Mercurial > repos > iuc > gatk2
comparison realigner_target_creator.xml @ 0:340633249b3d draft
Uploaded
author | bgruening |
---|---|
date | Mon, 02 Dec 2013 06:18:36 -0500 |
parents | |
children | f244b8209eb8 |
comparison
equal
deleted
inserted
replaced
-1:000000000000 | 0:340633249b3d |
---|---|
1 <tool id="gatk2_realigner_target_creator" name="Realigner Target Creator" version="0.0.7"> | |
2 <description>for use in local realignment</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 "${output_log}" | |
10 -d "-I" "${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 ' | |
15 @JAR_PATH@ | |
16 -T "RealignerTargetCreator" | |
17 -o "${output_interval}" | |
18 | |
19 \$GATK2_SITE_OPTIONS | |
20 | |
21 ## according to http://www.broadinstitute.org/gatk/guide/article?id=1975 | |
22 --num_cpu_threads_per_data_thread 1 | |
23 | |
24 @THREADS@ | |
25 | |
26 #if $reference_source.reference_source_selector != "history": | |
27 -R "${reference_source.ref_file.fields.path}" | |
28 #end if | |
29 ' | |
30 #set $rod_binding_names = dict() | |
31 #for $rod_binding in $rod_bind: | |
32 #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom': | |
33 #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name | |
34 #else | |
35 #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector | |
36 #end if | |
37 #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1 | |
38 -d "-known:${rod_bind_name},%(file_type)s" "${rod_binding.rod_bind_type.input_rod}" "${rod_binding.rod_bind_type.input_rod.ext}" "input_${rod_bind_name}_${rod_binding_names[$rod_bind_name]}" | |
39 #end for | |
40 | |
41 #include source=$standard_gatk_options# | |
42 ##start analysis specific options | |
43 #if $analysis_param_type.analysis_param_type_selector == "advanced": | |
44 -p ' | |
45 --minReadsAtLocus "${analysis_param_type.minReadsAtLocus}" | |
46 --windowSize "${analysis_param_type.windowSize}" | |
47 --mismatchFraction "${analysis_param_type.mismatchFraction}" | |
48 --maxIntervalSize "${analysis_param_type.maxIntervalSize}" | |
49 ' | |
50 #end if | |
51 </command> | |
52 <inputs> | |
53 <conditional name="reference_source"> | |
54 <expand macro="reference_source_selector_param" /> | |
55 <when value="cached"> | |
56 <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &lt;input_file&gt;"> | |
57 <validator type="unspecified_build" /> | |
58 <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 --> | |
59 </param> | |
60 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &lt;reference_sequence&gt;" > | |
61 <options from_data_table="gatk2_picard_indexes"> | |
62 <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/> | |
63 </options> | |
64 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/> | |
65 </param> | |
66 </when> | |
67 <when value="history"> | |
68 <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &lt;input_file&gt;" /> | |
69 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &lt;reference_sequence&gt;"> | |
70 <options> | |
71 <filter type="data_meta" key="dbkey" ref="input_bam" /> | |
72 </options> | |
73 </param> | |
74 </when> | |
75 </conditional> | |
76 | |
77 <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-known,--known &lt;known&gt;)"> | |
78 <conditional name="rod_bind_type"> | |
79 <param name="rod_bind_type_selector" type="select" label="Variant Type"> | |
80 <option value="dbsnp" selected="True">dbSNP</option> | |
81 <option value="snps">SNPs</option> | |
82 <option value="indels">INDELs</option> | |
83 <option value="custom">Custom</option> | |
84 </param> | |
85 <when value="dbsnp"> | |
86 <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" /> | |
87 </when> | |
88 <when value="snps"> | |
89 <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" /> | |
90 </when> | |
91 <when value="indels"> | |
92 <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" /> | |
93 </when> | |
94 <when value="custom"> | |
95 <param name="custom_rod_name" type="text" value="Unknown" label="Customer's variant file" /> | |
96 <param name="input_rod" type="data" format="vcf" label="Variant file (VCF format)" /> | |
97 </when> | |
98 </conditional> | |
99 </repeat> | |
100 | |
101 <expand macro="gatk_param_type_conditional" /> | |
102 | |
103 <expand macro="analysis_type_conditional"> | |
104 <param name="windowSize" type="integer" value="10" label="Window size for calculating entropy or SNP clusters (windowSize)" help="-window,--windowSize &lt;windowSize&gt;" /> | |
105 <param name="mismatchFraction" type="float" value="0.15" label="Fraction of base qualities needing to mismatch for a position to have high entropy (mismatchFraction)" help="to disable set to <= 0 or > 1 (-mismatch,--mismatchFraction &lt;mismatchFraction&gt;)"/> | |
106 <param name="minReadsAtLocus" type="integer" value="4" label="Minimum reads at a locus to enable using the entropy calculation (minReadsAtLocus)" help="-minReads,--minReadsAtLocus &lt;minReadsAtLocus&gt;" /> | |
107 <param name="maxIntervalSize" type="integer" value="500" label="Maximum interval size" help="-maxInterval,--maxIntervalSize &lt;maxIntervalSize&gt;" /> | |
108 </expand> | |
109 </inputs> | |
110 <outputs> | |
111 <data format="gatk_interval" name="output_interval" label="${tool.name} on ${on_string} (GATK intervals)" /> | |
112 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" /> | |
113 </outputs> | |
114 <tests> | |
115 <test> | |
116 <param name="reference_source_selector" value="history" /> | |
117 <param name="ref_file" value="phiX.fasta" ftype="fasta" /> | |
118 <param name="input_bam" value="gatk/fake_phiX_reads_1.bam" ftype="bam" /> | |
119 <param name="rod_bind_type_selector" value="dbsnp" /> | |
120 <param name="input_rod" value="gatk/fake_phiX_variant_locations.vcf" ftype="vcf" /> | |
121 <param name="gatk_param_type_selector" value="basic" /> | |
122 <param name="analysis_param_type_selector" value="advanced" /> | |
123 <param name="windowSize" value="10" /> | |
124 <param name="mismatchFraction" value="0.15" /> | |
125 <param name="minReadsAtLocus" value="4" /> | |
126 <param name="maxIntervalSize" value="500" /> | |
127 <output name="output_interval" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.gatk_interval" /> | |
128 <output name="output_log" file="gatk/gatk_realigner_target_creator/gatk_realigner_target_creator_out_1.log.contains" compare="contains"/> | |
129 </test> | |
130 </tests> | |
131 <help> | |
132 **What it does** | |
133 | |
134 Emits intervals for the Local Indel Realigner to target for cleaning. Ignores 454 reads, MQ0 reads, and reads with consecutive indel operators in the CIGAR string. | |
135 | |
136 For more information on local realignment around indels using the GATK, see this `tool specific page <http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_indels_RealignerTargetCreator.html>`_. | |
137 | |
138 To learn about best practices for variant detection using GATK, see this `overview <http://www.broadinstitute.org/gatk/guide/topic?name=best-practices>`_. | |
139 | |
140 If you encounter errors, please view the `GATK FAQ <http://www.broadinstitute.org/gatk/guide/topic?name=faqs>`_. | |
141 | |
142 ------ | |
143 | |
144 **Inputs** | |
145 | |
146 GenomeAnalysisTK: RealignerTargetCreator accepts an aligned BAM input file. | |
147 | |
148 | |
149 **Outputs** | |
150 | |
151 The output is in GATK Interval format. | |
152 | |
153 | |
154 Go `here <http://www.broadinstitute.org/gatk/guide/topic?name=intro>`_ for details on GATK file formats. | |
155 | |
156 ------- | |
157 | |
158 **Settings**:: | |
159 | |
160 windowSize window size for calculating entropy or SNP clusters | |
161 mismatchFraction fraction of base qualities needing to mismatch for a position to have high entropy; to disable set to <= 0 or > 1 | |
162 minReadsAtLocus minimum reads at a locus to enable using the entropy calculation | |
163 maxIntervalSize maximum interval size | |
164 | |
165 @CITATION_SECTION@ | |
166 </help> | |
167 </tool> |