
|
49 ## files
|
|
|
50 #if str($analysis_param_type.activeRegionIn) != 'None':
|
|
|
51 --activeRegionIn "$analysis_param_type.activeRegionIn"
|
|
|
52 #end if
|
|
|
53 #if str($analysis_param_type.comp) != 'None':
|
|
|
54 --comp "$analysis_param_type.comp"
|
|
|
55 #end if
|
|
|
56 ##
|
|
|
57 #if str( $analysis_param_type.annotation ) != "None":
|
|
|
58 #for $annotation in str( $analysis_param_type.annotation.fields.gatk_value ).split( ','):
|
|
|
59 --annotation "${annotation}"
|
|
|
60 #end for
|
|
|
61 #end if
|
|
|
62 #for $additional_annotation in $analysis_param_type.additional_annotations:
|
|
|
63 --annotation "${additional_annotation.additional_annotation_name}"
|
|
|
64 #end for
|
|
|
65 #if str( $analysis_param_type.group ) != "None":
|
|
|
66 #for $group in str( $analysis_param_type.group ).split( ','):
|
|
|
67 --group "${group}"
|
|
|
68 #end for
|
|
|
69 #end if
|
|
|
70 #if str( $analysis_param_type.exclude_annotations ) != "None":
|
|
|
71 #for $annotation in str( $analysis_param_type.exclude_annotations.fields.gatk_value ).split( ','):
|
|
|
72 --excludeAnnotation "${annotation}"
|
|
|
73 #end for
|
|
|
74 #end if
|
|
|
75
|
|
|
76 ## value setings
|
|
|
77 #if $analysis_param_type.contamination_fraction_to_filter.__str__.strip() != '':
|
|
|
78 --contamination_fraction_to_filter $analysis_param_type.contamination_fraction_to_filter
|
|
|
79 #end if
|
|
|
80 #if $analysis_param_type.minPruning.__str__.strip() != '':
|
|
|
81 --minPruning $analysis_param_type.minPruning
|
|
|
82 #end if
|
|
|
83 #if $analysis_param_type.standard_min_confidence_threshold_for_calling.__str__.strip() != '':
|
|
|
84 --standard_min_confidence_threshold_for_calling $analysis_param_type.standard_min_confidence_threshold_for_calling
|
|
|
85 #end if
|
|
|
86 #if $analysis_param_type.standard_min_confidence_threshold_for_emitting.__str__.strip() != '':
|
|
|
87 --standard_min_confidence_threshold_for_emitting $analysis_param_type.standard_min_confidence_threshold_for_emitting
|
|
|
88 #end if
|
|
|
89 #if $analysis_param_type.gcpHMM.__str__.strip() != '':
|
|
|
90 --gcpHMM $analysis_param_type.gcpHMM
|
|
|
91 #end if
|
|
|
92 #if $analysis_param_type.max_alternate_alleles.__str__.strip() != '':
|
|
|
93 --max_alternate_alleles $analysis_param_type.max_alternate_alleles
|
|
|
94 #end if
|
|
|
95 ## mode selections
|
|
2
|
96
|
|
0
|
97 #if $analysis_param_type.pair_hmm_implementation.__str__ != "None" and len($analysis_param_type.pair_hmm_implementation.__str__) > 0:
|
|
|
98 --pair_hmm_implementation $analysis_param_type.pair_hmm_implementation
|
|
|
99 #end if
|
|
|
100 ## optional outputs
|
|
|
101 #if $analysis_param_type.activeRegionOut:
|
|
|
102 --activeRegionOut $active_region_out
|
|
|
103 #end if
|
|
|
104 #if $analysis_param_type.graphOutput:
|
|
|
105 --graphOutput $graph_out
|
|
|
106 #end if
|
|
|
107 ## flags
|
|
|
108 $analysis_param_type.useAllelesTrigger
|
|
|
109 $analysis_param_type.fullHaplotype
|
|
|
110 $analysis_param_type.genotypeFullActiveRegion
|
|
|
111 $analysis_param_type.debug
|
|
|
112 '
|
|
|
113 #end if
|
|
|
114 </command>
|
|
|
115 <inputs>
|
|
2
|
116 <param name="input_recal" type="data" format="gatk_report" optional="true" label="Covariates table recalibration file" help="The input covariates table file which enables on-the-fly base quality score recalibration. Enables on-the-fly recalibrate of base qualities. The covariates tables are produced by the BaseQualityScoreRecalibrator tool. Please be aware that one should only run recalibration with the covariates file created on the same input bam(s) (-BQSR,--BQSR &lt;recal_file&gt;)" />
|
|
0
|
117 <conditional name="reference_source">
|
|
|
118 <expand macro="reference_source_selector_param" />
|
|
|
119 <when value="cached">
|
|
2
|
120 <repeat name="input_bams" title="BAM file" min="1" help="-I,--input_file &lt;input_file&gt;">
|
|
|
121 <param name="input_bam" type="data" format="bam" label="BAM file">
|
|
|
122 <validator type="unspecified_build" />
|
|
|
123 <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 -->
|
|
|
124 </param>
|
|
|
125 </repeat>
|
|
0
|
126 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;" >
|
|
|
127 <options from_data_table="gatk2_picard_indexes">
|
|
2
|
128 <!-- <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/> does not yet work in a repeat...-->
|
|
0
|
129 </options>
|
|
|
130 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
|
|
|
131 </param>
|
|
|
132 </when>
|
|
|
133 <when value="history">
|
|
2
|
134 <repeat name="input_bams" title="BAM file" min="1" help="-I,--input_file &lt;input_file&gt;">
|
|
|
135 <param name="input_bam" type="data" format="bam" label="BAM file" />
|
|
|
136 </repeat>
|
|
|
137 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" />
|
|
0
|
138 </when>
|
|
|
139 </conditional>
|
|
2
|
140 <expand macro="dbsnp_param" />
|
|
0
|
141
|
|
|
142 <expand macro="gatk_param_type_conditional" />
|
|
|
143
|
|
|
144 <conditional name="analysis_param_type">
|
|
|
145 <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options">
|
|
|
146 <option value="basic" selected="True">Basic</option>
|
|
|
147 <option value="advanced">Advanced</option>
|
|
|
148 </param>
|
|
|
149 <when value="basic">
|
|
|
150 <!-- Do nothing here -->
|
|
|
151 </when>
|
|
|
152 <when value="advanced">
|
|
|
153
|
|
|
154 <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"/>
|
|
|
155 <param name="activeRegionOut" type="boolean" checked="False" truevalue="" falsevalue="" label="activeRegionOut" help="--activeRegionOut / -ARO Output the active region to an interval list file"/>
|
|
|
156
|
|
|
157 <param name="annotation" type="select" multiple="True" display="checkboxes" label="Annotation Types" help="-A,--annotation &lt;annotation&gt;">
|
|
|
158 <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs -->
|
|
|
159 <options from_data_table="gatk2_annotations">
|
|
|
160 <filter type="multiple_splitter" column="tools_valid_for" separator=","/>
|
|
4
|
161 <filter type="static_value" value="HaplotypeCaller" column="tools_valid_for"/>
|
|
0
|
162 </options>
|
|
|
163 </param>
|
|
|
164 <repeat name="additional_annotations" title="Additional annotation" help="-A,--annotation &lt;annotation&gt;">
|
|
|
165 <param name="additional_annotation_name" type="text" value="" label="Annotation name" />
|
|
|
166 </repeat>
|
|
|
167 <!--
|
|
|
168 <conditional name="snpEff_rod_bind_type">
|
|
|
169 <param name="snpEff_rod_bind_type_selector" type="select" label="Provide a snpEff reference-ordered data file">
|
|
|
170 <option value="set_snpEff">Set snpEff</option>
|
|
|
171 <option value="exclude_snpEff" selected="True">Don't set snpEff</option>
|
|
|
172 </param>
|
|
|
173 <when value="exclude_snpEff">
|
|
|
174 </when>
|
|
|
175 <when value="set_snpEff">
|
|
|
176 <param name="snpEff_input_rod" type="data" format="vcf" label="ROD file" />
|
|
|
177 <param name="snpEff_rod_name" type="hidden" value="snpEff" label="ROD Name"/>
|
|
|
178 </when>
|
|
|
179 </conditional>
|
|
|
180 -->
|
|
|
181 <param name="group" type="select" multiple="True" display="checkboxes" label="Annotation Interfaces/Groups" help="-G,--group &lt;group&gt;">
|
|
|
182 <option value="RodRequiringAnnotation">RodRequiringAnnotation</option>
|
|
|
183 <option value="Standard">Standard</option>
|
|
|
184 <option value="Experimental">Experimental</option>
|
|
|
185 <option value="WorkInProgress">WorkInProgress</option>
|
|
|
186 <option value="RankSumTest">RankSumTest</option>
|
|
|
187 <!-- <option value="none">none</option> -->
|
|
|
188 </param>
|
|
|
189 <!-- <param name="family_string" type="text" value="" label="Family String"/> -->
|
|
|
190 <param name="exclude_annotations" type="select" multiple="True" display="checkboxes" label="Annotations to exclude" help="-XA,--excludeAnnotation &lt;excludeAnnotation&gt;" >
|
|
|
191 <!-- load the available annotations from an external configuration file, since additional ones can be added to local installs -->
|
|
|
192 <options from_data_table="gatk2_annotations">
|
|
|
193 <filter type="multiple_splitter" column="tools_valid_for" separator=","/>
|
|
4
|
194 <filter type="static_value" value="HaplotypeCaller" column="tools_valid_for"/>
|
|
0
|
195 </options>
|
|
|
196 </param>
|
|
|
197
|
|
|
198 <param name="comp" type="data" format="vcf" optional="true" label="comp" help="--comp / -comp comparison VCF file"/>
|
|
|
199 <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">
|
|
|
200 <validator type="in_range" message="value between 0.00 and 1.00" min="0" max="1"/>
|
|
|
201 </param>
|
|
|
202 <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"/>
|
|
|
203
|
|
|
204 <conditional name="genotyping_mode_type">
|
|
|
205 <param name="genotyping_mode" type="select" label="How to determine the alternate allele to use for genotyping" help="-gt_mode,--genotyping_mode &lt;genotyping_mode&gt;">
|
|
|
206 <option value="DISCOVERY" selected="True">DISCOVERY</option>
|
|
|
207 <option value="GENOTYPE_GIVEN_ALLELES">GENOTYPE_GIVEN_ALLELES</option>
|
|
|
208 </param>
|
|
|
209 <when value="DISCOVERY">
|
|
|
210 <!-- Do nothing here -->
|
|
|
211 </when>
|
|
|
212 <when value="GENOTYPE_GIVEN_ALLELES">
|
|
|
213 <param name="input_alleles_rod" type="data" format="vcf" label="Alleles ROD file" help="-alleles,--alleles &lt;alleles&gt;" />
|
|
|
214 </when>
|
|
|
215 </conditional>
|
|
|
216
|
|
|
217
|
|
|
218 <param name="graphOutput" type="boolean" checked="False" truevalue="" falsevalue="" label="graphOutput" help="--graphOutput / -graph File to which debug assembly graph information should be written"/>
|
|
|
219 <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"/>
|
|
|
220 <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">
|
|
|
221 <validator type="in_range" message="value between 0 and 127" min="0" max="127"/>
|
|
|
222 </param>
|
|
2
|
223 <!-- http://www.broadinstitute.org/gatk/guide/article?id=2940 -->
|
|
|
224 <param name="emitRefConfidence" type="select" optional="true" label="Output confidence estimates" help="Emitting a per-bp or summarized confidence estimate for a site being strictly homozygous-reference (--emitRefConfidence)">
|
|
|
225 <option value="NONE" selected="True">don't emit anything</option>
|
|
|
226 <option value="BP_RESOLUTION">BP_RESOLUTION (emit detailed information for each BP)</option>
|
|
|
227 <option value="GVCF">GVCF (emit a block summarized version of the BP_RESOLUTION data)</option>
|
|
0
|
228 </param>
|
|
|
229 <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">
|
|
|
230 <option value="EXACT">EXACT</option>
|
|
|
231 <option value="ORIGINAL">ORIGINAL</option>
|
|
|
232 <option value="CACHING">CACHING</option>
|
|
|
233 <option value="LOGLESS_CACHING" selected="True">LOGLESS_CACHING</option>
|
|
|
234 </param>
|
|
|
235 <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"/>
|
|
|
236 <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)"/>
|
|
|
237 <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"/>
|
|
|
238 <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"/>
|
|
|
239 <param name="gcpHMM" type="integer" value="10" optional="true" label="gcpHMM" help="--gcpHMM / -gcpHMM Flat gap continuation penalty for use in the Pair HMM"/>
|
|
|
240 <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"/>
|
|
|
241 <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"/>
|
|
|
242 </when>
|
|
|
243 </conditional>
|
|
|
244 </inputs>
|
|
|
245 <outputs>
|
|
|
246 <data format="vcf" name="output_vcf" label="${tool.name} on ${on_string} (VCF)" />
|
|
|
247 <data format="vcf" name="graph_out" label="${tool.name} on ${on_string} graph" >
|
|
|
248 <filter>analysis_param_type['analysis_param_type_selector'] == "advanced" and analysis_param_type['graphOutput'] == True</filter>
|
|
|
249 </data>
|
|
|
250 <data format="vcf" name="active_region_out" label="${tool.name} on ${on_string} activeRegion" >
|
|
|
251 <filter>analysis_param_type['analysis_param_type_selector'] == "advanced" and analysis_param_type['activeRegionOut'] == True</filter>
|
|
|
252 </data>
|
|
|
253 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
|
|
|
254 </outputs>
|
|
|
255 <tests>
|
|
|
256 <test>
|
|
|
257 <param name="input_recal" value="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" ftype="csv" />
|
|
|
258 <param name="reference_source_selector" value="history" />
|
|
|
259 <param name="ref_file" value="phiX.fasta" ftype="fasta" />
|
|
|
260 <param name="input_bam" value="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" />
|
|
|
261 <param name="gatk_param_type_selector" value="basic" />
|
|
|
262 <param name="analysis_param_type_selector" value="basic" />
|
|
|
263 <output name="output_bam" file="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.bam" ftype="bam" lines_diff="4" />
|
|
|
264 <output name="output_log" file="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.log.contains" compare="contains" />
|
|
|
265 </test>
|
|
|
266 </tests>
|
|
|
267 <help>
|
|
|
268 **What it does**
|
|
|
269
|
|
|
270 **HaplotypeCaller**
|
|
|
271 calls SNPs and indels simultaneously via local de-novo assembly of haplotypes in an active region.
|
|
|
272 Haplotypes are evaluated using an affine gap penalty Pair HMM.
|
|
|
273
|
|
|
274 For more information on using read based compression in the GATK, see this `tool specific page <http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_haplotypecaller_HaplotypeCaller.html>`_.
|
|
|
275
|
|
|
276 To learn about best practices for variant detection using GATK, see this `overview <http://www.broadinstitute.org/gatk/guide/topic?name=best-practices>`_.
|
|
|
277
|
|
|
278 If you encounter errors, please view the `GATK FAQ <http://www.broadinstitute.org/gatk/guide/topic?name=faqs>`_.
|
|
|
279
|
|
|
280 ------
|
|
|
281
|
|
|
282 **Inputs**
|
|
|
283
|
|
|
284 GenomeAnalysisTK: PrintReads accepts aligned BAM files.
|
|
|
285
|
|
|
286
|
|
|
287 **Outputs**
|
|
|
288
|
|
|
289 The output is a VCF file with raw, unrecalibrated SNP and indel calls.
|
|
|
290
|
|
|
291
|
|
|
292 Go `here <http://www.broadinstitute.org/gatk/guide/topic?name=intro>`_ for details on GATK file formats.
|
|
|
293
|
|
|
294 -------
|
|
|
295
|
|
|
296 **Settings**::
|
|
|
297
|
|
|
298 activeRegionIn Use this interval list file as the active regions to process
|
|
|
299 activeRegionOut Output the active region to this interval list file
|
|
|
300 alleles The set of alleles at which to genotype when --genotyping_mode is GENOTYPE_GIVEN_ALLELES
|
|
|
301 annotation One or more specific annotations to apply to variant calls
|
|
|
302 comp comparison VCF file
|
|
|
303 contamination Fraction of contamination in sequencing data (for all samples) to aggressively remove
|
|
|
304 dbsnp dbSNP file
|
|
|
305 debug If specified, print out very verbose debug information about each triggering active region
|
|
|
306 excludeAnnotation One or more specific annotations to exclude
|
|
|
307 genotyping_mode Specifies how to determine the alternate alleles to use for genotyping
|
|
|
308 graphOutput File to which debug assembly graph information should be written
|
|
|
309 group One or more classes/groups of annotations to apply to variant calls
|
|
|
310 heterozygosity Heterozygosity value used to compute prior likelihoods for any locus
|
|
|
311 minPruning The minimum allowed pruning factor in assembly graph. Paths with less than or equal supporting kmers are pruned from the graph
|
|
|
312 pair_hmm_implementation The PairHMM implementation to use for genotype likelihood calculations
|
|
|
313 stand_call_conf The minimum phred-scaled confidence threshold at which variants should be called
|
|
|
314 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)
|
|
|
315 useAllelesTrigger If specified, use additional trigger on variants found in an external alleles file
|
|
|
316 fullHaplotype If specified, output the full haplotype sequence instead of converting to individual variants w.r.t. the reference
|
|
|
317 gcpHMM Flat gap continuation penalty for use in the Pair HMM
|
|
|
318 genotypeFullActiveRegion If specified, alternate alleles are considered to be the full active region for the purposes of genotyping
|
|
|
319 max_alternate_alleles Maximum number of alternate alleles to genotype
|
|
|
320
|
|
|
321 @CITATION_SECTION@
|
|
|
322 </help>
|
|
4
|
323 <expand macro="citations" />
|
|
0
|
324 </tool>
|
