Mercurial > repos > iuc > miniprot

comparison miniprot.xml @ 0:ef712a5e9834 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/miniprot commit 931e98e27ac60b189e2dfbb1c99767bd17860c5e
author iuc
date Mon, 19 Sep 2022 12:30:10 +0000
parents
children ce04c239454b
comparison
equal deleted inserted replaced
-1:000000000000 0:ef712a5e9834
1 <?xml version="1.0"?>
2 <tool id="miniprot" name="Miniprot align" version="@TOOL_VERSION@+galaxy0" profile="21.05">
3 <description>align a protein sequence against a genome with affine gap penalty, splicing and frameshift</description>
4 <macros>
5 <import>macros.xml</import>
6 </macros>
7 <requirements>
8 <requirement type="package" version="@TOOL_VERSION@">miniprot</requirement>
9 </requirements>
10 <command detect_errors="exit_code"><![CDATA[
11 miniprot
12 -t \${GALAXY_SLOTS:-1}
13 #if str($adv.options) == "yes"
14 $adv.mapping.no_splicing
15 -c $adv.mapping.max_kmer
16 -G $adv.mapping.max_intron
17 -n $adv.mapping.min_syncmers
18 -m $adv.mapping.min_chain_score
19 -l $adv.mapping.second_round_kmer_size
20 -e $adv.mapping.max_extension
21 -p $adv.mapping.score_ratio
22 -N $adv.mapping.max_secondary_alignments
23 -O $adv.alignment.gap_open
24 -E $adv.alignment.gap_extension
25 -J $adv.alignment.intron_open
26 -C $adv.alignment.non_canonical_splice
27 -F $adv.alignment.frameshift
28 -B $adv.alignment.end_bonus
29 #end if
30 #if str($db.dbtype) == 'fasta'
31 '$db.genomic_fasta'
32 -k $db.kmer_size
33 -s $db.submer_size
34 -b $db.bits_per_block
35 #else
36 '$db.genomic_db'
37 #end if
38 #if str($output_format) == "gff"
39 --gff
40 #end if
41 '$protein_fasta'
42 >'$output_alignment'
43 ]]></command>
44 <inputs>
45 <conditional name="db">
46 <param name="dbtype" type="select" label="Database type" help="Build an index from FASTA or use a pre-indexed database">
47 <option value="fasta" selected="true">FASTA</option>
48 <option value="preindexed">Pre-indexed</option>
49 </param>
50 <when value="fasta">
51 <param name="genomic_fasta" type="data" format="fasta,fasta.gz" label="Genomic sequence (FASTA)" help="Genomic contigs / scaffolds to be aligned against in FASTA format" />
52 <param argument="-k" name="kmer_size" type="integer" min="1" value="6" label="K-mer size" />
53 <param argument="-s" name="submer_size" type="integer" min="1" value="4" label="Submer size" help="Submer size (density: 1/(2*(kmer_size-submer_size)+1))" />
54 <param argument="-b" name="bits_per_block" type="integer" min="1" value="8" label="Bits per block" />
55 </when>
56 <when value="preindexed">
57 <!-- refine the datatype here once Miniprot index data type is in Galaxy -->
58 <param name="genomic_db" type="data" format="binary" label="Pre-indexed genomic database" help="A pre-indexed database built by miniprot" />
59 </when>
60 </conditional>
61 <param name="protein_fasta" type="data" format="fasta,fasta.gz" label="Protein sequence (FASTA)" help="Protein sequences to be aligned in FASTA format" />
62 <param name="output_format" type="select" label="Output format" >
63 <option value="gff" selected="true">GFF3</option>
64 <option value="paf">PAF</option>
65 </param>
66 <conditional name="adv">
67 <param name="options" type="select" label="Advanced options">
68 <option value="yes">Show</option>
69 <option value="no" selected="true">Hide</option>
70 </param>
71 <when value="yes">
72 <section name="mapping" title="Mapping">
73 <param argument="-S" name="no_splicing" type="boolean" truevalue="-S" falsevalue="" checked="false" label="No splicing" help="No splicing (apply -G1000 -J1000 -e1000)" />
74 <param argument="-c" name="max_kmer" type="integer" min="1" value="50000" label="Max k-mer occurences" />
75 <param argument="-G" name="max_intron" type="integer" min="0" value="200000" label="Max intron size" />
76 <param argument="-n" name="min_syncmers" type="integer" min="1" value="5" label="Minimum number of syncmers in a chain" />
77 <param argument="-m" name="min_chain_score" type="integer" min="0" value="0" label="Minimum chaining score" />
78 <param argument="-l" name="second_round_kmer_size" type="integer" min="1" value="5" label="K-mer size for second round of chaining" />
79 <param argument="-e" name="max_extension" type="integer" min="0" value="10000" label="Max extension for second round of chaining" />
80 <param argument="-p" name="score_ratio" type="float" min="0" max="1" value="0.5" label="Minimum secondary-to-primary score ratio" />
81 <param argument="-N" name="max_secondary_alignments" type="integer" min="0" value="100" label="Max secondary alignments to consider" />
82 </section>
83 <section name="alignment" title="Alignment">
84 <param argument="-O" name="gap_open" type="integer" min="0" value="11" label="Gap open penalty" />
85 <param argument="-E" name="gap_extension" type="integer" min="0" value="1" label="Gap extension penalty" help="A k-long gap costs open_penalty+k*extension_penalty" />
86 <param argument="-J" name="intron_open" type="integer" min="0" value="31" label="Intron open penalty" />
87 <param argument="-C" name="non_canonical_splice" type="integer" min="0" value="11" label="Penalty for non-canonical splicing" />
88 <param argument="-F" name="frameshift" type="integer" min="0" value="15" label="Frameshift penalty" />
89 <param argument="-B" name="end_bonus" type="integer" min="0" value="5" label="End bonus" />
90 </section>
91 <param argument="-K" name="query_batch_size" type="integer" min="1" value="2000000" label="Query batch size" />
92 </when>
93 <when value="no">
94 </when>
95 </conditional>
96 </inputs>
97 <outputs>
98 <data name="output_alignment" format="gff3" label="Miniprot on ${on_string}">
99 <change_format>
100 <when input="output_format" value="paf" format="paf" />
101 </change_format>
102 </data>
103 </outputs>
104 <tests>
105 <test expect_num_outputs="1">
106 <conditional name="db">
107 <param name="dbtype" value="fasta" />
108 <param name="genomic_fasta" value="input_genome.fasta.gz" ftype="fasta" />
109 </conditional>
110 <param name="protein_fasta" value="input_query.fasta.gz" ftype="fasta" />
111 <output name="output_alignment" ftype="gff3">
112 <assert_contents>
113 <has_text text="ID=MP000001;Identity=1.0000;Positive=1.0000;Target=tr|O06302|O06302_MYCTU 1 126" />
114 <has_text text="Parent=MP000372;Target=tr|V5QPR5|V5QPR5_MYCTU 1 53" />
115 </assert_contents>
116 </output>
117 </test>
118 <test expect_num_outputs="1">
119 <conditional name="db">
120 <param name="dbtype" value="fasta" />
121 <param name="genomic_fasta" value="input_genome.fasta.gz" ftype="fasta" />
122 </conditional>
123 <param name="protein_fasta" value="input_query.fasta.gz" ftype="fasta" />
124 <param name="output_format" value="paf" />
125 <output name="output_alignment" ftype="paf">
126 <assert_contents>
127 <has_text text="tr|O06302|O06302_MYCTU" />
128 <has_text text="cs:Z::29*agcG:3*gtgA:5*ccgA:9*accS:1*gccV:4*cagL:1*gtcS:3*gtcA*gtcI*accA*gccG:8*gccS:2*ggtA:5*gccI*agcG:1*ctgA:4*gccV:5*gggL:1*gtgS:2" />
129 </assert_contents>
130 </output>
131 </test>
132 <test expect_num_outputs="1">
133 <conditional name="db">
134 <param name="dbtype" value="fasta" />
135 <param name="genomic_fasta" value="input_genome.fasta.gz" ftype="fasta" />
136 </conditional>
137 <param name="protein_fasta" value="input_query.fasta.gz" ftype="fasta" />
138 <param name="output_format" value="gff" />
139 <conditional name="adv">
140 <param name="options" value="yes" />
141 <param name="second_round_kmer_size" value="32" />
142 </conditional>
143 <output name="output_alignment" ftype="gff3">
144 <assert_contents>
145 <has_text text="##gff-version 3" />
146 </assert_contents>
147 </output>
148 </test>
149 </tests>
150 <help><![CDATA[
151 miniprot_ rapidly aligns a protein sequence against a genome with affine gap penalty, splicing and frameshift.
152 It is primarily intended for annotating protein-coding genes in a new species using known genes from other species.
153
154 **NOTE:** miniprot is in the early stages of development and should be considered experimental at this stage.
155 .. _miniprot: https://github.com/lh3/miniprot
156 ]]></help>
157 </tool>