Mercurial > repos > yating-l > ucsc_blat

comparison blat.xml @ 6:250a7f2b1331 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 630d4a316c3b72b0d92b32e8b37303d247a41190
author iuc
date Thu, 31 May 2018 11:25:53 -0400
parents dede60f744f4
children 10f3a5b6f2ac
comparison
equal deleted inserted replaced
5:dede60f744f4 6:250a7f2b1331
1 <?xml version="1.0"?> 1 <?xml version="1.0"?>
2 <tool id="ucsc_blat" name="UCSC BLAT Alignment Tool" version="35-0"> 2 <tool id="ucsc_blat" name="UCSC BLAT Alignment Tool" version="36">
3 <description>BLAST-like sequence alignment tool</description> 3 <description>BLAST-like sequence alignment tool</description>
4 <requirements> 4 <requirements>
5 <requirement type="package" version="35">blat</requirement> 5 <requirement type="package" version="36">blat</requirement>
6 </requirements> 6 </requirements>
7 <command detect_errors="exit_code"><![CDATA[ 7 <command detect_errors="exit_code"><![CDATA[
8 #if str($reference_source.reference_source_selector) == "history": 8 #if str($reference_source.reference_source_selector) == "history":
9 ## blat depends on file extension 9 ## blat depends on file extension
10 #if $reference_source.database.is_of_type("fasta"): 10 #if $reference_source.database.is_of_type("fasta"):
18 ln -s '${reference_source.database}' '${reference_fasta_filename}' && 18 ln -s '${reference_source.database}' '${reference_fasta_filename}' &&
19 #else: 19 #else:
20 #set $reference_fasta_filename = str($reference_source.database.fields.path) 20 #set $reference_fasta_filename = str($reference_source.database.fields.path)
21 #end if 21 #end if
22 22
23 blat 23 blat
24 #if $noHead == "yes"
25 -noHead
26 #end if
27 24
28 -q=$query_type 25 -q=$query_type
29 -t=$database_type 26 -t=$database_type
27
28 ${oneOff}
29
30 #if str($minScore)
31 -minScore=${minScore}
32 #end if
33
34 -maxGap=${maxGap}
35
36 #if str($repMatch)
37 -repMatch=${repMatch}
38 #end if
30 39
31 #if $mask_type.mask == "file.out": 40 #if $mask_type.mask == "file.out":
32 -mask='${mask_type.mask_file}' 41 -mask='${mask_type.mask_file}'
33 #else: 42 #else:
34 -mask=$mask_type.mask 43 -mask=$mask_type.mask
35 #end if 44 #end if
36 45
46 #if str($dots)
47 -dots=${dots}
48 #end if
49
50 ${trimT}
51 ${noTrimA}
52 ${trimHardA}
53 ${fastMap}
54 ${fine}
55
56 #if str($maxIntron)
57 -maxIntron=${maxIntron}
58 #end if
59
60 ${extendThroughN}
61
37 '$reference_fasta_filename' 62 '$reference_fasta_filename'
38 '${query}' 63 '${query}'
39 64
40 output 65 -out=${out}
41 && sort -k 10,10 -k 12,12n output > '${output_sorted}' 66
67 '${output}'
42 ]]></command> 68 ]]></command>
43 <inputs> 69 <inputs>
44 <conditional name="reference_source"> 70 <conditional name="reference_source">
45 <param name="reference_source_selector" type="select" label="Choose the source for the database"> 71 <param name="reference_source_selector" type="select" label="Choose the source for the database">
46 <option value="cached">Locally cached</option> 72 <option value="cached">Locally cached</option>
69 <option value="rna">rna - RNA sequence</option> 95 <option value="rna">rna - RNA sequence</option>
70 <option value="prot">prot - protein sequence</option> 96 <option value="prot">prot - protein sequence</option>
71 <option value="dnax">dnax - DNA sequence translated in six frames to protein</option> 97 <option value="dnax">dnax - DNA sequence translated in six frames to protein</option>
72 <option value="rnax" selected="true">rnax - DNA sequence translated in three frames to protein</option> 98 <option value="rnax" selected="true">rnax - DNA sequence translated in three frames to protein</option>
73 </param> 99 </param>
74 <param type="select" name="noHead" format="text" label="Suppresses .psl header (so it's just a tab-separated file)." > 100 <param argument="-oneOff" type="boolean" truevalue="-oneOff=1" falsevalue="" label="If set, this allows one mismatch in tile and still triggers an alignments" />
75 <option value="yes" selected="true">Yes, suppresses .psl header</option> 101 <param argument="-minScore" type="integer" value="30" label="Minimum score" help="It is the matches minus the mismatches minus some sort of gap penalty" />
76 <option value="no">No, do not suppresses .psl header</option> 102 <param argument="-maxGap" type="integer" value="2" min="0" max="3" label="Maximum gap between tiles in a clump" help="Usually set from 0 to 3. Only relevant for minMatch > 1" />
77 </param> 103 <param argument="-repMatch" type="integer" value="" optional="true" label="Number of repetitions of a tile allowed before it is marked as overused" help="Typically this is 256 for tileSize 12, 1024 for tileSize 11, 4096 for tileSize 10. Also affected by stepSize. When stepSize is halved repMatch is doubled to compensate" />
78 <conditional name="mask_type"> 104 <conditional name="mask_type">
79 <param name="mask" type="select" label="Mask out repeats" help="Alignments won't be started in masked region 105 <param name="mask" type="select" label="Mask out repeats" help="Alignments won't be started in masked region
80 but may extend through it in nucleotide searches. Masked areas 106 but may extend through it in nucleotide searches. Masked areas
81 are ignored entirely in protein or translated searches. Default is lower" argument="-mask"> 107 are ignored entirely in protein or translated searches. Default is lower" argument="-mask">
82 <option value="lower" selected="true">lower - mask out lower-cased sequence</option> 108 <option value="lower" selected="true">lower - mask out lower-cased sequence</option>
87 <when value="lower" /> 113 <when value="lower" />
88 <when value="upper" /> 114 <when value="upper" />
89 <when value="out" /> 115 <when value="out" />
90 <when value="file.out"> 116 <when value="file.out">
91 <param name="mask_file" type="data" format="txt" label="RepeatMasker file.out" /> 117 <param name="mask_file" type="data" format="txt" label="RepeatMasker file.out" />
92 </when> 118 </when>
93 </conditional> 119 </conditional>
120 <param argument="-dots" type="integer" value="" optional="true" label="Output a dot every N sequences in log" help="Dots show program's progress" />
121 <param argument="-trimT" type="boolean" truevalue="-trimT" falsevalue="" label="Trim leading poly-T" />
122 <param argument="-noTrimA" type="boolean" truevalue="-noTrimA" falsevalue="" label="Don't trim trailing poly-A" />
123 <param argument="-trimHardA" type="boolean" truevalue="-trimHardA" falsevalue="" label="Remove poly-A tail from qSize and alignments in .psl output" />
124 <param argument="-fastMap" type="boolean" truevalue="-fastMap" falsevalue="" label="Run for fast DNA/DNA remapping" help="It does not allow introns and require high %ID. Query sizes must not exceed 5000" />
125 <param argument="-fine" type="boolean" truevalue="-fine" falsevalue="" label="Refine search for small initial and terminal exons" help="For high-quality mRNAs. Not recommended for ESTs" />
126 <param argument="-maxIntron" type="integer" value="750000" optional="true" label="Maximum intron size" />
127 <param argument="-extendThroughN" type="boolean" truevalue="-extendThroughN" falsevalue="" label="Allow extension of alignment through large blocks of N's" />
128 <param name="out" type="select" label="Select output file format (-out)">
129 <option value="psl">Tab-separated format, no sequence (psl)</option>
130 <option value="psl -noHead">Tab-separated format, no sequence, no header (psl -noHead)</option>
131 <option value="axt">Blastz-associated axt format (axt)</option>
132 <option value="maf">Multiz-associated maf format (maf)</option>
133 <option value="sim4">Similar to sim4 format (sim4)</option>
134 <option value="wublast">Similar to WU-BLAST format (wublast)</option>
135 <option value="blast">Similar to NCBI BLAST format (blast)</option>
136 <option value="blast8">NCBI BLAST tabular format (blast8)</option>
137 <option value="blast9">NCBI BLAST tabular format with comments (blast9)</option>
138 </param>
94 </inputs> 139 </inputs>
95 <outputs> 140 <outputs>
96 <data format="psl" name="output_sorted" label="${tool.name} on ${on_string}"></data> 141 <data name="output" format="tabular" label="${tool.name} on ${on_string}">
142 <change_format>
143 <when input="out" value="psl" format="psl" />
144 <when input="out" value="axt" format="axt" />
145 <when input="out" value="maf" format="maf" />
146 <when input="out" value="sim4" format="txt" />
147 <when input="out" value="wublast" format="txt" />
148 <when input="out" value="blast" format="txt" />
149 </change_format>
150 </data>
97 </outputs> 151 </outputs>
98 <tests> 152 <tests>
99 <!-- test on query of GenBank RefSeq records for Gallus gallus and database of Amazona vittata --> 153 <!-- test on query of GenBank RefSeq records for Gallus gallus and database of Amazona vittata -->
100 <test> 154 <test>
101 <param name="reference_source_selector" value="history" /> 155 <param name="reference_source_selector" value="history" />
102 <param name="database" value="amaVit1_Gallus/amaVit1.fa" /> 156 <param name="database" value="amaVit1_Gallus/amaVit1.fa" />
103 <param name="query" value="amaVit1_Gallus/Gallus_gallus_RefSeq.fa" /> 157 <param name="query" value="amaVit1_Gallus/Gallus_gallus_RefSeq.fa" />
104 <param name="database_type" value="dnax" /> 158 <param name="database_type" value="dnax" />
105 <param name="query_type" value="rnax" /> 159 <param name="query_type" value="rnax" />
106 <param name="noHead" value="yes" />
107 <param name="mask" value="lower" /> 160 <param name="mask" value="lower" />
108 <output name="output_sorted" value="amaVit1_Gallus/amaVit1_Gallus_gallus.psl" /> 161 <param name="out" value="psl -noHead" />
162 <output name="output" value="amaVit1_Gallus/amaVit1_Gallus_gallus_sorted.psl" sort="True"/>
109 </test> 163 </test>
110 <!-- test on query of partial mRNA of Drosophila melanogaster and the database of Drosophila biamipes dot chromosome --> 164 <!-- test on query of partial mRNA of Drosophila melanogaster and the database of Drosophila biamipes dot chromosome -->
111 <test> 165 <test>
112 <param name="reference_source_selector" value="history" /> 166 <param name="reference_source_selector" value="history" />
113 <param name="database" value="dbia3/dbia3.fa" /> 167 <param name="database" value="dbia3/dbia3.fa" />
114 <param name="query" value="dbia3/dmel-transcript.fa" /> 168 <param name="query" value="dbia3/dmel-transcript.fa" />
115 <param name="database_type" value="dnax" /> 169 <param name="database_type" value="dnax" />
116 <param name="query_type" value="rnax" /> 170 <param name="query_type" value="rnax" />
117 <param name="noHead" value="yes" />
118 <param name="mask" value="lower" /> 171 <param name="mask" value="lower" />
119 <output name="output_sorted" value="dbia3/dbia3.sorted.psl" /> 172 <param name="out" value="psl -noHead" />
173 <param name="maxIntron" value="" />
174 <output name="output" value="dbia3/dbia3.sorted.psl" sort="True"/>
120 </test> 175 </test>
121 <!-- test on the database masked by repeat masker --> 176 <!-- test on the database masked by repeat masker -->
122 <test> 177 <test>
123 <param name="reference_source_selector" value="history" /> 178 <param name="reference_source_selector" value="history" />
124 <param name="database" value="dbia3/dbia3_masked.2bit" /> 179 <param name="database" value="dbia3/dbia3_masked.2bit" />
125 <param name="query" value="dbia3/dmel-transcript.fa" /> 180 <param name="query" value="dbia3/dmel-transcript.fa" />
126 <param name="database_type" value="dnax" /> 181 <param name="database_type" value="dnax" />
127 <param name="query_type" value="rnax" /> 182 <param name="query_type" value="rnax" />
128 <param name="noHead" value="yes" /> 183 <param name="oneOff" value="false" />
184 <param name="minScore" value="30" />
185 <param name="maxGap" value="2" />
186 <param name="trimT" value="false" />
187 <param name="noTrimA" value="false" />
188 <param name="fine" value="false" />
189 <param name="maxIntron" value="750000" />
190 <param name="extendThroughN" value="false" />
129 <param name="mask" value="file.out" /> 191 <param name="mask" value="file.out" />
130 <param name="mask_file" value="dbia3/dbia3_RM.out" /> 192 <param name="mask_file" value="dbia3/dbia3_RM.out" />
131 <output name="output_sorted" value="dbia3/dbia3_masked.sorted.psl" /> 193 <param name="out" value="psl -noHead" />
194 <output name="output" value="dbia3/dbia3_masked.sorted.psl"/>
132 </test> 195 </test>
133 196
134 </tests> 197 </tests>
135 <help> 198 <help>
136 <![CDATA[ 199 <![CDATA[
137 BLAT 200 BLAT
138 ==== 201 ====
139 BLAT is a bioinformatics software a tool which performs rapid mRNA/DNA and cross-species protein alignments. 202 BLAT is a bioinformatics software a tool which performs rapid mRNA/DNA and cross-species protein alignments.
140 203
141 blat (version: v340)- Standalone blat sequence search command line tool. 204 blat (version: v36)- Standalone blat sequence search command line tool.
142 ------------------------------------------------------------------------- 205 -------------------------------------------------------------------------
143 206
144 usage: 207 usage:
145 ++++++ 208 ++++++
146 209
150 database and query are each either a .fa, .nib or .2bit file, 213 database and query are each either a .fa, .nib or .2bit file,
151 or a list of these files with one file name per line. 214 or a list of these files with one file name per line.
152 -ooc=11.ooc tells the program to load over-occurring 11-mers from 215 -ooc=11.ooc tells the program to load over-occurring 11-mers from
153 an external file. This will increase the speed 216 an external file. This will increase the speed
154 by a factor of 40 in many cases, but is not required. 217 by a factor of 40 in many cases, but is not required.
155 output.psl is the name of the output file. 218 output.psl is the name of the output file.
156 219
157 documentation: 220 documentation:
158 ++++++++++++++ 221 ++++++++++++++
159 222
160 See Blat documentation (http://genome.ucsc.edu/goldenPath/help/blatSpec.html) 223 See Blat documentation (http://genome.ucsc.edu/goldenPath/help/blatSpec.html)
161 224
162 Source code: 225 Source code:
163 ++++++++++++ 226 ++++++++++++
164 227
165 http://hgdownload.cse.ucsc.edu/admin/exe/ 228 http://hgdownload.cse.ucsc.edu/admin/exe/
166 229
167 ]]></help> 230 ]]></help>
168 <citations> 231 <citations>
169 <citation type="doi">10.1101/gr.229202</citation> 232 <citation type="doi">10.1101/gr.229202</citation>
170 </citations> 233 </citations>
171 </tool> 234 </tool>
172
173
174
175
176