Mercurial > repos > jjohnson > gatk2
annotate variant_eval.xml @ 35:a29c7d7d962d draft
Fix default read filter screening
| author | Jim Johnson <jj@umn.edu> |
|---|---|
| date | Fri, 01 Mar 2013 07:52:02 -0600 |
| parents | b99c25b0ad4d |
| children | 14d47237bb0a |
| rev | line source |
|---|---|
| 34 | 1 <tool id="gatk2_variant_eval" name="Eval Variants" version="0.0.7"> |
| 0 | 2 <description></description> |
| 3 <requirements> | |
|
18
7533db8dfb5b
Update tool_dependencies to GATK v 2.3
Jim Johnson <jj@umn.edu>
parents:
0
diff
changeset
|
4 <requirement type="package" version="2.3">gatk</requirement> |
| 0 | 5 </requirements> |
| 6 <command interpreter="python">gatk2_wrapper.py | |
| 7 #from binascii import hexlify | |
| 8 --max_jvm_heap_fraction "1" | |
| 9 --stdout "${output_log}" | |
| 10 #for $var_count, $variant in enumerate( $reference_source.variants ): | |
| 11 -d "--eval:input_${var_count},%(file_type)s" "${variant.input_variant}" "${variant.input_variant.ext}" "input_variants_${var_count}" | |
| 12 #end for | |
| 13 -p 'java | |
| 14 -jar "\$GATK2_PATH/GenomeAnalysisTK.jar" | |
| 15 -T "VariantEval" | |
| 16 --out "${output_report}" | |
| 17 \$GATK2_SITE_OPTIONS | |
|
28
6ef8eb568700
Move GATK tool_dependency to a repository dependency, use 2 env vars: GATK2_NUM_THREADS and GATK2_NUM_CPU_THREADS for site threading values
Jim Johnson <jj@umn.edu>
parents:
18
diff
changeset
|
18 \$GATK2_NUM_THREADS |
| 0 | 19 ##--num_threads 4 ##hard coded, for now |
| 20 ##-et "NO_ET" -K "\$GATK2_BASE/gatk2_key_file" ##ET no phone home | |
| 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 ' | |
| 26 | |
| 27 #for $rod_binding in $comp_rod_bind: | |
| 28 -d "--comp:${rod_binding.comp_rod_name},%(file_type)s" "${rod_binding.comp_input_rod}" "${rod_binding.comp_input_rod.ext}" "input_comp_${rod_binding.comp_rod_name}" | |
| 29 #if str( $rod_binding.comp_known_names ): | |
| 30 -p '--known_names "${rod_binding.comp_rod_name}"' | |
| 31 #end if | |
| 32 #end for | |
| 33 | |
| 34 #if str( $dbsnp_rod_bind_type.dbsnp_rod_bind_type_selector ) == 'set_dbsnp': | |
| 35 -d "--dbsnp:${dbsnp_rod_bind_type.dbsnp_rod_name},%(file_type)s" "${dbsnp_rod_bind_type.dbsnp_input_rod}" "${dbsnp_rod_bind_type.dbsnp_input_rod.ext}" "input_dbsnp_${dbsnp_rod_bind_type.dbsnp_rod_name}" | |
| 36 #if str( $dbsnp_rod_bind_type.dbsnp_known_names ): | |
| 37 -p '--known_names "${dbsnp_rod_bind_type.dbsnp_rod_name}"' | |
| 38 #end if | |
| 39 #end if | |
| 40 | |
| 41 ##start standard gatk options | |
| 42 #if $gatk_param_type.gatk_param_type_selector == "advanced": | |
| 43 #for $pedigree in $gatk_param_type.pedigree: | |
| 44 -p '--pedigree "${pedigree.pedigree_file}"' | |
| 45 #end for | |
| 46 #for $pedigree_string in $gatk_param_type.pedigree_string_repeat: | |
| 47 -p '--pedigreeString "${pedigree_string.pedigree_string}"' | |
| 48 #end for | |
| 49 -p '--pedigreeValidationType "${gatk_param_type.pedigree_validation_type}"' | |
| 35 | 50 #set default_read_filters = ['DuplicateRead','FailsVendorQualityCheck','NotPrimaryAlignment','MalformedRead','UnmappedRead'] |
| 0 | 51 #for $read_filter in $gatk_param_type.read_filter: |
|
33
76f8ca47b810
Omit --read_filter param included by default
Jim Johnson <jj@umn.edu>
parents:
28
diff
changeset
|
52 -p ' |
| 35 | 53 #if $read_filter.read_filter_type.read_filter_type_selector not in $default_read_filters: |
|
33
76f8ca47b810
Omit --read_filter param included by default
Jim Johnson <jj@umn.edu>
parents:
28
diff
changeset
|
54 --read_filter "${read_filter.read_filter_type.read_filter_type_selector}" |
|
76f8ca47b810
Omit --read_filter param included by default
Jim Johnson <jj@umn.edu>
parents:
28
diff
changeset
|
55 #end_if |
| 0 | 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} | |
|
18
7533db8dfb5b
Update tool_dependencies to GATK v 2.3
Jim Johnson <jj@umn.edu>
parents:
0
diff
changeset
|
89 ${gatk_param_type.fix_misencoded_quality_scores} |
| 0 | 90 ${gatk_param_type.non_deterministic_random_seed} |
| 91 ' | |
| 92 #for $rg_black_list_count, $rg_black_list in enumerate( $gatk_param_type.read_group_black_list_repeat ): | |
| 93 #if $rg_black_list.read_group_black_list_type.read_group_black_list_type_selector == "file": | |
| 94 -d "--read_group_black_list" "${rg_black_list.read_group_black_list_type.read_group_black_list}" "txt" "input_read_group_black_list_${rg_black_list_count}" | |
| 95 #else | |
| 96 -p '--read_group_black_list "${rg_black_list.read_group_black_list_type.read_group_black_list}"' | |
| 97 #end if | |
| 98 #end for | |
| 99 #end if | |
| 100 | |
| 101 #if $reference_source.reference_source_selector == "history": | |
| 102 -d "-R" "${reference_source.ref_file}" "${reference_source.ref_file.ext}" "gatk_input" | |
| 103 #end if | |
| 104 ##end standard gatk options | |
| 105 | |
| 106 | |
| 107 ##start analysis specific options | |
| 108 #if $analysis_param_type.analysis_param_type_selector == "advanced": | |
| 109 #for $stratification in $analysis_param_type.stratifications: | |
| 110 #set $select_string = "--select_exps '%s' --select_names '%s'" % ( str( $stratification.select_exps ), str( $stratification.select_name ) ) | |
| 111 -o '${ hexlify( $select_string ) }' | |
| 112 #end for | |
| 113 -p ' | |
| 114 | |
| 115 #for $sample in $analysis_param_type.samples: | |
| 116 --sample "${sample.sample}" | |
| 117 #end for | |
| 118 | |
| 119 #if str( $analysis_param_type.stratification_modules ) != "None": | |
| 120 #for $stratification_module in str( $analysis_param_type.stratification_modules).split( ',' ): | |
| 121 --stratificationModule "${stratification_module}" | |
| 122 #end for | |
| 123 #end if | |
| 124 | |
| 125 ${analysis_param_type.do_not_use_all_standard_stratifications} | |
| 126 | |
| 127 #for $variant_type in $analysis_param_type.only_variants_of_type: | |
| 128 --onlyVariantsOfType "${variant_type.variant_type}" | |
| 129 #end for | |
| 130 | |
| 131 #if str( $analysis_param_type.eval_modules ) != "None": | |
| 132 #for $eval_module in str( $analysis_param_type.eval_modules).split( ',' ): | |
| 133 --evalModule "${eval_module}" | |
| 134 #end for | |
| 135 #end if | |
| 136 | |
| 137 ${analysis_param_type.do_not_use_all_standard_modules} | |
| 138 | |
| 139 #if str( $analysis_param_type.num_samples ) != "0": | |
| 140 --numSamples "${analysis_param_type.num_samples}" | |
| 141 #end if | |
| 142 | |
| 143 --minPhaseQuality "${analysis_param_type.min_phase_quality}" | |
| 144 | |
| 145 #if str( $analysis_param_type.family ): | |
| 146 --family_structure "${analysis_param_type.family}" | |
| 147 #end if | |
| 148 | |
| 149 --mendelianViolationQualThreshold "${analysis_param_type.mendelian_violation_qual_threshold}" | |
| 150 | |
| 151 #if str( $analysis_param_type.ancestral_alignments ) != "None": | |
| 152 --ancestralAlignments "${analysis_param_type.ancestral_alignments}" | |
| 153 #end if | |
| 154 ' | |
| 155 #if str( $analysis_param_type.known_cnvs ) != "None": | |
| 156 -d "--knownCNVs" "${analysis_param_type.known_cnvs}" "${analysis_param_type.known_cnvs.ext}" "input_known_cnvs" | |
| 157 #end if | |
| 158 | |
| 159 #if str( $analysis_param_type.strat_intervals ) != "None": | |
| 160 -d "--stratIntervals" "${analysis_param_type.strat_intervals}" "${analysis_param_type.strat_intervals.ext}" "input_strat_intervals" | |
| 161 #end if | |
| 162 #end if | |
| 163 </command> | |
| 164 <inputs> | |
| 165 | |
| 166 <conditional name="reference_source"> | |
| 167 <param name="reference_source_selector" type="select" label="Choose the source for the reference list"> | |
| 168 <option value="cached">Locally cached</option> | |
| 169 <option value="history">History</option> | |
| 170 </param> | |
| 171 <when value="cached"> | |
| 172 <repeat name="variants" title="Variant" min="1" help="-eval,--eval &lt;eval&gt;"> | |
| 173 <param name="input_variant" type="data" format="vcf" label="Input variant file" /> | |
| 174 </repeat> | |
| 175 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;"> | |
| 176 <options from_data_table="gatk2_picard_indexes"> | |
| 177 <!-- <filter type="data_meta" key="dbkey" ref="input_variant" column="dbkey"/> --> | |
| 178 </options> | |
| 179 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/> | |
| 180 </param> | |
| 181 </when> | |
| 182 <when value="history"> <!-- FIX ME!!!! --> | |
| 183 <repeat name="variants" title="Variant" min="1" help="-eval,--eval &lt;eval&gt;"> | |
| 184 <param name="input_variant" type="data" format="vcf" label="Input variant file" /> | |
| 185 </repeat> | |
| 186 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;" /> | |
| 187 </when> | |
| 188 </conditional> | |
| 189 | |
| 190 <repeat name="comp_rod_bind" title="Binding for reference-ordered comparison data" help="-comp,--comp &lt;comp&gt;"> | |
| 191 <param name="comp_input_rod" type="data" format="vcf" label="Comparison ROD file" /> | |
| 192 <param name="comp_rod_name" type="text" value="Unnamed" label="Comparison ROD Name"/> | |
| 193 <param name="comp_known_names" type="boolean" truevalue="--known_names" falsevalue="" label="Use Comparison ROD as known_names" help="-knownName,--known_names &lt;known_names&gt;"/> | |
| 194 </repeat> | |
| 195 | |
| 196 <conditional name="dbsnp_rod_bind_type"> | |
| 197 <param name="dbsnp_rod_bind_type_selector" type="select" label="Provide a dbSNP reference-ordered data file" help="-D,--dbsnp &lt;dbsnp&gt;"> | |
| 198 <option value="set_dbsnp" selected="True">Set dbSNP</option> | |
| 199 <option value="exclude_dbsnp">Don't set dbSNP</option> | |
| 200 </param> | |
| 201 <when value="exclude_dbsnp"> | |
| 202 <!-- Do nothing here --> | |
| 203 </when> | |
| 204 <when value="set_dbsnp"> | |
| 205 <param name="dbsnp_input_rod" type="data" format="vcf" label="dbSNP ROD file" /> | |
| 206 <param name="dbsnp_rod_name" type="hidden" value="dbsnp" label="dbSNP ROD Name"/> | |
| 207 <param name="dbsnp_known_names" type="boolean" truevalue="--known_names" falsevalue="" label="Use dbSNP ROD as known_names" help="-knownName,--known_names &lt;known_names&gt;" /> | |
| 208 </when> | |
| 209 </conditional> | |
| 210 | |
| 211 <conditional name="gatk_param_type"> | |
| 212 <param name="gatk_param_type_selector" type="select" label="Basic or Advanced GATK options"> | |
| 213 <option value="basic" selected="True">Basic</option> | |
| 214 <option value="advanced">Advanced</option> | |
| 215 </param> | |
| 216 <when value="basic"> | |
| 217 <!-- Do nothing here --> | |
| 218 </when> | |
| 219 <when value="advanced"> | |
| 220 <repeat name="pedigree" title="Pedigree file" help="-ped,--pedigree &lt;pedigree&gt;"> | |
| 221 <param name="pedigree_file" type="data" format="txt" label="Pedigree files for samples"/> | |
| 222 </repeat> | |
| 223 <repeat name="pedigree_string_repeat" title="Pedigree string" help="-pedString,--pedigreeString &lt;pedigreeString&gt;"> | |
| 224 <param name="pedigree_string" type="text" value="" label="Pedigree string for samples"/> | |
| 225 </repeat> | |
| 226 <param name="pedigree_validation_type" type="select" label="How strict should we be in validating the pedigree information" help="-pedValidationType,--pedigreeValidationType &lt;pedigreeValidationType&gt;"> | |
| 227 <option value="STRICT" selected="True">STRICT</option> | |
| 228 <option value="SILENT">SILENT</option> | |
| 229 </param> | |
| 230 <repeat name="read_filter" title="Read Filter" help="-rf,--read_filter &lt;read_filter&gt;"> | |
| 231 <conditional name="read_filter_type"> | |
| 232 <param name="read_filter_type_selector" type="select" label="Read Filter Type"> | |
| 233 <option value="BadCigar">BadCigar</option> | |
| 234 <option value="BadMate">BadMate</option> | |
| 235 <option value="DuplicateRead">DuplicateRead</option> | |
| 236 <option value="FailsVendorQualityCheck">FailsVendorQualityCheck</option> | |
| 237 <option value="MalformedRead">MalformedRead</option> | |
| 238 <option value="MappingQuality">MappingQuality</option> | |
| 239 <option value="MappingQualityUnavailable">MappingQualityUnavailable</option> | |
| 240 <option value="MappingQualityZero">MappingQualityZero</option> | |
| 241 <option value="MateSameStrand">MateSameStrand</option> | |
| 242 <option value="MaxInsertSize">MaxInsertSize</option> | |
| 243 <option value="MaxReadLength" selected="True">MaxReadLength</option> | |
| 244 <option value="MissingReadGroup">MissingReadGroup</option> | |
| 245 <option value="NoOriginalQualityScores">NoOriginalQualityScores</option> | |
| 246 <option value="NotPrimaryAlignment">NotPrimaryAlignment</option> | |
| 247 <option value="Platform454">Platform454</option> | |
| 248 <option value="Platform">Platform</option> | |
| 249 <option value="PlatformUnit">PlatformUnit</option> | |
| 250 <option value="ReadGroupBlackList">ReadGroupBlackList</option> | |
| 251 <option value="ReadName">ReadName</option> | |
| 252 <option value="ReadStrand">ReadStrand</option> | |
| 253 <option value="ReassignMappingQuality">ReassignMappingQuality</option> | |
| 254 <option value="Sample">Sample</option> | |
| 255 <option value="SingleReadGroup">SingleReadGroup</option> | |
| 256 <option value="UnmappedRead">UnmappedRead</option> | |
| 257 </param> | |
| 258 <when value="BadCigar"> | |
| 259 <!-- no extra options --> | |
| 260 </when> | |
| 261 <when value="BadMate"> | |
| 262 <!-- no extra options --> | |
| 263 </when> | |
| 264 <when value="DuplicateRead"> | |
| 265 <!-- no extra options --> | |
| 266 </when> | |
| 267 <when value="FailsVendorQualityCheck"> | |
| 268 <!-- no extra options --> | |
| 269 </when> | |
| 270 <when value="MalformedRead"> | |
| 35 | 271 <param name="filter_mismatching_base_and_quals" type="boolean" truevalue="--filter_mismatching_base_and_quals" falsevalue="" checked="false" label="filter out the reads with mismatching number of bases and base qualities" help="filter out the mismatch reads instead of quitting with an error"/> |
| 0 | 272 </when> |
| 273 <when value="MappingQuality"> | |
| 274 <param name="min_mapping_quality_score" type="integer" value="10" label="Minimum read mapping quality required to consider a read for calling"/> | |
| 275 </when> | |
| 276 <when value="MappingQualityUnavailable"> | |
| 277 <!-- no extra options --> | |
| 278 </when> | |
| 279 <when value="MappingQualityZero"> | |
| 280 <!-- no extra options --> | |
| 281 </when> | |
| 282 <when value="MateSameStrand"> | |
| 283 <!-- no extra options --> | |
| 284 </when> | |
| 285 <when value="MaxInsertSize"> | |
| 286 <param name="maxInsertSize" type="integer" value="1000000" label="Discard reads with insert size greater than the specified value"/> | |
| 287 </when> | |
| 288 <when value="MaxReadLength"> | |
| 289 <param name="maxReadLength" type="integer" value="76" label="Max Read Length"/> | |
| 290 </when> | |
| 291 <when value="MissingReadGroup"> | |
| 292 <!-- no extra options --> | |
| 293 </when> | |
| 294 <when value="NoOriginalQualityScores"> | |
| 295 <!-- no extra options --> | |
| 296 </when> | |
| 297 <when value="NotPrimaryAlignment"> | |
| 298 <!-- no extra options --> | |
| 299 </when> | |
| 300 <when value="Platform454"> | |
| 301 <!-- no extra options --> | |
| 302 </when> | |
| 303 <when value="Platform"> | |
| 304 <param name="PLFilterName" type="text" value="" label="Discard reads with RG:PL attribute containing this string"/> | |
| 305 </when> | |
| 306 <when value="PlatformUnit"> | |
| 307 <!-- no extra options --> | |
| 308 </when> | |
| 309 <when value="ReadGroupBlackList"> | |
| 310 <!-- no extra options --> | |
| 311 </when> | |
| 312 <when value="ReadName"> | |
| 313 <param name="readName" type="text" value="" label="Filter out all reads except those with this read name"/> | |
| 314 </when> | |
| 315 <when value="ReadStrand"> | |
| 316 <param name="filterPositive" type="boolean" truevalue="--filterPositive" falsevalue="" label="Discard reads on the forward strand"/> | |
| 317 </when> | |
| 318 <when value="ReassignMappingQuality"> | |
| 319 <param name="default_mapping_quality" type="integer" value="60" label="Default read mapping quality to assign to all reads"/> | |
| 320 </when> | |
| 321 <when value="Sample"> | |
| 322 <param name="sample_to_keep" type="text" value="" label="The name of the sample(s) to keep, filtering out all others"/> | |
| 323 </when> | |
| 324 <when value="SingleReadGroup"> | |
| 325 <param name="read_group_to_keep" type="integer" value="76" label="The name of the read group to keep, filtering out all others"/> | |
| 326 </when> | |
| 327 <when value="UnmappedRead"> | |
| 328 <!-- no extra options --> | |
| 329 </when> | |
| 330 </conditional> | |
| 331 </repeat> | |
| 332 <repeat name="input_interval_repeat" title="Operate on Genomic intervals" help="-L,--intervals &lt;intervals&gt;"> | |
| 333 <param name="input_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" /> | |
| 334 </repeat> | |
| 335 <repeat name="input_exclude_interval_repeat" title="Exclude Genomic intervals" help="-XL,--excludeIntervals &lt;excludeIntervals&gt;"> | |
| 336 <param name="input_exclude_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" /> | |
| 337 </repeat> | |
| 338 | |
| 339 <param name="interval_set_rule" type="select" label="Interval set rule" help="-isr,--interval_set_rule &lt;interval_set_rule&gt;"> | |
| 340 <option value="UNION" selected="True">UNION</option> | |
| 341 <option value="INTERSECTION">INTERSECTION</option> | |
| 342 </param> | |
| 343 | |
| 344 <conditional name="downsampling_type"> | |
| 345 <param name="downsampling_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="-dt,--downsampling_type &lt;downsampling_type&gt;"> | |
| 346 <option value="NONE" selected="True">NONE</option> | |
| 347 <option value="ALL_READS">ALL_READS</option> | |
| 348 <option value="BY_SAMPLE">BY_SAMPLE</option> | |
| 349 </param> | |
| 350 <when value="NONE"> | |
| 351 <!-- no more options here --> | |
| 352 </when> | |
| 353 <when value="ALL_READS"> | |
| 354 <conditional name="downsample_to_type"> | |
| 355 <param name="downsample_to_type_selector" type="select" label="Downsample method"> | |
| 356 <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option> | |
| 357 <option value="downsample_to_coverage">Downsample by Coverage</option> | |
| 358 </param> | |
| 359 <when value="downsample_to_fraction"> | |
| 360 <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 &lt;downsample_to_fraction&gt;"/> | |
| 361 </when> | |
| 362 <when value="downsample_to_coverage"> | |
| 363 <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0" help="-dcov,--downsample_to_coverage &lt;downsample_to_coverage&gt;"/> | |
| 364 </when> | |
| 365 </conditional> | |
| 366 </when> | |
| 367 <when value="BY_SAMPLE"> | |
| 368 <conditional name="downsample_to_type"> | |
| 369 <param name="downsample_to_type_selector" type="select" label="Downsample method"> | |
| 370 <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option> | |
| 371 <option value="downsample_to_coverage">Downsample by Coverage</option> | |
| 372 </param> | |
| 373 <when value="downsample_to_fraction"> | |
| 374 <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 &lt;downsample_to_fraction&gt;"/> | |
| 375 </when> | |
| 376 <when value="downsample_to_coverage"> | |
| 377 <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0" help="-dcov,--downsample_to_coverage &lt;downsample_to_coverage&gt;"/> | |
| 378 </when> | |
| 379 </conditional> | |
| 380 </when> | |
| 381 </conditional> | |
| 382 <param name="baq" type="select" label="Type of BAQ calculation to apply in the engine" help="-baq,--baq &lt;baq&gt;"> | |
| 383 <option value="OFF" selected="True">OFF</option> | |
| 384 <option value="CALCULATE_AS_NECESSARY">CALCULATE_AS_NECESSARY</option> | |
| 385 <option value="RECALCULATE">RECALCULATE</option> | |
| 386 </param> | |
| 387 <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 &lt;baqGapOpenPenalty&gt;" /> | |
| 388 <param name="use_original_qualities" type="boolean" truevalue="--useOriginalQualities" falsevalue="" label="Use the original base quality scores from the OQ tag" help="-OQ,--useOriginalQualities" /> | |
| 389 <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 &lt;defaultBaseQualities&gt;"/> | |
| 390 <param name="validation_strictness" type="select" label="How strict should we be with validation" help="-S,--validation_strictness &lt;validation_strictness&gt;"> | |
| 391 <option value="STRICT" selected="True">STRICT</option> | |
| 392 <option value="LENIENT">LENIENT</option> | |
| 393 <option value="SILENT">SILENT</option> | |
| 394 <!-- <option value="DEFAULT_STRINGENCY">DEFAULT_STRINGENCY</option> listed in docs, but not valid value...--> | |
| 395 </param> | |
| 396 <param name="interval_merging" type="select" label="Interval merging rule" help="-im,--interval_merging &lt;interval_merging&gt;"> | |
| 397 <option value="ALL" selected="True">ALL</option> | |
| 398 <option value="OVERLAPPING_ONLY">OVERLAPPING_ONLY</option> | |
| 399 </param> | |
| 400 | |
| 401 <repeat name="read_group_black_list_repeat" title="Read group black list" help="-rgbl,--read_group_black_list &lt;read_group_black_list&gt;"> | |
| 402 <conditional name="read_group_black_list_type"> | |
| 403 <param name="read_group_black_list_type_selector" type="select" label="Type of reads read group black list"> | |
| 404 <option value="file" selected="True">Filters in file</option> | |
| 405 <option value="text">Specify filters as a string</option> | |
| 406 </param> | |
| 407 <when value="file"> | |
| 408 <param name="read_group_black_list" type="data" format="txt" label="Read group black list file" /> | |
| 409 </when> | |
| 410 <when value="text"> | |
| 411 <param name="read_group_black_list" type="text" value="tag:string" label="Read group black list tag:string" /> | |
| 412 </when> | |
| 413 </conditional> | |
| 414 </repeat> | |
| 415 | |
| 416 <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"/> | |
| 417 <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"/> | |
|
18
7533db8dfb5b
Update tool_dependencies to GATK v 2.3
Jim Johnson <jj@umn.edu>
parents:
0
diff
changeset
|
418 <param name="fix_misencoded_quality_scores" type="boolean" truevalue="--fix_misencoded_quality_scores" falsevalue="" label="Fix mis-encoded base quality scores. Q0 == ASCII 33 according to the SAM specification, whereas Illumina encoding starts at Q64. The idea here is simple: we just iterate over all reads and subtract 31 from every quality score." checked="False" help="-fixMisencodedQuals / --fix_misencoded_quality_scores"/> |
| 0 | 419 |
| 420 </when> | |
| 421 </conditional> | |
| 422 | |
| 423 | |
| 424 <conditional name="analysis_param_type"> | |
| 425 <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options"> | |
| 426 <option value="basic" selected="True">Basic</option> | |
| 427 <option value="advanced">Advanced</option> | |
| 428 </param> | |
| 429 <when value="basic"> | |
| 430 <!-- Do nothing here --> | |
| 431 </when> | |
| 432 <when value="advanced"> | |
| 433 <repeat name="stratifications" title="Stratification"> | |
| 434 <param name="select_exps" value="" type="text" label="Stratification Expression" help="-select,--select_exps &lt;select_exps&gt;"> | |
| 435 <sanitizer> | |
| 436 <valid initial="string.printable"> | |
| 437 <remove value="'"/> | |
| 438 </valid> | |
| 439 <mapping initial="none"/> | |
| 440 </sanitizer> | |
| 441 </param> | |
| 442 <param name="select_name" value="" type="text" label="Name" help="-selectName,--select_names &lt;select_names&gt;"/> | |
| 443 </repeat> | |
| 444 | |
| 445 <repeat name="samples" title="Sample" help="-sn,--sample &lt;sample&gt;"> | |
| 446 <param name="sample" value="" type="text" label="Derive eval and comp contexts using only these sample genotypes, when genotypes are available in the original context"/> | |
| 447 </repeat> | |
| 448 | |
| 449 <param name="stratification_modules" type="select" multiple="True" display="checkboxes" label="Stratification modules to apply to the eval track(s)" help="-ST,--stratificationModule &lt;stratificationModule&gt;" > | |
| 450 <!-- do these need individual options also? gatk wiki has little info --> | |
| 451 <option value="AlleleFrequency" /> | |
| 452 <option value="AlleleCount" /> | |
| 453 <option value="CompRod" /> | |
| 454 <option value="Contig" /> | |
| 455 <option value="CpG" /> | |
| 456 <option value="Degeneracy" /> | |
| 457 <option value="EvalRod" /> | |
| 458 <option value="Filter" /> | |
| 459 <option value="FunctionalClass" /> | |
| 460 <option value="JexlExpression" /> | |
| 461 <option value="Sample" /> | |
| 462 <option value="IntervalStratification" /> | |
| 463 </param> | |
| 464 <param name="do_not_use_all_standard_stratifications" checked="false" type="boolean" truevalue="--doNotUseAllStandardStratifications" falsevalue="" label="Do not use the standard stratification modules by default" help="-noST,--doNotUseAllStandardStratifications" /> | |
| 465 | |
| 466 <repeat name="only_variants_of_type" title="only Variants Of Type" help="--onlyVariantsOfType"> | |
| 467 <param name="variant_type" type="text" value="" label="only variants of these types will be considered during the evaluation"/> | |
| 468 </repeat> | |
| 469 | |
| 470 <param name="eval_modules" type="select" multiple="True" display="checkboxes" label="Eval modules to apply to the eval track(s)" help="-EV,--evalModule &lt;evalModule&gt;" > | |
| 471 <!-- do these need individual options also? gatk wiki has little info --> | |
| 472 <option value="ACTransitionTable" /> | |
| 473 <option value="AlleleFrequencyComparison" /> | |
| 474 <option value="AminoAcidTransition" /> | |
| 475 <option value="CompOverlap" /> | |
| 476 <option value="CountVariants" /> | |
| 477 <option value="GenotypeConcordance" /> | |
| 478 <option value="GenotypePhasingEvaluator" /> | |
| 479 <option value="IndelMetricsByAC" /> | |
| 480 <option value="IndelStatistics" /> | |
| 481 <option value="MendelianViolationEvaluator" /> | |
| 482 <option value="PrintMissingComp" /> | |
| 483 <option value="PrivatePermutations" /> | |
| 484 <option value="SimpleMetricsByAC" /> | |
| 485 <option value="ThetaVariantEvaluator" /> | |
| 486 <option value="TiTvVariantEvaluator" /> | |
| 487 <option value="VariantQualityScore" /> | |
| 488 </param> | |
| 489 <param name="do_not_use_all_standard_modules" checked="false" type="boolean" truevalue="--doNotUseAllStandardModules" falsevalue="" label="Do not use the standard eval modules by default" help="-noEV,--doNotUseAllStandardModules" /> | |
| 490 | |
| 491 <param name="num_samples" type="integer" label="Number of samples (used if no samples are available in the VCF file" value="0" help="-ns,--numSamples &lt;numSamples&gt;"/> | |
| 492 <param name="min_phase_quality" type="float" label="Minimum phasing quality " value="10.0" help="-mpq,--minPhaseQuality &lt;minPhaseQuality&gt;"/> | |
| 493 <param name="family" type="text" value="" label="If provided, genotypes in will be examined for mendelian violations: this argument is a string formatted as dad+mom=child where these parameters determine which sample names are examined" help="--family_structure"/> | |
| 494 <param name="mendelian_violation_qual_threshold" type="integer" label="Minimum genotype QUAL score for each trio member required to accept a site as a violation" value="50" help="-mvq,--mendelianViolationQualThreshold &lt;mendelianViolationQualThreshold&gt;"/> | |
| 495 <param name="ancestral_alignments" type="data" format="fasta" optional="True" label="Fasta file with ancestral alleles" help="-aa,--ancestralAlignments &lt;ancestralAlignments&gt;" /> | |
| 496 <param name="known_cnvs" type="data" format="bed,gatk_interval,picard_interval_list" optional="True" label="File containing tribble-readable features describing a known list of copy number variants" help="-knownCNVs,--knownCNVs &lt;knownCNVs&gt;" /> | |
| 497 <param name="strat_intervals" type="data" format="bed,gatk_interval,picard_interval_list" optional="True" label="File containing tribble-readable features for the IntervalStratificiation" help="-stratIntervals,--stratIntervals &lt;stratIntervals&gt;" /> | |
| 498 | |
| 499 </when> | |
| 500 </conditional> | |
| 501 | |
| 502 | |
| 503 </inputs> | |
| 504 <outputs> | |
| 505 <data format="gatk_report" name="output_report" label="${tool.name} on ${on_string} (report)" /> | |
| 506 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" /> | |
| 507 </outputs> | |
| 508 <tests> | |
| 509 <test> | |
| 510 <param name="reference_source_selector" value="history" /> | |
| 511 <param name="ref_file" value="phiX.fasta" ftype="fasta" /> | |
| 512 <param name="input_variant" value="gatk/gatk_variant_annotator/gatk_variant_annotator_out_1.vcf" ftype="vcf" /> | |
| 513 <param name="dbsnp_rod_bind_type_selector" value="set_dbsnp" /> | |
| 514 <param name="dbsnp_input_rod" value="gatk/fake_phiX_variant_locations.vcf" ftype="vcf" /> | |
| 515 <param name="dbsnp_known_names" value="True"/> | |
| 516 <param name="comp_rod_bind" value="0" /> | |
| 517 <param name="gatk_param_type_selector" value="basic" /> | |
| 518 <param name="analysis_param_type_selector" value="basic" /> | |
| 519 <output name="output_report" file="gatk/gatk_variant_eval/gatk_variant_eval_out_1.gatk_report" /> | |
| 520 <output name="output_log" file="gatk/gatk_variant_eval/gatk_variant_eval_out_1.log.contains" compare="contains" /> | |
| 521 </test> | |
| 522 </tests> | |
| 523 <help> | |
| 524 **What it does** | |
| 525 | |
| 526 General-purpose tool for variant evaluation (% in dbSNP, genotype concordance, Ti/Tv ratios, and a lot more) | |
| 527 | |
| 528 For more information on using the VariantEval module, see this `tool specific page <http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_varianteval_VariantEval.html>`_. | |
| 529 | |
| 530 To learn about best practices for variant detection using GATK, see this `overview <http://www.broadinstitute.org/gatk/guide/topic?name=best-practices>`_. | |
| 531 | |
| 532 If you encounter errors, please view the `GATK FAQ <http://www.broadinstitute.org/gatk/guide/topic?name=faqs>`_. | |
| 533 | |
| 534 ------ | |
| 535 | |
| 536 **Inputs** | |
| 537 | |
| 538 GenomeAnalysisTK: VariantEval accepts variant files as input. | |
| 539 | |
| 540 | |
| 541 **Outputs** | |
| 542 | |
| 543 The output is a table of variant evaluation. | |
| 544 | |
| 545 | |
| 546 Go `here <http://www.broadinstitute.org/gatk/guide/topic?name=intro>`_ for details on GATK file formats. | |
| 547 | |
| 548 ------- | |
| 549 | |
| 550 **Settings**:: | |
| 551 | |
| 552 out An output file presented to the walker. Will overwrite contents if file exists. | |
| 553 list List the available eval modules and exit | |
| 554 select_exps One or more stratifications to use when evaluating the data | |
| 555 select_names Names to use for the list of stratifications (must be a 1-to-1 mapping) | |
| 556 sample Derive eval and comp contexts using only these sample genotypes, when genotypes are available in the original context | |
| 557 known_names Name of ROD bindings containing variant sites that should be treated as known when splitting eval rods into known and novel subsets | |
| 558 stratificationModule One or more specific stratification modules to apply to the eval track(s) (in addition to the standard stratifications, unless -noS is specified) | |
| 559 doNotUseAllStandardStratifications Do not use the standard stratification modules by default (instead, only those that are specified with the -S option) | |
| 560 onlyVariantsOfType If provided, only variants of these types will be considered during the evaluation, in | |
| 561 evalModule One or more specific eval modules to apply to the eval track(s) (in addition to the standard modules, unless -noE is specified) | |
| 562 doNotUseAllStandardModules Do not use the standard modules by default (instead, only those that are specified with the -E option) | |
| 563 numSamples Number of samples (used if no samples are available in the VCF file | |
| 564 minPhaseQuality Minimum phasing quality | |
| 565 family_structure If provided, genotypes in will be examined for mendelian violations: this argument is a string formatted as dad+mom=child where these parameters determine which sample names are examined | |
| 566 mendelianViolationQualThreshold Minimum genotype QUAL score for each trio member required to accept a site as a violation | |
| 567 ancestralAlignments Fasta file with ancestral alleles | |
| 568 | |
| 569 ------ | |
| 570 | |
| 571 **Citation** | |
| 572 | |
| 573 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. <http://www.ncbi.nlm.nih.gov/pubmed/21478889>`_ | |
| 574 | |
| 575 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. <http://www.ncbi.nlm.nih.gov/pubmed/20644199>`_ | |
| 576 | |
| 577 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. <http://www.ncbi.nlm.nih.gov/pubmed/20069535>`_ | |
| 578 | |
| 579 </help> | |
| 580 </tool> |