annotate base_recalibrator.xml @ 0:340633249b3d draft

Uploaded
author bgruening
date Mon, 02 Dec 2013 06:18:36 -0500
parents
children 8bcc13094767
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
340633249b3d Uploaded
bgruening
parents:
diff changeset
1 <tool id="gatk2_base_recalibrator" name="Base Recalibrator" version="0.0.7">
340633249b3d Uploaded
bgruening
parents:
diff changeset
2 <description>calculates covariates used to recalibrate base quality scores of reads</description>
340633249b3d Uploaded
bgruening
parents:
diff changeset
3 <expand macro="requirements" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
4 <macros>
340633249b3d Uploaded
bgruening
parents:
diff changeset
5 <import>gatk2_macros.xml</import>
340633249b3d Uploaded
bgruening
parents:
diff changeset
6 </macros>
340633249b3d Uploaded
bgruening
parents:
diff changeset
7 <command interpreter="python">
340633249b3d Uploaded
bgruening
parents:
diff changeset
8 gatk2_wrapper.py
340633249b3d Uploaded
bgruening
parents:
diff changeset
9 --stdout "${output_log}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
10 -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
340633249b3d Uploaded
bgruening
parents:
diff changeset
11 #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
340633249b3d Uploaded
bgruening
parents:
diff changeset
12 -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
340633249b3d Uploaded
bgruening
parents:
diff changeset
13 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
14 -p '
340633249b3d Uploaded
bgruening
parents:
diff changeset
15 @JAR_PATH@
340633249b3d Uploaded
bgruening
parents:
diff changeset
16 -T "BaseRecalibrator"
340633249b3d Uploaded
bgruening
parents:
diff changeset
17 \$GATK2_SITE_OPTIONS
340633249b3d Uploaded
bgruening
parents:
diff changeset
18
340633249b3d Uploaded
bgruening
parents:
diff changeset
19 ## according to http://www.broadinstitute.org/gatk/guide/article?id=1975
340633249b3d Uploaded
bgruening
parents:
diff changeset
20 --num_cpu_threads_per_data_thread 8
340633249b3d Uploaded
bgruening
parents:
diff changeset
21
340633249b3d Uploaded
bgruening
parents:
diff changeset
22 @THREADS@
340633249b3d Uploaded
bgruening
parents:
diff changeset
23 ## we set non standards at every run and the user can choose which ones are preferred
340633249b3d Uploaded
bgruening
parents:
diff changeset
24 ## in our select box both standard options (ContextCovariate, CycleCovariate) are selected by default
340633249b3d Uploaded
bgruening
parents:
diff changeset
25 --no_standard_covs
340633249b3d Uploaded
bgruening
parents:
diff changeset
26
340633249b3d Uploaded
bgruening
parents:
diff changeset
27 #if $reference_source.reference_source_selector != "history":
340633249b3d Uploaded
bgruening
parents:
diff changeset
28 -R "${reference_source.ref_file.fields.path}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
29 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
30 #if str($input_recal) != 'None':
340633249b3d Uploaded
bgruening
parents:
diff changeset
31 --BQSR "${input_recal}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
32 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
33 --out "${output_recal}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
34 #if str( $covariates ) != "None":
340633249b3d Uploaded
bgruening
parents:
diff changeset
35 #for $cov in str( $covariates ).split( ',' ):
340633249b3d Uploaded
bgruening
parents:
diff changeset
36 -cov "${cov}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
37 #end for
340633249b3d Uploaded
bgruening
parents:
diff changeset
38 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
39 '
340633249b3d Uploaded
bgruening
parents:
diff changeset
40
340633249b3d Uploaded
bgruening
parents:
diff changeset
41 #set $snp_dataset_provided = False
340633249b3d Uploaded
bgruening
parents:
diff changeset
42 #set $rod_binding_names = dict()
340633249b3d Uploaded
bgruening
parents:
diff changeset
43 #for $rod_binding in $rod_bind:
340633249b3d Uploaded
bgruening
parents:
diff changeset
44 #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'custom':
340633249b3d Uploaded
bgruening
parents:
diff changeset
45 #set $rod_bind_name = $rod_binding.rod_bind_type.custom_rod_name
340633249b3d Uploaded
bgruening
parents:
diff changeset
46 #else
340633249b3d Uploaded
bgruening
parents:
diff changeset
47 #set $rod_bind_name = $rod_binding.rod_bind_type.rod_bind_type_selector
340633249b3d Uploaded
bgruening
parents:
diff changeset
48 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
49 #if str( $rod_binding.rod_bind_type.rod_bind_type_selector ) == 'dbsnp':
340633249b3d Uploaded
bgruening
parents:
diff changeset
50 #set $snp_dataset_provided = True
340633249b3d Uploaded
bgruening
parents:
diff changeset
51 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
52 #set $rod_binding_names[$rod_bind_name] = $rod_binding_names.get( $rod_bind_name, -1 ) + 1
340633249b3d Uploaded
bgruening
parents:
diff changeset
53 -d "--knownSites:${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[$rod_bind_name]}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
54 #end for
340633249b3d Uploaded
bgruening
parents:
diff changeset
55
340633249b3d Uploaded
bgruening
parents:
diff changeset
56 #include source=$standard_gatk_options#
340633249b3d Uploaded
bgruening
parents:
diff changeset
57
340633249b3d Uploaded
bgruening
parents:
diff changeset
58 ##start analysis specific options
340633249b3d Uploaded
bgruening
parents:
diff changeset
59 #if $analysis_param_type.analysis_param_type_selector == "advanced":
340633249b3d Uploaded
bgruening
parents:
diff changeset
60 -p '
340633249b3d Uploaded
bgruening
parents:
diff changeset
61 #if $analysis_param_type.default_read_group_type.default_read_group_type_selector == "set":
340633249b3d Uploaded
bgruening
parents:
diff changeset
62 --default_read_group "${analysis_param_type.default_read_group_type.default_read_group}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
63 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
64 #if str( $analysis_param_type.default_platform ) != "default":
340633249b3d Uploaded
bgruening
parents:
diff changeset
65 --default_platform "${analysis_param_type.default_platform}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
66 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
67 #if str( $analysis_param_type.force_read_group_type.force_read_group_type_selector ) == "set":
340633249b3d Uploaded
bgruening
parents:
diff changeset
68 --force_read_group "${analysis_param_type.force_read_group_type.force_read_group}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
69 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
70 #if str( $analysis_param_type.force_platform ) != "default":
340633249b3d Uploaded
bgruening
parents:
diff changeset
71 --force_platform "${analysis_param_type.force_platform}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
72 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
73 ${analysis_param_type.exception_if_no_tile}
340633249b3d Uploaded
bgruening
parents:
diff changeset
74 #if str( $analysis_param_type.solid_options_type.solid_options_type_selector ) == "set":
340633249b3d Uploaded
bgruening
parents:
diff changeset
75 #if str( $analysis_param_type.solid_options_type.solid_recal_mode ) != "default":
340633249b3d Uploaded
bgruening
parents:
diff changeset
76 --solid_recal_mode "${analysis_param_type.solid_options_type.solid_recal_mode}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
77 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
78 #if str( $analysis_param_type.solid_options_type.solid_nocall_strategy ) != "default":
340633249b3d Uploaded
bgruening
parents:
diff changeset
79 --solid_nocall_strategy "${analysis_param_type.solid_options_type.solid_nocall_strategy}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
80 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
81 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
82 --window_size_nqs "${analysis_param_type.window_size_nqs}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
83 --homopolymer_nback "${analysis_param_type.homopolymer_nback}"
340633249b3d Uploaded
bgruening
parents:
diff changeset
84 '
340633249b3d Uploaded
bgruening
parents:
diff changeset
85 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
86 #if not $snp_dataset_provided:
340633249b3d Uploaded
bgruening
parents:
diff changeset
87 -p '--run_without_dbsnp_potentially_ruining_quality'
340633249b3d Uploaded
bgruening
parents:
diff changeset
88 #end if
340633249b3d Uploaded
bgruening
parents:
diff changeset
89 </command>
340633249b3d Uploaded
bgruening
parents:
diff changeset
90 <inputs>
340633249b3d Uploaded
bgruening
parents:
diff changeset
91 <conditional name="reference_source">
340633249b3d Uploaded
bgruening
parents:
diff changeset
92 <expand macro="reference_source_selector_param" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
93 <when value="cached">
340633249b3d Uploaded
bgruening
parents:
diff changeset
94 <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &amp;lt;input_file&amp;gt;">
340633249b3d Uploaded
bgruening
parents:
diff changeset
95 <validator type="unspecified_build" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
96 <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 -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
97 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
98 <param name="ref_file" type="select" label="Using reference genome" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;" >
340633249b3d Uploaded
bgruening
parents:
diff changeset
99 <options from_data_table="gatk2_picard_indexes">
340633249b3d Uploaded
bgruening
parents:
diff changeset
100 <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
101 </options>
340633249b3d Uploaded
bgruening
parents:
diff changeset
102 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
103 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
104 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
105 <when value="history">
340633249b3d Uploaded
bgruening
parents:
diff changeset
106 <param name="input_bam" type="data" format="bam" label="BAM file" help="-I,--input_file &amp;lt;input_file&amp;gt;" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
107 <param name="ref_file" type="data" format="fasta" label="Using reference file" help="-R,--reference_sequence &amp;lt;reference_sequence&amp;gt;">
340633249b3d Uploaded
bgruening
parents:
diff changeset
108 <options>
340633249b3d Uploaded
bgruening
parents:
diff changeset
109 <filter type="data_meta" key="dbkey" ref="input_bam" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
110 </options>
340633249b3d Uploaded
bgruening
parents:
diff changeset
111 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
112 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
113 </conditional>
340633249b3d Uploaded
bgruening
parents:
diff changeset
114 <param name="input_recal" type="data" format="gatk_report" optional="true" label="Covariates table recalibration file" help="-BQSR,--BQSR &amp;lt;recal_file&amp;gt;" >
340633249b3d Uploaded
bgruening
parents:
diff changeset
115 <help>The input covariates table file which enables on-the-fly base quality score recalibration.
340633249b3d Uploaded
bgruening
parents:
diff changeset
116 Enables on-the-fly recalibrate of base qualities. The covariates tables are produced by the BaseQualityScoreRecalibrator tool.
340633249b3d Uploaded
bgruening
parents:
diff changeset
117 Please be aware that one should only run recalibration with the covariates file created on the same input bam(s).
340633249b3d Uploaded
bgruening
parents:
diff changeset
118 </help>
340633249b3d Uploaded
bgruening
parents:
diff changeset
119 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
120
340633249b3d Uploaded
bgruening
parents:
diff changeset
121 <param name="covariates" type="select" multiple="True" display="checkboxes" label="Covariates to be used in the recalibration" help="-cov,--covariate &amp;lt;covariate&amp;gt;" >
340633249b3d Uploaded
bgruening
parents:
diff changeset
122 <!-- might we want to load the available covariates from an external configuration file, since additional ones can be added to local installs? -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
123 <option value="ContextCovariate" selected="true"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
124 <option value="CycleCovariate" selected="true"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
125 <option value="RepeatLengthCovariate" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
126 <option value="RepeatUnitCovariate" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
127 <option value="RepeatUnitAndLengthCovariate" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
128 <!--
340633249b3d Uploaded
bgruening
parents:
diff changeset
129 Note: ReadGroupCovariate and QualityScoreCovariate are required covariates and will
340633249b3d Uploaded
bgruening
parents:
diff changeset
130 be added for the user regardless of whether or not they were specified.
340633249b3d Uploaded
bgruening
parents:
diff changeset
131 <option value="QualityScoreCovariate" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
132 <option value="ReadGroupCovariate" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
133 -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
134 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
135
340633249b3d Uploaded
bgruening
parents:
diff changeset
136 <repeat name="rod_bind" title="Known Variants" help="Using data sets of known variants (-knownSites,--knownSites &amp;lt;knownSites&amp;gt;)">
340633249b3d Uploaded
bgruening
parents:
diff changeset
137 <conditional name="rod_bind_type">
340633249b3d Uploaded
bgruening
parents:
diff changeset
138 <param name="rod_bind_type_selector" type="select" label="Variant Type">
340633249b3d Uploaded
bgruening
parents:
diff changeset
139 <option value="dbsnp" selected="True">dbSNP</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
140 <option value="snps">SNPs</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
141 <option value="indels">INDELs</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
142 <option value="mask">Mask</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
143 <option value="custom">Custom</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
144 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
145 <when value="dbsnp">
340633249b3d Uploaded
bgruening
parents:
diff changeset
146 <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
147 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
148 <when value="snps">
340633249b3d Uploaded
bgruening
parents:
diff changeset
149 <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
150 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
151 <when value="indels">
340633249b3d Uploaded
bgruening
parents:
diff changeset
152 <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
153 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
154 <when value="mask">
340633249b3d Uploaded
bgruening
parents:
diff changeset
155 <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
156 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
157 <when value="custom">
340633249b3d Uploaded
bgruening
parents:
diff changeset
158 <param name="custom_rod_name" type="text" value="Unknown" label="Customer's variant file"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
159 <param name="input_rod" type="data" format="vcf,gatk_dbsnp,bed" label="Variant file" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
160 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
161 </conditional>
340633249b3d Uploaded
bgruening
parents:
diff changeset
162 </repeat>
340633249b3d Uploaded
bgruening
parents:
diff changeset
163
340633249b3d Uploaded
bgruening
parents:
diff changeset
164 <expand macro="gatk_param_type_conditional" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
165
340633249b3d Uploaded
bgruening
parents:
diff changeset
166 <conditional name="analysis_param_type">
340633249b3d Uploaded
bgruening
parents:
diff changeset
167 <param name="analysis_param_type_selector" type="select" label="Basic or Advanced Analysis options">
340633249b3d Uploaded
bgruening
parents:
diff changeset
168 <option value="basic" selected="True">Basic</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
169 <option value="advanced">Advanced</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
170 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
171 <when value="basic">
340633249b3d Uploaded
bgruening
parents:
diff changeset
172 <!-- Do nothing here -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
173 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
174 <when value="advanced">
340633249b3d Uploaded
bgruening
parents:
diff changeset
175 <conditional name="default_read_group_type">
340633249b3d Uploaded
bgruening
parents:
diff changeset
176 <param name="default_read_group_type_selector" type="select" label="Set default Read Group" help="--default_read_group">
340633249b3d Uploaded
bgruening
parents:
diff changeset
177 <option value="default" selected="True">Don't Set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
178 <option value="set">Set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
179 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
180 <when value="default">
340633249b3d Uploaded
bgruening
parents:
diff changeset
181 <!-- do nothing here -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
182 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
183 <when value="set">
340633249b3d Uploaded
bgruening
parents:
diff changeset
184 <param name="default_read_group" type="text" value="Unknown" label="If a read has no read group then default to the provided String"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
185 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
186 </conditional>
340633249b3d Uploaded
bgruening
parents:
diff changeset
187 <param name="default_platform" type="select" label="Set default Platform" help="--default_platform">
340633249b3d Uploaded
bgruening
parents:
diff changeset
188 <option value="default" selected="True">Don't Set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
189 <option value="illumina">illumina</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
190 <option value="454">454</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
191 <option value="solid">solid</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
192 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
193 <conditional name="force_read_group_type">
340633249b3d Uploaded
bgruening
parents:
diff changeset
194 <param name="force_read_group_type_selector" type="select" label="Force Read Group" help="--force_read_group">
340633249b3d Uploaded
bgruening
parents:
diff changeset
195 <option value="default" selected="True">Don't Force</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
196 <option value="set">Force</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
197 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
198 <when value="default">
340633249b3d Uploaded
bgruening
parents:
diff changeset
199 <!-- do nothing here -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
200 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
201 <when value="set">
340633249b3d Uploaded
bgruening
parents:
diff changeset
202 <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."/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
203 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
204 </conditional>
340633249b3d Uploaded
bgruening
parents:
diff changeset
205 <param name="force_platform" type="select" label="Force Platform" help="--force_platform">
340633249b3d Uploaded
bgruening
parents:
diff changeset
206 <option value="default" selected="True">Don't Force</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
207 <option value="illumina">illumina</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
208 <option value="454">454</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
209 <option value="solid">solid</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
210 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
211 <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"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
212 <conditional name="solid_options_type">
340633249b3d Uploaded
bgruening
parents:
diff changeset
213 <param name="solid_options_type_selector" type="select" label="Set SOLiD specific options">
340633249b3d Uploaded
bgruening
parents:
diff changeset
214 <option value="default" selected="True">Don't Set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
215 <option value="set">Set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
216 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
217 <when value="default">
340633249b3d Uploaded
bgruening
parents:
diff changeset
218 <!-- do nothing here -->
340633249b3d Uploaded
bgruening
parents:
diff changeset
219 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
220 <when value="set">
340633249b3d Uploaded
bgruening
parents:
diff changeset
221 <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;">
340633249b3d Uploaded
bgruening
parents:
diff changeset
222 <option value="default" selected="True">Don't set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
223 <option value="DO_NOTHING">DO_NOTHING</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
224 <option value="SET_Q_ZERO">SET_Q_ZERO</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
225 <option value="SET_Q_ZERO_BASE_N">SET_Q_ZERO_BASE_N</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
226 <option value="REMOVE_REF_BIAS">REMOVE_REF_BIAS</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
227 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
228 <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;">
340633249b3d Uploaded
bgruening
parents:
diff changeset
229 <option value="default" selected="True">Don't set</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
230 <option value="THROW_EXCEPTION">THROW_EXCEPTION</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
231 <option value="LEAVE_READ_UNRECALIBRATED">LEAVE_READ_UNRECALIBRATED</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
232 <option value="PURGE_READ">PURGE_READ</option>
340633249b3d Uploaded
bgruening
parents:
diff changeset
233 </param>
340633249b3d Uploaded
bgruening
parents:
diff changeset
234 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
235 </conditional>
340633249b3d Uploaded
bgruening
parents:
diff changeset
236 <param name="window_size_nqs" type="integer" value="5" label="Window size used by MinimumNQSCovariate" help="window_size_nqs"/>
340633249b3d Uploaded
bgruening
parents:
diff changeset
237 <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;" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
238 </when>
340633249b3d Uploaded
bgruening
parents:
diff changeset
239 </conditional>
340633249b3d Uploaded
bgruening
parents:
diff changeset
240 </inputs>
340633249b3d Uploaded
bgruening
parents:
diff changeset
241 <outputs>
340633249b3d Uploaded
bgruening
parents:
diff changeset
242 <data format="gatk_report" name="output_recal" label="${tool.name} on ${on_string} (Covariate File)" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
243 <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
244 </outputs>
340633249b3d Uploaded
bgruening
parents:
diff changeset
245 <tests>
340633249b3d Uploaded
bgruening
parents:
diff changeset
246 <test>
340633249b3d Uploaded
bgruening
parents:
diff changeset
247 <param name="reference_source_selector" value="history" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
248 <param name="ref_file" value="phiX.fasta" ftype="fasta" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
249 <param name="input_bam" value="gatk/gatk_indel_realigner/gatk_indel_realigner_out_1.bam" ftype="bam" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
250 <param name="rod_bind_type_selector" value="dbsnp" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
251 <param name="input_rod" value="gatk/fake_phiX_variant_locations.bed" ftype="bed" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
252 <param name="standard_covs" value="True" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
253 <param name="covariates" value="ReadGroupCovariate,HomopolymerCovariate,MinimumNQSCovariate,PositionCovariate" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
254 <param name="gatk_param_type_selector" value="basic" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
255 <param name="analysis_param_type_selector" value="basic" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
256 <output name="output_recal" file="gatk/gatk_count_covariates/gatk_count_covariates_out_1.csv" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
257 <output name="output_log" file="gatk/gatk_count_covariates/gatk_count_covariates_out_1.log.contains" compare="contains" />
340633249b3d Uploaded
bgruening
parents:
diff changeset
258 </test>
340633249b3d Uploaded
bgruening
parents:
diff changeset
259 </tests>
340633249b3d Uploaded
bgruening
parents:
diff changeset
260 <help>
340633249b3d Uploaded
bgruening
parents:
diff changeset
261 .. class:: warningmark
340633249b3d Uploaded
bgruening
parents:
diff changeset
262
340633249b3d Uploaded
bgruening
parents:
diff changeset
263 "This calculation is critically dependent on being able to skip over known variant sites. Please provide a dbSNP ROD or a VCF file containing known sites of genetic variation."
340633249b3d Uploaded
bgruening
parents:
diff changeset
264 However, if you do not provide this file, the '--run_without_dbsnp_potentially_ruining_quality' flag will be automatically used, and the command will be allowed to run.
340633249b3d Uploaded
bgruening
parents:
diff changeset
265
340633249b3d Uploaded
bgruening
parents:
diff changeset
266 **What it does**
340633249b3d Uploaded
bgruening
parents:
diff changeset
267
340633249b3d Uploaded
bgruening
parents:
diff changeset
268 This walker is designed to work as the first pass in a two-pass processing step. It does a by-locus traversal operating only at sites that are not in dbSNP. We assume that all reference mismatches we see are therefore errors and indicative of poor base quality. This walker generates tables based on various user-specified covariates (such as read group, reported quality score, cycle, and dinucleotide) Since there is a large amount of data one can then calculate an empirical probability of error given the particular covariates seen at this site, where p(error) = num mismatches / num observations The output file is a CSV list of (the several covariate values, num observations, num mismatches, empirical quality score) The first non-comment line of the output file gives the name of the covariates that were used for this calculation. Note: ReadGroupCovariate and QualityScoreCovariate are required covariates and will be added for the user regardless of whether or not they were specified Note: This walker is designed to be used in conjunction with TableRecalibrationWalker.
340633249b3d Uploaded
bgruening
parents:
diff changeset
269
340633249b3d Uploaded
bgruening
parents:
diff changeset
270 For more information on base quality score recalibration using the GATK, see this `tool specific page &lt;http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_bqsr_BaseRecalibrator.html&gt;`_.
340633249b3d Uploaded
bgruening
parents:
diff changeset
271
340633249b3d Uploaded
bgruening
parents:
diff changeset
272 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;`_.
340633249b3d Uploaded
bgruening
parents:
diff changeset
273
340633249b3d Uploaded
bgruening
parents:
diff changeset
274 If you encounter errors, please view the `GATK FAQ &lt;http://www.broadinstitute.org/gatk/guide/topic?name=faqs&gt;`_.
340633249b3d Uploaded
bgruening
parents:
diff changeset
275
340633249b3d Uploaded
bgruening
parents:
diff changeset
276 ------
340633249b3d Uploaded
bgruening
parents:
diff changeset
277
340633249b3d Uploaded
bgruening
parents:
diff changeset
278 **Inputs**
340633249b3d Uploaded
bgruening
parents:
diff changeset
279
340633249b3d Uploaded
bgruening
parents:
diff changeset
280 GenomeAnalysisTK: BaseRecalibrator accepts an aligned BAM input file.
340633249b3d Uploaded
bgruening
parents:
diff changeset
281
340633249b3d Uploaded
bgruening
parents:
diff changeset
282
340633249b3d Uploaded
bgruening
parents:
diff changeset
283 **Outputs**
340633249b3d Uploaded
bgruening
parents:
diff changeset
284
340633249b3d Uploaded
bgruening
parents:
diff changeset
285 The output is in CSV format.
340633249b3d Uploaded
bgruening
parents:
diff changeset
286
340633249b3d Uploaded
bgruening
parents:
diff changeset
287
340633249b3d Uploaded
bgruening
parents:
diff changeset
288 Go `here &lt;http://www.broadinstitute.org/gatk/guide/topic?name=intro&gt;`_ for details on GATK file formats.
340633249b3d Uploaded
bgruening
parents:
diff changeset
289
340633249b3d Uploaded
bgruening
parents:
diff changeset
290 -------
340633249b3d Uploaded
bgruening
parents:
diff changeset
291
340633249b3d Uploaded
bgruening
parents:
diff changeset
292 **Settings**::
340633249b3d Uploaded
bgruening
parents:
diff changeset
293
340633249b3d Uploaded
bgruening
parents:
diff changeset
294
340633249b3d Uploaded
bgruening
parents:
diff changeset
295 default_read_group If a read has no read group then default to the provided String.
340633249b3d Uploaded
bgruening
parents:
diff changeset
296 default_platform If a read has no platform then default to the provided String. Valid options are illumina, 454, and solid.
340633249b3d Uploaded
bgruening
parents:
diff changeset
297 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.
340633249b3d Uploaded
bgruening
parents:
diff changeset
298 force_platform If provided, the platform of EVERY read will be forced to be the provided String. Valid options are illumina, 454, and solid.
340633249b3d Uploaded
bgruening
parents:
diff changeset
299 window_size_nqs The window size used by MinimumNQSCovariate for its calculation
340633249b3d Uploaded
bgruening
parents:
diff changeset
300 homopolymer_nback The number of previous bases to look at in HomopolymerCovariate
340633249b3d Uploaded
bgruening
parents:
diff changeset
301 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
340633249b3d Uploaded
bgruening
parents:
diff changeset
302 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)
340633249b3d Uploaded
bgruening
parents:
diff changeset
303 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)
340633249b3d Uploaded
bgruening
parents:
diff changeset
304 recal_file Filename for the input covariates table recalibration .csv file
340633249b3d Uploaded
bgruening
parents:
diff changeset
305 out The output CSV file
340633249b3d Uploaded
bgruening
parents:
diff changeset
306 standard_covs Use the standard set of covariates in addition to the ones listed using the -cov argument
340633249b3d Uploaded
bgruening
parents:
diff changeset
307 run_without_dbsnp_potentially_ruining_quality If specified, allows the recalibrator to be used without a dbsnp rod. Very unsafe and for expert users only.
340633249b3d Uploaded
bgruening
parents:
diff changeset
308
340633249b3d Uploaded
bgruening
parents:
diff changeset
309 @CITATION_SECTION@
340633249b3d Uploaded
bgruening
parents:
diff changeset
310 </help>
340633249b3d Uploaded
bgruening
parents:
diff changeset
311 </tool>