Mercurial > repos > crs4 > bwa_mem

comparison bwa_mem.xml @ 1:ebb02ba5987c draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Rewrite of param handling. interPairEnd param moved to "paired" section. Add param for '-a' option. Remove basic parallelism tag, which does not work with multiple inputs (thanks Bjoern Gruening for the notice). Simplify Python code.
author crs4
date Fri, 21 Mar 2014 12:56:15 -0400
parents 6820983ba5d5
children
comparison
equal deleted inserted replaced
0:6820983ba5d5 1:ebb02ba5987c
1 <tool id="bwa_mem" name="Map with BWA-MEM" version="0.7.7"> 1 <tool id="bwa_mem" name="Map with BWA-MEM" version="0.8.0">
2 <requirements> 2 <requirements>
3 <requirement type="package" version="0.7.7">bwa</requirement> 3 <requirement type="package" version="0.7.7">bwa</requirement>
4 </requirements> 4 </requirements>
5 <description></description> 5 <description></description>
6 <parallelism method="basic"></parallelism>
7 <version_command>bwa 2&gt;&amp;1 | grep "Version: " | sed -e 's/Version: //'</version_command> 6 <version_command>bwa 2&gt;&amp;1 | grep "Version: " | sed -e 's/Version: //'</version_command>
8 <command interpreter="python"> 7 <command interpreter="python">
9 bwa_mem.py 8 bwa_mem.py
10 --threads="\${GALAXY_SLOTS:-1}" 9 --threads="\${GALAXY_SLOTS:-1}"
11 --fileSource="${genomeSource.refGenomeSource}" 10 --fileSource="${genomeSource.refGenomeSource}"
12 #if $genomeSource.refGenomeSource == "history": 11 #if $genomeSource.refGenomeSource == "history"
13 ##build index on the fly 12 ##build index on the fly
14 --ref="${genomeSource.ownFile}" 13 --ref="${genomeSource.ownFile}"
15 --dbkey="${dbkey}" 14 --dbkey="${dbkey}"
16 #else: 15 #else
17 ##use precomputed indexes 16 ##use precomputed indexes
18 --ref="${genomeSource.indices.fields.path}" 17 --ref="${genomeSource.indices.fields.path}"
19 #end if 18 #end if
20 19
21 ## input file(s) 20 ## input file(s)
22 --fastq="${paired.fastq}" 21 --fastq="${paired.fastq}"
23 #if $paired.sPaired == "paired": 22 #if $paired.sPaired == "single"
23 #if $paired.interPairEnd
24 --interPairEnd
25 #end if
26 #else
24 --rfastq="${paired.rfastq}" 27 --rfastq="${paired.rfastq}"
25 #end if 28 #end if
26 29
27 ## output file 30 ## output file
28 --output="${output}" 31 --output="${output}"
29 32
30 ## run parameters 33 ## run parameters
31 --genAlignType="${paired.sPaired}" 34 --genAlignType="${paired.sPaired}"
32 --params="${params.source_select}" 35 --params="${params.source_select}"
33 #if $params.source_select != "pre_set": 36 #if $params.source_select != "pre_set"
34 --minEditDistSeed="${params.minEditDistSeed}" 37 #if str($params.minEditDistSeed)
35 --bandWidth="${params.bandWidth}" 38 --minSeedLength ${params.minEditDistSeed}
36 --offDiagonal="${params.offDiagonal}" 39 #end if
37 --internalSeeds="${params.internalSeeds}" 40 #if str($params.bandWidth)
38 --seedsOccurrence="${params.seedsOccurrence}" 41 --bandWidth ${params.bandWidth}
39 --mateRescue="${params.mateRescue}" 42 #end if
40 --skipPairing="${params.skipPairing}" 43 #if str($params.offDiagonal)
41 --seqMatch="${params.seqMatch}" 44 --offDiagonal ${params.offDiagonal}
42 --mismatch="${params.mismatch}" 45 #end if
43 --gapOpen="${params.gapOpen}" 46 #if str($params.internalSeeds)
44 --gapExtension="${params.gapExtension}" 47 --internalSeeds ${params.internalSeeds}
45 --clipping="${params.clipping}" 48 #end if
46 --unpairedReadpair="${params.unpairedReadpair}" 49 #if str($params.seedsOccurrence)
47 --interPairEnd="${params.interPairEnd}" 50 --seedsOccurrence ${params.seedsOccurrence}
48 --minScore="${params.minScore}" 51 #end if
49 --mark="${params.mark}" 52 #if $params.mateRescue
53 --mateRescue
54 #end if
55 #if $params.skipPairing
56 --skipPairing
57 #end if
58 #if str($params.seqMatch)
59 --seqMatch ${params.seqMatch}
60 #end if
61 #if str($params.mismatch)
62 --mismatch ${params.mismatch}
63 #end if
64 #if str($params.gapOpen)
65 --gapOpen ${params.gapOpen}
66 #end if
67 #if str($params.gapExtension)
68 --gapExtension ${params.gapExtension}
69 #end if
70 #if $params.clipping
71 --clipping "${params.clipping}"
72 #end if
73 #if str($params.unpairedReadpair)
74 --unpairedReadpair ${params.unpairedReadpair}
75 #end if
76 #if str($params.minScore)
77 --minScore ${params.minScore}
78 #end if
79 #if $params.outputAll
80 --outputAll
81 #end if
82 #if $params.mark
83 --mark
84 #end if
50 85
51 #if $params.readGroup.specReadGroup == "yes" 86 #if $params.readGroup.specReadGroup == "yes"
52 --rgid="${params.readGroup.rgid}" 87 --rgid="${params.readGroup.rgid}"
53 --rgsm="${params.readGroup.rgsm}" 88 --rgsm="${params.readGroup.rgsm}"
54 --rgpl="${params.readGroup.rgpl}" 89 --rgpl ${params.readGroup.rgpl}
55 --rglb="${params.readGroup.rglb}" 90 --rglb="${params.readGroup.rglb}"
56 --rgpu="${params.readGroup.rgpu}" 91 --rgpu="${params.readGroup.rgpu}"
57 --rgcn="${params.readGroup.rgcn}" 92 --rgcn="${params.readGroup.rgcn}"
58 --rgds="${params.readGroup.rgds}" 93 --rgds="${params.readGroup.rgds}"
59 --rgdt="${params.readGroup.rgdt}" 94 --rgdt="${params.readGroup.rgdt}"
63 --rgpi="${params.readGroup.rgpi}" 98 --rgpi="${params.readGroup.rgpi}"
64 #end if 99 #end if
65 #end if 100 #end if
66 101
67 ## suppress output SAM header 102 ## suppress output SAM header
68 --suppressHeader="${suppressHeader}" 103 #if $suppressHeader
104 --suppressHeader
105 #end if
69 </command> 106 </command>
70 107
71 <inputs> 108 <inputs>
72 <conditional name="genomeSource"> 109 <conditional name="genomeSource">
73 <param name="refGenomeSource" type="select" label="Will you select a reference genome from your history or use a built-in index?"> 110 <param name="refGenomeSource" type="select" label="Will you select a reference genome from your history or use a built-in index?">
86 <param name="ownFile" type="data" format="fasta" metadata_name="dbkey" label="Select a reference from history" /> 123 <param name="ownFile" type="data" format="fasta" metadata_name="dbkey" label="Select a reference from history" />
87 </when> 124 </when>
88 </conditional> 125 </conditional>
89 <conditional name="paired"> 126 <conditional name="paired">
90 <param name="sPaired" type="select" label="Is this library mate-paired?"> 127 <param name="sPaired" type="select" label="Is this library mate-paired?">
91 <option value="single">Single-end</option> 128 <option value="single">Single-end or interleaved paired-end</option>
92 <option value="paired">Paired-end</option> 129 <option value="paired">Paired-end</option>
93 </param> 130 </param>
94 <when value="single"> 131 <when value="single">
95 <param name="fastq" type="data" format="fastqsanger,fastqillumina" label="FASTQ file" help="FASTQ with either Sanger-scaled quality values (fastqsanger) or Illumina-scaled quality values (fastqillumina)" /> 132 <param name="fastq" type="data" format="fastqsanger,fastqillumina" label="FASTQ file" help="FASTQ with either Sanger-scaled quality values (fastqsanger) or Illumina-scaled quality values (fastqillumina)" />
133 <param name="interPairEnd" type="boolean" checked="false" label="FASTQ file consists of interleaved paired-end sequences (-p)" />
96 </when> 134 </when>
97 <when value="paired"> 135 <when value="paired">
98 <param name="fastq" type="data" format="fastqsanger,fastqillumina" label="Forward FASTQ file" help="FASTQ with either Sanger-scaled quality values (fastqsanger) or Illumina-scaled quality values (fastqillumina)" /> 136 <param name="fastq" type="data" format="fastqsanger,fastqillumina" label="Forward FASTQ file" help="FASTQ with either Sanger-scaled quality values (fastqsanger) or Illumina-scaled quality values (fastqillumina)" />
99 <param name="rfastq" type="data" format="fastqsanger,fastqillumina" label="Reverse FASTQ file" help="FASTQ with either Sanger-scaled quality values (fastqsanger) or Illumina-scaled quality values (fastqillumina)" /> 137 <param name="rfastq" type="data" format="fastqsanger,fastqillumina" label="Reverse FASTQ file" help="FASTQ with either Sanger-scaled quality values (fastqsanger) or Illumina-scaled quality values (fastqillumina)" />
100 </when> 138 </when>
104 <option value="pre_set">Commonly Used</option> 142 <option value="pre_set">Commonly Used</option>
105 <option value="full">Full Parameter List</option> 143 <option value="full">Full Parameter List</option>
106 </param> 144 </param>
107 <when value="pre_set" /> 145 <when value="pre_set" />
108 <when value="full"> 146 <when value="full">
109 <param name="minEditDistSeed" type="integer" value="19" label="Minimum seed length" /> 147 <param name="minEditDistSeed" type="integer" value="19" optional="true" label="Minimum seed length (-k)" />
110 <param name="bandWidth" type="integer" value="100" label="Band width for banded alignment" /> 148 <param name="bandWidth" type="integer" value="100" optional="true" label="Band width for banded alignment (-w)" />
111 <param name="offDiagonal" type="integer" value="100" label="off-diagonal X-dropoff" /> 149 <param name="offDiagonal" type="integer" value="100" optional="true" label="Off-diagonal X-dropoff (-d)" />
112 <param name="internalSeeds" type="float" value="1.5" label="look for internal seeds inside a seed longer than {-k} * FLOAT [1.5]" /> 150 <param name="internalSeeds" type="float" value="1.5" optional="true" label="Look for internal seeds inside a seed longer than the minimum seed length times this value (-r)" help="This is a key heuristic parameter for tuning the performance. Larger value yields fewer seeds, which leads to faster alignment speed but lower accuracy" />
113 <param name="seedsOccurrence" type="integer" value="10000" label="skip seeds with more than INT occurrences" /> 151 <param name="seedsOccurrence" type="integer" value="10000" optional="true" label="Skip seeds with more occurrences than this value (-c)" />
114 <param name="mateRescue" type="boolean" truevalue="True" falsevalue="False" checked="False" label="skip seeds with more than INT occurrences" /> 152 <param name="mateRescue" type="boolean" checked="false" label="Skip mate rescue (-S)" />
115 <param name="skipPairing" type="boolean" truevalue="True" falsevalue="False" checked="False" label="skpe pairing, mate rescue performed unless -S also in use" /> 153 <param name="skipPairing" type="boolean" truevalue="True" falsevalue="False" checked="False" label="Skip pairing (-P)" help="In the paired-end mode, perform Smith-Waterman to rescue missing hits only, but do not try to find hits that fit a proper pair" />
116 <param name="seqMatch" type="integer" value="1" label="score of a sequence match" /> 154 <param name="seqMatch" type="integer" value="1" optional="true" label="Score for a sequence match (-A)" />
117 <param name="mismatch" type="integer" value="4" label="penalty for a mismatch" /> 155 <param name="mismatch" type="integer" value="4" optional="true" label="Penalty for a mismatch (-B)" />
118 <param name="gapOpen" type="integer" value="6" label="gap open penalty" /> 156 <param name="gapOpen" type="integer" value="6" optional="true" label="Gap open penalty (-O)" />
119 <param name="gapExtension" type="text" value="None" label="gap extension penalty; a gap of size k cost {-O} + {-E}*k [1]" /> 157 <param name="gapExtension" type="integer" value="1" optional="true" label="Gap extension penalty (-E)" help="A gap of length k costs {gap open penalty} + k*{this value}" />
120 <param name="clipping" type="integer" value="5" label="penalty for clipping" /> 158 <param name="clipping" type="text" value="5" optional="true" label="Penalty for clipping (-L)" help="When performing Smith-Waterman extension, BWA-MEM keeps track of the best score reaching the end of query. If this score is larger than the best Smith-Waterman score minus the clipping penalty, clipping will not be applied. Note that in this case, the SAM AS tag reports the best Smith-Waterman score; clipping penalty is not deduced. If two comma-separated numbers are provided, the first is for 5'-end clipping and second for 3'-end clipping">
121 <param name="unpairedReadpair" type="integer" value="17" label="penalty for an unpaired read pair" /> 159 <validator type="regex" message="Invalid clipping, format is INT[,INT]">\d+(,\d+)?$</validator>
122 <param name="interPairEnd" type="boolean" truevalue="True" falsevalue="False" checked="False" label="first query file consists of interleaved paired-end sequences" /> 160 </param>
123 <param name="minScore" type="integer" value="30" label="minimum score to output" /> 161 <param name="unpairedReadpair" type="integer" value="17" optional="true" label="Penalty for an unpaired read pair (-U)" help="" />
124 <param name="mark" type="boolean" truevalue="True" falsevalue="False" checked="False" label="mark shorter split hits as secondary (for Picard/GATK compatibility)" /> 162 <param name="minScore" type="integer" value="30" optional="true" label="Minimum score to output (-T)" />
125 163 <param name="outputAll" type="boolean" checked="false" label="Output all found alignments for single-end or unpaired paired-end reads (-a)" help="These alignments will be flagged as secondary alignments" />
164 <param name="mark" type="boolean" checked="false" label="Mark shorter split hits as secondary (-M)" help="For Picard/GATK compatibility" />
126 <conditional name="readGroup"> 165 <conditional name="readGroup">
127 <param name="specReadGroup" type="select" label="Specify the read group for this file? (samse/sampe -r)"> 166 <param name="specReadGroup" type="select" label="Specify the read group for this file? (-R)">
128 <option value="yes">Yes</option> 167 <option value="yes">Yes</option>
129 <option value="no" selected="True">No</option> 168 <option value="no" selected="True">No</option>
130 </param> 169 </param>
131 <when value="yes"> 170 <when value="yes">
132 <param name="rgid" type="text" size="25" label="[Essential]Read group identifier (ID). Each @RG line must have a unique ID. The value of ID is used in the RG 171 <param name="rgid" type="text" size="25" label="Read group identifier (ID). Each @RG line must have a unique ID. The value of ID is used in the RG tags of alignment records. Must be unique among all read groups in header section." help="Required if RG specified. Read group IDs may be modified when merging SAM files in order to handle collisions.">
133 tags of alignment records. Must be unique among all read groups in header section." help="Required if RG specified. Read group 172 <validator type="empty_field" />
134 IDs may be modified when merging SAM files in order to handle collisions." /> 173 </param>
135 <param name="rgpl" type="text" size="25" label="[Essential]Platform/technology used to produce the reads (PL)" help="Required if RG specified. Valid values : CAPILLARY, LS454, ILLUMINA, 174 <param name="rgpl" type="select" label="Platform/technology used to produce the reads (PL)">
136 SOLID, HELICOS, IONTORRENT and PACBIO" /> 175 <option value="CAPILLARY">CAPILLARY</option>
137 <param name="rglb" type="text" size="25" label="[Essential]Library name (LB)" help="Required if RG specified" /> 176 <option value="LS454">LS454</option>
138 <param name="rgsm" type="text" size="25" label="[Essential]Sample (SM)" help="Required if RG specified. Use pool name where a pool is being sequenced" /> 177 <option value="ILLUMINA">ILLUMINA</option>
178 <option value="SOLID">SOLID</option>
179 <option value="HELICOS">HELICOS</option>
180 <option value="IONTORRENT">IONTORRENT</option>
181 <option value="PACBIO">PACBIO</option>
182 </param>
183 <param name="rglb" type="text" size="25" label="Library name (LB)" help="Required if RG specified">
184 <validator type="empty_field" />
185 </param>
186 <param name="rgsm" type="text" size="25" label="Sample (SM)" help="Required if RG specified. Use pool name where a pool is being sequenced">
187 <validator type="empty_field" />
188 </param>
139 <param name="rgpu" type="text" size="25" label="Platform unit (PU)" help="Optional. Unique identifier (e.g. flowcell-barcode.lane for Illumina or slide for SOLiD)" /> 189 <param name="rgpu" type="text" size="25" label="Platform unit (PU)" help="Optional. Unique identifier (e.g. flowcell-barcode.lane for Illumina or slide for SOLiD)" />
140 <param name="rgcn" type="text" size="25" label="Sequencing center that produced the read (CN)" help="Optional" /> 190 <param name="rgcn" type="text" size="25" label="Sequencing center that produced the read (CN)" help="Optional" />
141 <param name="rgds" type="text" size="25" label="Description (DS)" help="Optional" /> 191 <param name="rgds" type="text" size="25" label="Description (DS)" help="Optional" />
142 <param name="rgdt" type="text" size="25" label="Date that run was produced (DT)" help="Optional. ISO8601 format date or date/time, like YYYY-MM-DD" /> 192 <param name="rgdt" type="text" size="25" label="Date that run was produced (DT)" help="Optional. ISO8601 format date or date/time, like YYYY-MM-DD" />
143 <param name="rgfo" type="text" size="25" label="Flow order (FO). The array of nucleotide bases that correspond to the nucleotides used for each 193 <param name="rgfo" type="text" size="25" optional="true" label="Flow order (FO). The array of nucleotide bases that correspond to the nucleotides used for each flow of each read" help="Optional. Multi-base flows are encoded in IUPAC format, and non-nucleotide flows by various other characters. Format: /\*|[ACMGRSVTWYHKDBN]+/">
144 flow of each read." help="Optional. Multi-base flows are encoded in IUPAC format, and non-nucleotide flows by 194 <validator type="regex">\*|[ACMGRSVTWYHKDBN]+$</validator>
145 various other characters. Format : /\*|[ACMGRSVTWYHKDBN]+/" /> 195 </param>
146 <param name="rgks" type="text" size="25" label="The array of nucleotide bases that correspond to the key sequence of each read (KS)" help="Optional" /> 196 <param name="rgks" type="text" size="25" label="The array of nucleotide bases that correspond to the key sequence of each read (KS)" help="Optional" />
147 <param name="rgpg" type="text" size="25" label="Programs used for processing the read group (PG)" help="Optional" /> 197 <param name="rgpg" type="text" size="25" label="Programs used for processing the read group (PG)" help="Optional" />
148 <param name="rgpi" type="text" size="25" label="Predicted median insert size (PI)" help="Optional" /> 198 <param name="rgpi" type="text" size="25" label="Predicted median insert size (PI)" help="Optional" />
149 </when> 199 </when>
150 <when value="no" /> 200 <when value="no" />
151 </conditional> 201 </conditional>
152 </when> 202 </when>
153 </conditional> 203 </conditional>
154 <param name="suppressHeader" type="boolean" truevalue="true" falsevalue="false" checked="False" label="Suppress the header in the output SAM file" help="BWA produces SAM with several lines of header information" /> 204 <param name="suppressHeader" type="boolean" checked="false" label="Suppress the header in the output SAM file" help="BWA produces SAM with several lines of header information" />
155 </inputs> 205 </inputs>
156 206
157 <outputs> 207 <outputs>
158 <data format="sam" name="output" label="${tool.name} on ${on_string}: mapped reads"> 208 <data format="sam" name="output" label="${tool.name} on ${on_string}: mapped reads">
159 <actions> 209 <actions>