|
0
|
1 <tool id="blat" name="Align sequences with BLAT" version="1.0.0">
|
|
|
2 <!-- Additional info: wrapper compatible with versions ..... -->
|
|
|
3 <description>
|
|
|
4 to a reference genome
|
|
|
5 </description>
|
|
|
6
|
|
|
7 <version_command/>
|
|
|
8
|
|
|
9 <requirements>
|
|
|
10 <requirement type="package" version="1.0.0">blat</requirement>
|
|
|
11 </requirements>
|
|
|
12
|
|
|
13 <command interpreter="perl">
|
|
|
14 ## it is recommended that you write a wrapper for your tool
|
|
|
15 ## and pass all parameters to that tool, which parses them.
|
|
|
16 blat_wrapper.pl $configfile
|
|
|
17
|
|
|
18 </command>
|
|
|
19
|
|
|
20 <inputs>
|
|
|
21 <param format="fasta" name="input" type="data" label="Query fasta sequences" />
|
|
|
22 <param name="q" type="select" label="Process the query as">
|
|
|
23 <option value="dna" selected="True">DNA</option>
|
|
|
24 <option value="rna">RNA</option>
|
|
|
25 <option value="prot">Protein</option>
|
|
|
26 <option value="dnax">DNA sequence translated in six frames to protein</option>
|
|
|
27 <option value="rnax">RNA sequence translated in six frames to protein</option>
|
|
|
28 </param>
|
|
|
29 <conditional name="refGenomeSource">
|
|
|
30 <param name="genomeSource" type="select" label="Will you select a sequence database from your history or use a built-in index?" help="Built-ins were indexed using default options">
|
|
|
31 <option value="indexed">Use a built-in index</option>
|
|
|
32 <option value="history" selected="True">Use one from the history</option>
|
|
|
33 </param>
|
|
|
34 <when value="indexed">
|
|
|
35 <param name="index" type="select" label="Select a reference genome" help="If your genome of interest is not listed, contact the Galaxy team">
|
|
|
36 <options from_data_table="twobit.loc">
|
|
|
37 <filter type="sort_by" column="1"/>
|
|
|
38 <validator type="no_options" message="No indexes are available for the selected input dataset"/>
|
|
|
39 </options>
|
|
|
40 <param name="range" type="text" size="25" value="" label="Restrict the search space to this genomic range in the sequence database" help="Format = 'chr1:250000-260000' (i.e. name:startposition-endposition)" />
|
|
|
41
|
|
|
42 </param>
|
|
|
43 </when>
|
|
|
44 <when value="history">
|
|
|
45 <param name="ownFile" type="data" format="fasta" metadata_name="dbkey" label="Select the sequence database" />
|
|
|
46 </when> <!-- history -->
|
|
|
47 </conditional> <!-- refGenomeSource -->
|
|
|
48
|
|
|
49 <param name="t" type="select" label="Process the sequence database as">
|
|
|
50 <option value="dna" selected="True">DNA</option>
|
|
|
51 <option value="prot">Protein</option>
|
|
|
52 <option value="dnax">DNA sequence translated in six frames to protein</option>
|
|
|
53 </param>
|
|
|
54
|
|
|
55 <conditional name="advanced_params">
|
|
|
56 <param name="use" type="select" label="Show advanced parameters">
|
|
|
57 <option value="no" selected="True">No</option>
|
|
|
58 <option value="yes">Yes</option>
|
|
|
59 </param>
|
|
|
60 <when value="no" />
|
|
|
61 <when value="yes" >
|
|
|
62 <param name="tileSize" type="text" size="3" value="11" label="size of match that triggers an alignment" help="Usually between 8 and 12. Default is 11 for DNA and 5 for protein." />
|
|
|
63 <param name="stepSize" type="text" size="3" value="11" label="spacing between tiles" />
|
|
|
64 <param name="oneOff" type="boolean" truevalue="1" falsevalue="0" checked="False" label="Allowing one mismatch in tile?" />
|
|
|
65 <param name="minMatch" type="select" label="number of tile matches" >
|
|
|
66 <option value="1">1 (for proteins)</option>
|
|
|
67 <option value="2" selected="True">2</option>
|
|
|
68 <option value="3">3</option>
|
|
|
69 <option value="4">4</option>
|
|
|
70 </param>
|
|
|
71 <param name="minScore" type="text" size="3" value="30" label="minimum score" help="This is the matches minus the mismatches minus some sort of gap penalty. Default is 30" />
|
|
|
72 <param name="maxGap" type="select" label="size of maximum gap between tiles in a clump" help=" Usually set from 0 to 3. Default is 2. Only relevent for minMatch > 1">
|
|
|
73 <option value="0">0</option>
|
|
|
74 <option value="1">1</option>
|
|
|
75 <option value="2" selected="True">2</option>
|
|
|
76 <option value="3">3</option>
|
|
|
77 </param>
|
|
|
78 <param name="mask" type="select" label="Mask out repeats in reference sequence" help="Alignments won't be started in masked region but may extend through it in nucleotide searches. Masked areas are ignored entirely in protein or translated searches.">
|
|
|
79 <option value="lower">mask out lower cased sequence</option>
|
|
|
80 <option value="upper">mask out upper cased sequence</option>
|
|
|
81 </param>
|
|
|
82 <param name="qMask" type="select" label="Mask out repeats in query sequence" help="Alignments won't be started in masked region but may extend through it in nucleotide searches. Masked areas are ignored entirely in protein or translated searches.">
|
|
|
83 <option value="lower">mask out lower cased sequence</option>
|
|
|
84 <option value="upper">mask out upper cased sequence</option>
|
|
|
85 </param>
|
|
|
86 <param name="repeats" type="boolean" truevalue="yes" falsevalue="no" checked="False" label="Report matches in repeats separately?" />
|
|
|
87 <param name="trimT" type="boolean" truevalue="yes" falsevalue="no" checked="False" label="Trim leading poly-T" />
|
|
|
88 <param name="noTrimA" type="boolean" truevalue="no" falsevalue="yes" checked="True" label="Trim trailing poly-A" />
|
|
|
89 <param name="fine" type="boolean" truevalue="yes" falsevalue="no" checked="False" label="For high quality mRNAs look harder for small initial and terminal exons. Not recommended for ESTs" />
|
|
|
90 <param name="maxIntron" type="text" size="10" value="750000" label="Maximum intron size" />
|
|
|
91 <param name="extendThroughN" type="boolean" truevalue="yes" falsevalue="no" checked="False" label="Extent alignment through large blocks of N's? Not recommended for ESTs" />
|
|
|
92 </when>
|
|
|
93 </conditional>
|
|
|
94 </inputs>
|
|
|
95
|
|
|
96 <outputs>
|
|
|
97 <data format="tabular" name="output" label="${tool.name} on ${on_string}: alignments"/>
|
|
|
98 </outputs>
|
|
|
99
|
|
|
100 <configfiles>
|
|
|
101 <!-- this config file collects all parameter settings -->
|
|
|
102 <configfile name="configfile">
|
|
|
103 ## first we pass some galaxy environment variables
|
|
|
104 galtemp==${__new_file_path__}
|
|
|
105
|
|
|
106 #if $refGenomeSource.genomeSource == "indexed"
|
|
|
107 referencepath==${refGenomeSource.index.fields.path}
|
|
|
108 range=$refGenomeSource.range
|
|
|
109 #else
|
|
|
110 referencepath==${refGenomeSource.ownFile}
|
|
|
111 #end if
|
|
|
112
|
|
|
113 input==$input
|
|
|
114 output==$output
|
|
|
115 q==$q
|
|
|
116 t==$t
|
|
|
117
|
|
|
118 advanced_params.use==$advanced_params.use
|
|
|
119 #if $advanced_params.use=="yes"
|
|
|
120 tileSize==$advanced_params.tileSize
|
|
|
121 stepSize==$advanced_params.stepSize
|
|
|
122 oneOff==$advanced_params.oneOff
|
|
|
123 minMatch==$advanced_params.minMatch
|
|
|
124 minScore==$advanced_params.minScore
|
|
|
125 maxGap==$advanced_params.maxGap
|
|
|
126 mask==$advanced_params.mask
|
|
|
127 qMask==$advanced_params.qMask
|
|
|
128 repeats==$advanced_params.repeats
|
|
|
129 trimT==$advanced_params.trimT
|
|
|
130 noTrimA==$advanced_params.noTrimA
|
|
|
131 fine==$advanced_params.fine
|
|
|
132 maxIntron==$advanced_params.maxIntron
|
|
|
133 extendThroughN==$advanced_params.extendThroughN
|
|
|
134 #end if
|
|
|
135
|
|
|
136 </configfile>
|
|
|
137 </configfiles>
|
|
|
138
|
|
|
139 <tests>
|
|
|
140 <test>
|
|
|
141 <param name="input" value="input83.fas" />
|
|
|
142 <param name="refGenomeSource.genomeSource" value="history" />
|
|
|
143 <param name="refGenomeSource.ownFile" value="databasetest1.txt" />
|
|
|
144 <param name="q" value="dna" />
|
|
|
145 <param name="t" value="dna" />
|
|
|
146 <param name="advanced_params.use" value="yes" />
|
|
|
147 <param name="advanced_params.tileSize" value="11" />
|
|
|
148 <param name="advanced_params.stepSize" value="11" />
|
|
|
149 <param name="advanced_params.oneOff" value="0" />
|
|
|
150 <param name="advanced_params.minMatch" value="2" />
|
|
|
151 <param name="advanced_params.minScore" value="30" />
|
|
|
152 <param name="advanced_params.maxGap" value="2" />
|
|
|
153 <param name="advanced_params.mask" value="lower" />
|
|
|
154 <param name="advanced_params.qMask" value="lower" />
|
|
|
155 <param name="advanced_params.repeats" value="no" />
|
|
|
156 <param name="advanced_params.trimT" value="no" />
|
|
|
157 <param name="advanced_params.noTrimA" value="yes" />
|
|
|
158 <param name="advanced_params.fine" value="no" />
|
|
|
159 <param name="advanced_params.maxIntron" value="750000" />
|
|
|
160 <param name="advanced_params.extendThroughN" value="no" />
|
|
|
161 <output name="output" file="outputtest1.txt" lines_diff="2"/>
|
|
|
162 </test>
|
|
|
163 </tests>
|
|
|
164
|
|
|
165 <help>
|
|
|
166 **BLAT Overview**
|
|
|
167
|
|
|
168 BLAT_ is an alignment tool like BLAST, but it is structured differently.
|
|
|
169 On DNA, Blat works by keeping an index of an entire genome in memory.
|
|
|
170 Thus, the target database of BLAT is not a set of GenBank sequences, but
|
|
|
171 instead an index derived from the assembly of the entire genome. By
|
|
|
172 default, the index consists of all non-overlapping 11-mers except for
|
|
|
173 those heavily involved in repeats.
|
|
|
174
|
|
|
175 .. _BLAT: http://genome.ucsc.edu/FAQ/FAQblat.html
|
|
|
176
|
|
|
177 </help>
|
|
|
178 </tool>
|
|
|
179
|
