|
7
|
1 <tool id="meme_meme" name="MEME" version="1.0.0">
|
|
|
2 <requirements><requirement type='package'>meme</requirement></requirements>
|
|
|
3 <description>motif discovery</description>
|
|
|
4 <command>meme "$input1" -o "${html_outfile.files_path}"
|
|
|
5 -nostatus
|
|
|
6
|
|
|
7 ##-p 8 ##number of processors
|
|
|
8
|
|
|
9 #if str( $options_type.options_type_selector ) == 'advanced':
|
|
|
10 -sf "${ str( $options_type.sf ).replace( ' ', '_' ) }"
|
|
|
11 -${options_type.alphabet_type.alphabet_type_selector}
|
|
|
12 -mod "${options_type.mod_type.mod_type_selector}"
|
|
|
13 -nmotifs "${options_type.nmotifs}"
|
|
|
14 -wnsites "${options_type.wnsites}"
|
|
|
15 -maxsize "${options_type.maxsize}"
|
|
|
16
|
|
|
17 #if $options_type.evt < float('inf'):
|
|
|
18 -evt "${options_type.evt}"
|
|
|
19 #end if
|
|
|
20
|
|
|
21 #if str( $options_type.mod_type.mod_type_selector ) != 'oops':
|
|
|
22 #if str( $options_type.mod_type.motif_occurrence_type.motif_occurrence_type_selector ) == 'nsites':
|
|
|
23 -nsites "${options_type.mod_type.motif_occurrence_type.nsites}"
|
|
|
24 #elif str( $options_type.mod_type.motif_occurrence_type.motif_occurrence_type_selector ) == 'min_max_sites':
|
|
|
25 -minsites "${options_type.mod_type.motif_occurrence_type.minsites}" -maxsites "${options_type.mod_type.motif_occurrence_type.maxsites}"
|
|
|
26 #end if
|
|
|
27 #end if
|
|
|
28
|
|
|
29 #if str( $options_type.motif_width_type.motif_width_type_selector ) == 'exact':
|
|
|
30 -w "${options_type.motif_width_type.width}"
|
|
|
31 #else
|
|
|
32 -minw "${options_type.motif_width_type.minw}" -maxw "${options_type.motif_width_type.maxw}"
|
|
|
33 #end if
|
|
|
34
|
|
|
35 #if str( $options_type.motif_trim_type.motif_trim_type_selector ) == 'nomatrim':
|
|
|
36 -nomatrim
|
|
|
37 #else
|
|
|
38 -wg "${options_type.motif_trim_type.wg}" -ws "${options_type.motif_trim_type.ws}" ${options_type.motif_trim_type.noendgaps}
|
|
|
39 #end if
|
|
|
40
|
|
|
41 #if str( $options_type.bfile ) != 'None':
|
|
|
42 -bfile "${options_type.bfile}"
|
|
|
43 #end if
|
|
|
44
|
|
|
45 #if str( $options_type.pspfile ) != 'None':
|
|
|
46 -psp "${options_type.pspfile}"
|
|
|
47 #end if
|
|
|
48
|
|
|
49 #if str( $options_type.alphabet_type.alphabet_type_selector ) == "dna":
|
|
|
50 ${options_type.alphabet_type.revcomp} ${options_type.alphabet_type.pal}
|
|
|
51 #end if
|
|
|
52
|
|
|
53 -maxiter "${options_type.maxiter}" -distance "${options_type.distance}"
|
|
|
54
|
|
|
55 -prior "${options_type.alphabet_type.prior_type.prior_type_selector}"
|
|
|
56 #if str( $options_type.alphabet_type.prior_type.prior_type_selector ) != 'addone':
|
|
|
57 -b "${options_type.alphabet_type.prior_type.prior_b}"
|
|
|
58 #if str( $options_type.alphabet_type.prior_type.plib ) != 'None':
|
|
|
59 -plib "${options_type.alphabet_type.prior_type.plib}"
|
|
|
60 #end if
|
|
|
61 #end if
|
|
|
62
|
|
|
63 #if str( $options_type.alphabet_type.spmap_type.spmap_type_selector ) == 'cons':
|
|
|
64 -cons "${options_type.alphabet_type.spmap_type.cons}"
|
|
|
65 #else
|
|
|
66 -spmap "${options_type.alphabet_type.spmap_type.spmap_type_selector}"
|
|
|
67 -spfuzz "${options_type.alphabet_type.spmap_type.spfuzz}"
|
|
|
68 #end if
|
|
|
69
|
|
|
70 #if str( $options_type.branching_type.branching_type_selector ) == 'x_branch':
|
|
|
71 -x_branch -bfactor "${options_type.branching_type.bfactor}" -heapsize "${options_type.branching_type.heapsize}"
|
|
|
72 #end if
|
|
|
73
|
|
|
74 ##-maxsize "1000000" ##remove hardcoded maxsize? should increase number of processors instead
|
|
|
75
|
|
|
76 #end if
|
|
|
77
|
|
|
78 2>&1 || echo "Error running MEME."
|
|
|
79
|
|
|
80
|
|
|
81 && mv ${html_outfile.files_path}/meme.html ${html_outfile}
|
|
|
82
|
|
|
83 && mv ${html_outfile.files_path}/meme.txt ${txt_outfile}
|
|
|
84
|
|
|
85 && mv ${html_outfile.files_path}/meme.xml ${xml_outfile}
|
|
|
86
|
|
|
87 </command>
|
|
|
88 <inputs>
|
|
|
89 <param format="fasta" name="input1" type="data" label="Sequences"/>
|
|
|
90
|
|
|
91 <conditional name="options_type">
|
|
|
92 <param name="options_type_selector" type="select" label="Options Configuration">
|
|
|
93 <option value="basic" selected="true">Basic</option>
|
|
|
94 <option value="advanced">Advanced</option>
|
|
|
95 </param>
|
|
|
96 <when value="basic">
|
|
|
97 <!-- do nothing here -->
|
|
|
98 </when>
|
|
|
99 <when value="advanced">
|
|
|
100
|
|
|
101 <param name="sf" type="text" value="Galaxy FASTA Input" label="Name of sequence set" />
|
|
|
102
|
|
|
103 <conditional name="alphabet_type">
|
|
|
104 <param name="alphabet_type_selector" type="select" label="Sequence Alphabet">
|
|
|
105 <option value="protein">Protein</option>
|
|
|
106 <option value="dna" selected="true">DNA</option>
|
|
|
107 </param>
|
|
|
108 <when value="protein">
|
|
|
109 <conditional name="prior_type">
|
|
|
110 <param name="prior_type_selector" type="select" label="Choice of prior">
|
|
|
111 <option value="dirichlet">simple Dirichlet prior</option>
|
|
|
112 <option value="dmix" selected="true">mixture of Dirichlets prior</option>
|
|
|
113 <option value="mega">extremely low variance dmix</option>
|
|
|
114 <option value="megap">mega for all but last iteration of EM; dmix on last iteration</option>
|
|
|
115 <option value="addone">add +1 to each observed count</option>
|
|
|
116 </param>
|
|
|
117 <when value="dirichlet">
|
|
|
118 <param name="prior_b" type="float" value="0.01" label="strength of prior on model parameters" />
|
|
|
119 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
120 </when>
|
|
|
121 <when value="dmix">
|
|
|
122 <param name="prior_b" type="float" value="0" label="strength of prior on model parameters" />
|
|
|
123 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
124 </when>
|
|
|
125 <when value="mega">
|
|
|
126 <param name="prior_b" type="float" value="0" label="strength of prior on model parameters" />
|
|
|
127 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
128 </when>
|
|
|
129 <when value="megap">
|
|
|
130 <param name="prior_b" type="float" value="0" label="strength of prior on model parameters" />
|
|
|
131 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
132 </when>
|
|
|
133 <when value="addone">
|
|
|
134 <!-- no values here? -->
|
|
|
135 </when>
|
|
|
136 </conditional>
|
|
|
137 <conditional name="spmap_type">
|
|
|
138 <param name="spmap_type_selector" type="select" label="EM starting points">
|
|
|
139 <option value="uni">uni</option>
|
|
|
140 <option value="pam" selected="true">pam</option>
|
|
|
141 <option value="cons">Use starting point from string</option>
|
|
|
142 </param>
|
|
|
143 <when value="uni">
|
|
|
144 <param name="spfuzz" type="float" value="0.5" label="Fuzziness of the mapping" />
|
|
|
145 </when>
|
|
|
146 <when value="pam">
|
|
|
147 <param name="spfuzz" type="integer" value="120" label="Fuzziness of the mapping" />
|
|
|
148 </when>
|
|
|
149 <when value="cons">
|
|
|
150 <param name="cons" type="text" value="" label="Starting point from string" />
|
|
|
151 </when>
|
|
|
152 </conditional>
|
|
|
153 </when>
|
|
|
154 <when value="dna">
|
|
|
155 <param name="revcomp" label="Check reverse complement" type="boolean" truevalue="-revcomp" falsevalue="" checked="False"/>
|
|
|
156 <param name="pal" label="Check for palindromes" type="boolean" truevalue="-pal" falsevalue="" checked="False"/>
|
|
|
157 <conditional name="prior_type">
|
|
|
158 <param name="prior_type_selector" type="select" label="Sequence Alphabet">
|
|
|
159 <option value="dirichlet" selected="true">simple Dirichlet prior</option>
|
|
|
160 <option value="dmix">mixture of Dirichlets prior</option>
|
|
|
161 <option value="mega">extremely low variance dmix</option>
|
|
|
162 <option value="megap">mega for all but last iteration of EM; dmix on last iteration</option>
|
|
|
163 <option value="addone">add +1 to each observed count</option>
|
|
|
164 </param>
|
|
|
165 <when value="dirichlet">
|
|
|
166 <param name="prior_b" type="float" value="0.01" label="strength of prior on model parameters" />
|
|
|
167 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
168 </when>
|
|
|
169 <when value="dmix">
|
|
|
170 <param name="prior_b" type="float" value="0" label="strength of prior on model parameters" />
|
|
|
171 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
172 </when>
|
|
|
173 <when value="mega">
|
|
|
174 <param name="prior_b" type="float" value="0" label="strength of prior on model parameters" />
|
|
|
175 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
176 </when>
|
|
|
177 <when value="megap">
|
|
|
178 <param name="prior_b" type="float" value="0" label="strength of prior on model parameters" />
|
|
|
179 <param name="plib" type="data" format="txt" optional="True" label="Dirichlet prior file" />
|
|
|
180 </when>
|
|
|
181 <when value="addone">
|
|
|
182 <!-- no values here? -->
|
|
|
183 </when>
|
|
|
184 </conditional>
|
|
|
185 <conditional name="spmap_type">
|
|
|
186 <param name="spmap_type_selector" type="select" label="EM starting points">
|
|
|
187 <option value="uni" selected="true">uni</option>
|
|
|
188 <option value="pam">pam</option>
|
|
|
189 <option value="cons">Use starting point from string</option>
|
|
|
190 </param>
|
|
|
191 <when value="uni">
|
|
|
192 <param name="spfuzz" type="float" value="0.5" label="Fuzziness of the mapping" />
|
|
|
193 </when>
|
|
|
194 <when value="pam">
|
|
|
195 <param name="spfuzz" type="integer" value="120" label="Fuzziness of the mapping" />
|
|
|
196 </when>
|
|
|
197 <when value="cons">
|
|
|
198 <param name="cons" type="text" value="" label="Starting point from string" />
|
|
|
199 </when>
|
|
|
200 </conditional>
|
|
|
201 </when>
|
|
|
202 </conditional>
|
|
|
203
|
|
|
204 <param name="nmotifs" type="integer" value="1" label="Number of different motifs to search" />
|
|
|
205 <param name="maxsize" type="integer" value="1000000" label="Max number of characters in the sequence file"/>
|
|
|
206 <param name="evt" type="float" value="inf" label="E-value to stop looking for motifs" />
|
|
|
207 <conditional name="mod_type">
|
|
|
208 <param name="mod_type_selector" type="select" label="Expected motif distribution">
|
|
|
209 <option value="oops">One Occurrence Per Sequence</option>
|
|
|
210 <option value="zoops" selected="true">Zero or One Occurrence Per Sequence</option>
|
|
|
211 <option value="anr">Any Number of Repetitions</option>
|
|
|
212 </param>
|
|
|
213 <when value="oops">
|
|
|
214 <!-- no values here -->
|
|
|
215 </when>
|
|
|
216 <when value="zoops">
|
|
|
217 <conditional name="motif_occurrence_type">
|
|
|
218 <param name="motif_occurrence_type_selector" type="select" label="Number of motif occurrences">
|
|
|
219 <option value="default" selected="true">Use defaults</option>
|
|
|
220 <option value="nsites">nsites</option>
|
|
|
221 <option value="min_max_sites">min and max sites</option>
|
|
|
222 </param>
|
|
|
223 <when value="default">
|
|
|
224 <!-- no values here -->
|
|
|
225 </when>
|
|
|
226 <when value="nsites">
|
|
|
227 <param name="nsites" type="integer" value="1" label="Search nsites number of occurrences" />
|
|
|
228 </when>
|
|
|
229 <when value="min_max_sites">
|
|
|
230 <param name="minsites" type="integer" value="1" label="minsites" />
|
|
|
231 <param name="maxsites" type="integer" value="50" label="maxsites" />
|
|
|
232 </when>
|
|
|
233 </conditional>
|
|
|
234 </when>
|
|
|
235 <when value="anr">
|
|
|
236 <conditional name="motif_occurrence_type">
|
|
|
237 <param name="motif_occurrence_type_selector" type="select" label="Number of motif occurrences">
|
|
|
238 <option value="default" selected="true">Use defaults</option>
|
|
|
239 <option value="nsites">nsites</option>
|
|
|
240 <option value="min_max_sites">min and max sites</option>
|
|
|
241 </param>
|
|
|
242 <when value="default">
|
|
|
243 <!-- no values here -->
|
|
|
244 </when>
|
|
|
245 <when value="nsites">
|
|
|
246 <param name="nsites" type="integer" value="1" label="Search nsites number of occurrences" />
|
|
|
247 </when>
|
|
|
248 <when value="min_max_sites">
|
|
|
249 <param name="minsites" type="integer" value="1" label="minsites" />
|
|
|
250 <param name="maxsites" type="integer" value="50" label="maxsites" />
|
|
|
251 </when>
|
|
|
252 </conditional>
|
|
|
253 </when>
|
|
|
254 </conditional>
|
|
|
255 <param name="wnsites" type="float" value="0.8" label="Weight on the prior on nsites" />
|
|
|
256
|
|
|
257 <conditional name="motif_width_type">
|
|
|
258 <param name="motif_width_type_selector" type="select" label="Motif width type">
|
|
|
259 <option value="exact">Exact width</option>
|
|
|
260 <option value="range" selected="true">Specify a range</option>
|
|
|
261 </param>
|
|
|
262 <when value="exact">
|
|
|
263 <param name="width" type="integer" value="10" label="Width of motif to search" />
|
|
|
264 </when>
|
|
|
265 <when value="range">
|
|
|
266 <param name="minw" type="integer" value="8" label="Min width of motif to search" />
|
|
|
267 <param name="maxw" type="integer" value="50" label="Max width of motif to search" />
|
|
|
268 </when>
|
|
|
269 </conditional>
|
|
|
270
|
|
|
271 <conditional name="motif_trim_type">
|
|
|
272 <param name="motif_trim_type_selector" type="select" label="Motif trim type">
|
|
|
273 <option value="nomatrim">No motif trim</option>
|
|
|
274 <option value="trim" selected="true">Trim motif</option>
|
|
|
275 </param>
|
|
|
276 <when value="nomatrim">
|
|
|
277 <!-- no values here -->
|
|
|
278 </when>
|
|
|
279 <when value="trim">
|
|
|
280 <param name="wg" type="integer" value="11" label="Gap cost" />
|
|
|
281 <param name="ws" type="integer" value="1" label="Space cost" />
|
|
|
282 <param name="noendgaps" label="Do not penalize endgaps" type="boolean" truevalue="-noendgaps" falsevalue="" checked="False"/>
|
|
|
283 </when>
|
|
|
284 </conditional>
|
|
|
285
|
|
|
286 <param name="bfile" type="data" format="txt" optional="True" label="Background Model" />
|
|
|
287 <param name="pspfile" type="data" format="txt" optional="True" label="Position-Specific Prior" />
|
|
|
288
|
|
|
289 <param name="maxiter" type="integer" value="50" label="Number of iterations of EM to run" />
|
|
|
290 <param name="distance" type="float" value="0.001" label="Convergence criterion" />
|
|
|
291
|
|
|
292 <conditional name="branching_type">
|
|
|
293 <param name="branching_type_selector" type="select" label="x-branching type">
|
|
|
294 <option value="x_branch">Perform x-branching</option>
|
|
|
295 <option value="no_x_branch" selected="true">No x-branching</option>
|
|
|
296 </param>
|
|
|
297 <when value="no_x_branch">
|
|
|
298 <!-- no values here -->
|
|
|
299 </when>
|
|
|
300 <when value="x_branch">
|
|
|
301 <param name="bfactor" type="integer" value="3" label="Number of iterations of branching" />
|
|
|
302 <param name="heapsize" type="integer" value="64" label="Maximum number of heaps to use" />
|
|
|
303 </when>
|
|
|
304 </conditional>
|
|
|
305
|
|
|
306 </when>
|
|
|
307 </conditional>
|
|
|
308
|
|
|
309 <param name="non_commercial_use" label="I certify that I am not using this tool for commercial purposes." type="boolean" truevalue="NON_COMMERCIAL_USE" falsevalue="COMMERCIAL_USE" checked="False">
|
|
|
310 <validator type="expression" message="This tool is only available for non-commercial use.">value == True</validator>
|
|
|
311 </param>
|
|
|
312
|
|
|
313 </inputs>
|
|
|
314 <outputs>
|
|
|
315 <data format="html" name="html_outfile" label="${tool.name} on ${on_string} (html)"/>
|
|
|
316 <data format="txt" name="txt_outfile" label="${tool.name} on ${on_string} (text)"/>
|
|
|
317 <data format="memexml" name="xml_outfile" label="${tool.name} on ${on_string} (xml)"/>
|
|
|
318 </outputs>
|
|
|
319 <tests>
|
|
|
320 <test>
|
|
|
321 <param name="input1" value="meme/meme/meme_input_1.fasta" ftype="fasta" dbkey="hg19"/>
|
|
|
322 <param name="options_type_selector" value="basic"/>
|
|
|
323 <param name="non_commercial_use" value="True"/>
|
|
|
324 <output name="html_outfile" file="meme/meme/meme_output_html_1.html" lines_diff="12"/>
|
|
|
325 <output name="txt_outfile" file="meme/meme/meme_output_txt_1.txt" lines_diff="12"/>
|
|
|
326 <output name="xml_outfile" file="meme/meme/meme_output_xml_1.xml" lines_diff="8"/>
|
|
|
327 </test>
|
|
|
328 </tests>
|
|
|
329 <help>
|
|
|
330
|
|
|
331 .. class:: warningmark
|
|
|
332
|
|
|
333 **WARNING: This tool is only available for non-commercial use. Use for educational, research and non-profit purposes is permitted. Before using, be sure to review, agree, and comply with the license.**
|
|
|
334
|
|
|
335 If you want to specify sequence weights, you must include them at the top of your input FASTA file.
|
|
|
336
|
|
|
337 .. class:: infomark
|
|
|
338
|
|
|
339 **To cite MEME:**
|
|
|
340 Timothy L. Bailey and Charles Elkan, "Fitting a mixture model by expectation maximization to discover motifs in biopolymers", Proceedings of the Second International Conference on Intelligent Systems for Molecular Biology, pp. 28-36, AAAI Press, Menlo Park, California, 1994.
|
|
|
341
|
|
|
342
|
|
|
343 For detailed information on MEME, click here_. To view the license_.
|
|
|
344
|
|
|
345 .. _here: http://meme.nbcr.net/meme/meme-intro.html
|
|
|
346 .. _license: http://meme.nbcr.net/meme/COPYRIGHT.html
|
|
|
347
|
|
|
348 </help>
|
|
|
349 </tool>
|
