|
0
|
1 <tool id="pick_otus" name="pick_otus" version="2.0.0">
|
|
|
2 <description>OTU picking</description>
|
|
|
3 <requirements>
|
|
|
4 <requirement type="binary">pick_otus.py</requirement>
|
|
|
5 </requirements>
|
|
|
6 <command interpreter="python">
|
|
|
7 qiime_wrapper.py
|
|
|
8 --galaxy_outputdir='$log.extra_files_path'
|
|
|
9 #if $pick.otu_picking_method == 'uclust' and $pick.refseqs_fp.__str__ != 'None':
|
|
|
10 --galaxy_datasets='^\S+_otus\.txt$:'$otus,'^\S+_otus\.log$:'$log,'^\S+_failures\.txt$:'$failures
|
|
|
11 #else:
|
|
|
12 --galaxy_datasets='^\S+_otus\.txt$:'$otus,'^\S+_otus\.log$:'$log
|
|
|
13 #end if
|
|
|
14 pick_otus.py
|
|
|
15 --input_seqs_filepath=$input_seqs_filepath
|
|
|
16 #if $pick.otu_picking_method.__str__ == 'uclust':
|
|
|
17 #if $pick.refseqs_fp.__str__ != 'None':
|
|
|
18 --refseqs_fp=$pick.refseqs_fp
|
|
|
19 --otu_picking_method='uclust_ref'
|
|
|
20 $pick.suppress_new_clusters
|
|
|
21 #else:
|
|
|
22 --otu_picking_method=$pick.otu_picking_method
|
|
|
23 #end if
|
|
|
24 --similarity=$pick.similarity
|
|
|
25 $pick.enable_rev_strand_match
|
|
|
26 $pick.optimal_uclust
|
|
|
27 $pick.exact_uclust
|
|
|
28 $pick.user_sort
|
|
|
29 $pick.suppress_presort_by_abundance_uclust
|
|
|
30 --max_accepts=$pick.max_accepts
|
|
|
31 --max_rejects=$pick.max_rejects
|
|
|
32 #if $pick.uclust_otu_id_prefix != None and $pick.uclust_otu_id_prefix.__str__ != 'None' and $pick.uclust_otu_id_prefix.__str__ != '':
|
|
|
33 --uclust_otu_id_prefix=$pick.uclust_otu_id_prefix
|
|
|
34 #end if
|
|
|
35 $pick.uclust_stable_sort
|
|
|
36 $pick.save_uc_files
|
|
|
37 #elif $pick.otu_picking_method.__str__ == 'mothur':
|
|
|
38 --otu_picking_method=$pick.otu_picking_method
|
|
|
39 --clustering_algorithm=$pick.clustering_algorithm
|
|
|
40 --similarity=$pick.similarity
|
|
|
41 #elif $pick.otu_picking_method.__str__ == 'trie':
|
|
|
42 --otu_picking_method=$pick.otu_picking_method
|
|
|
43 $pick.trie_reverse_seqs
|
|
|
44 #elif $pick.otu_picking_method.__str__ == 'prefix_suffix':
|
|
|
45 --otu_picking_method=$pick.otu_picking_method
|
|
|
46 --prefix_length=$pick.prefix_length
|
|
|
47 --suffix_length=$pick.suffix_length
|
|
|
48 #elif pick.otu_picking_method.__str__ == 'blast':
|
|
|
49 --otu_picking_method=$pick.otu_picking_method
|
|
|
50 #if $refseqs_fp.__str__ != 'None':
|
|
|
51 --refseqs_fp=$pick.refseqs_fp
|
|
|
52 #end if
|
|
|
53 #if $pick.blast_db != None and $pick.blast_db.__str__ != 'None' and $pick.blast_db.__str__ != '':
|
|
|
54 --blast_db=$pick.blast_db
|
|
|
55 #end if
|
|
|
56 --similarity=$pick.similarity
|
|
|
57 --max_e_value=$pick.max_e_value
|
|
|
58 --min_aligned_percent=$pick.min_aligned_percent
|
|
|
59 #elif $pick.otu_picking_method == 'cdhit':
|
|
|
60 --otu_picking_method=$pick.otu_picking_method
|
|
|
61 --similarity=$pick.similarity
|
|
|
62 --max_cdhit_memory=$pick.max_cdhit_memory
|
|
|
63 #if $pick.prefix_prefilter_length != 0:
|
|
|
64 --prefix_prefilter_length=$pick.prefix_prefilter_length
|
|
|
65 #end if
|
|
|
66 $pick.trie_prefilter
|
|
|
67 #end if
|
|
|
68 --output_dir='$log.extra_files_path'
|
|
|
69 </command>
|
|
|
70 <inputs>
|
|
|
71 <param name="input_seqs_filepath" type="data" format="fasta" label="input_seqs_filepath"
|
|
|
72 help="Input sequences [REQUIRED]"/>
|
|
|
73 <conditional name="pick">
|
|
|
74 <param name="otu_picking_method" type="select" label="otu_picking_method"
|
|
|
75 help="Method for picking OTUs. Valid choices are: mothur, trie, uclust_ref, prefix_suffix, blast, cdhit, uclust. The mothur method requires an input file of aligned sequences [default: uclust]">
|
|
|
76 <option value="uclust" selected="true">uclust</option>
|
|
|
77 <option value="mothur">mothur</option>
|
|
|
78 <option value="trie">trie</option>
|
|
|
79 <option value="prefix_suffix">prefix_suffix</option>
|
|
|
80 <option value="blast">blast</option>
|
|
|
81 <option value="cdhit">cdhit</option>
|
|
|
82 </param>
|
|
|
83 <when value="mothur">
|
|
|
84 <param name="clustering_algorithm" type="select" label="clustering_algorithm"
|
|
|
85 help="Clustering algorithm for mothur otu picking method. [default: furthest]">
|
|
|
86 <option value="furthest" selected="true">furthest</option>
|
|
|
87 <option value="nearest">nearest</option>
|
|
|
88 <option value="average">average</option>
|
|
|
89 </param>
|
|
|
90 <param name="similarity" type="float" value="0.97" label="similarity"
|
|
|
91 help="Sequence similarity threshold (for cdhit, uclust, or uclust_ref) [default: 0.97]"/>
|
|
|
92 </when> <!-- mothur -->
|
|
|
93 <when value="trie">
|
|
|
94 <param name="trie_reverse_seqs" type="boolean" truevalue="--trie_reverse_seqs" falsevalue="" checked="false" label="trie_reverse_seqs"
|
|
|
95 help="Reverse seqs before picking OTUs with the Trie OTU picker for suffix (rather than prefix) collapsing [default: False]"/>
|
|
|
96 </when> <!-- trie -->
|
|
|
97 <when value="uclust">
|
|
|
98 <param name="refseqs_fp" type="data" format="fasta" label="refseqs_fp" optional="true"
|
|
|
99 help="Reference sequences to search against when using blast, uclust_ref, or usearch_ref [OPTIONAL]"/>
|
|
|
100 <param name="suppress_new_clusters" type="boolean" truevalue="--suppress_new_clusters" falsevalue="" checked="false" label="suppress_new_clusters"
|
|
|
101 help="Suppress creation of new clusters using seqs that don't match reference when using -m uclust_ref [default: False]"/>
|
|
|
102 <param name="suppress_presort_by_abundance_uclust" type="boolean" truevalue="--suppress_presort_by_abundance_uclust" falsevalue="" checked="false" label="suppress_presort_by_abundance_uclust"
|
|
|
103 help="Suppress presorting of sequences by abundance when picking OTUs with uclust or uclust_ref [default: False]"/>
|
|
|
104 <param name="similarity" type="float" value="0.97" label="similarity"
|
|
|
105 help="Sequence similarity threshold (for cdhit, uclust, or uclust_ref) [default: 0.97]"/>
|
|
|
106 <param name="enable_rev_strand_match" type="boolean" truevalue="--enable_rev_strand_match" falsevalue="" checked="false" label="enable_rev_strand_match"
|
|
|
107 help="Enable reverse strand matching for uclust otu picking, will double the amount of memory used. [default: False]"/>
|
|
|
108 <param name="optimal_uclust" type="boolean" truevalue="--optimal_uclust" falsevalue="" checked="false" label="optimal_uclust"
|
|
|
109 help="Pass the --optimal flag to uclust for uclust otu picking. [default: False]"/>
|
|
|
110 <param name="exact_uclust" type="boolean" truevalue="--exact_uclust" falsevalue="" checked="false" label="exact_uclust"
|
|
|
111 help="Pass the --exact flag to uclust for uclust otu picking. [default: False]"/>
|
|
|
112 <param name="user_sort" type="boolean" truevalue="--user_sort" falsevalue="" checked="false" label="user_sort"
|
|
|
113 help="Pass the --user_sort flag to uclust for uclust otu picking. [default: False]"/>
|
|
|
114 <param name="max_accepts" type="integer" value="20" label="max_accepts"
|
|
|
115 help="max_accepts value to uclust and uclust_ref [default: 20]"/>
|
|
|
116 <param name="max_rejects" type="integer" value="500" label="max_rejects"
|
|
|
117 help="max_rejects value to uclust and uclust_ref [default: 500]"/>
|
|
|
118 <param name="uclust_otu_id_prefix" type="text" label="uclust_otu_id_prefix"
|
|
|
119 help="OTU identifier prefix (string) for the de novo uclust OTU picker [default: None, OTU ids are ascending integers]"/>
|
|
|
120 <param name="uclust_stable_sort" type="boolean" truevalue="--uclust_stable_sort" falsevalue="" checked="false" label="uclust_stable_sort"
|
|
|
121 help="pass --stable_sort to uclust (uclust versions uclustq1.2.15 and later only) [default: False]"/>
|
|
|
122 <param name="save_uc_files" type="boolean" truevalue="" falsevalue="--save_uc_files" checked="false" label="save_uc_files"
|
|
|
123 help="Enable preservation of intermediate uclust (.uc) files that are used to generate clusters via uclust. [default: True]"/>
|
|
|
124 </when> <!-- uclust -->
|
|
|
125 <when value="prefix_suffix">
|
|
|
126 <param name="prefix_length" type="integer" value="50" label="prefix_length"
|
|
|
127 help="Prefix length when using the prefix_suffix otu picker; WARNING: CURRENTLY DIFFERENT FROM prefix_prefilter_length (-n)! [default: 50]"/>
|
|
|
128 <param name="suffix_length" type="integer" value="50" label="suffix_length"
|
|
|
129 help="Suffix length when using the prefix_suffix otu picker [default: 50]"/>
|
|
|
130
|
|
|
131 </when> <!-- prefix_suffix -->
|
|
|
132 <when value="blast">
|
|
|
133 <param name="refseqs_fp" type="data" format="fasta" label="refseqs_fp" optional="true"
|
|
|
134 help="Reference sequences to search against"/>
|
|
|
135 <param name="blast_db" type="data" format="txt" optional="True" label="blast_db"
|
|
|
136 help="Pre-existing database to blast against when using -m blast [OPTIONAL]"/>
|
|
|
137 <param name="min_aligned_percent" type="float" value="0.5" label="min_aligned_percent"
|
|
|
138 help="Minimum percent of query sequence that can be aligned to consider a hit (BLAST OTU picker only) [default: 0.5]"/>
|
|
|
139 <param name="max_e_value" type="float" value="1e-10" label="max_e_value"
|
|
|
140 help="Max E-value when clustering with BLAST [default: 1e-10]"/>
|
|
|
141 </when> <!-- blast -->
|
|
|
142 <when value="cdhit">
|
|
|
143 <param name="max_cdhit_memory" type="integer" value="400" label="max_cdhit_memory"
|
|
|
144 help="Maximum available memory to cd-hit-est (via the program's -M option) for cdhit OTU picking method (units of Mbyte) [default: 400]"/>
|
|
|
145 <param name="similarity" type="float" value="0.97" label="similarity"
|
|
|
146 help="Sequence similarity threshold (for cdhit, uclust, or uclust_ref) [default: 0.97]"/>
|
|
|
147 <param name="prefix_prefilter_length" type="integer" value="0" label="prefix_prefilter_length"
|
|
|
148 help="Prefilter data so seqs with identical first prefix_prefilter_length are automatically grouped into a single OTU. This is useful for large sequence collections where OTU picking doesn't scale well [default: None; 100 is a good value]"/>
|
|
|
149 <param name="trie_prefilter" type="boolean" truevalue="--trie_prefilter" falsevalue="" checked="false" label="trie_prefilter"
|
|
|
150 help="prefilter data so seqs which are identical prefixes of a longer seq are automatically grouped into a single OTU; useful for large sequence collections where OTU picking doesn't scale well [default: False]"/>
|
|
|
151 </when> <!-- cdhit -->
|
|
|
152 </conditional> <!-- pick -->
|
|
|
153 </inputs>
|
|
|
154 <outputs>
|
|
|
155 <data format="txt" name="log" label="${tool.name} on ${on_string}: log" />
|
|
|
156 <data format="qiimeotu" name="otus" label="${tool.name} on ${on_string}: otus" />
|
|
|
157 <data format="txt" name="failures" label="${tool.name} on ${on_string}: failures.txt" >
|
|
|
158 <filter>(pick['otu_picking_method'] == 'uclust' and pick['refseqs_fp'])</filter>
|
|
|
159 </data>
|
|
|
160 </outputs>
|
|
|
161 <tests>
|
|
|
162 </tests>
|
|
|
163 <help>For more information, see pick_otus_ in the Qiime documentation.
|
|
|
164
|
|
|
165 Updated and validated 01/16/12 by Amanda Zuzolo, Microbiome Analysis Center, George Mason University, Fairfax, VA
|
|
|
166
|
|
|
167 Qiime integration courtesy Jim Johnson, Minnesota Supercomputing Institute, University of Minnesota, Minneapolis, MN
|
|
|
168
|
|
|
169 .. _pick_otus: http://qiime.org/scripts/pick_otus.html</help>
|
|
|
170 </tool>
|
|
|
171
|
