Mercurial > repos > bgruening > flye

comparison flye.xml @ 8:2522110f3cb4 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flye commit 7c956f5b897dc366b2e5b7e37a2fea4b51a814f3"
author bgruening
date Tue, 23 Nov 2021 08:34:56 +0000
parents 28898f6f2352
children 5f2671cec451
comparison
equal deleted inserted replaced
7:28898f6f2352 8:2522110f3cb4
18 #set $ext = 'fasta' 18 #set $ext = 'fasta'
19 #end if 19 #end if
20 ln -s '$input' ./input_${counter}.${ext} && 20 ln -s '$input' ./input_${counter}.${ext} &&
21 #end for 21 #end for
22 flye 22 flye
23 $mode 23 $mode_conditional.mode
24 #for $counter, $input in enumerate($inputs): 24 #for $counter, $input in enumerate($inputs):
25 ./input_${counter}.$ext 25 ./input_${counter}.$ext
26 #end for 26 #end for
27 -o out_dir 27 -o out_dir
28 -t \${GALAXY_SLOTS:-4} 28 -t \${GALAXY_SLOTS:-4}
29 -i $iterations 29 -i $iterations
30 #if $hifi_error: 30 #if $mode_conditional.mode == '--pacbio-hifi' and $mode_conditional.hifi_error:
31 --hifi-error $hifi_error 31 --hifi-error $mode_conditional.hifi_error
32 #end if 32 #end if
33 #if $min_overlap: 33 #if $min_overlap:
34 -m $min_overlap 34 -m $min_overlap
35 #end if 35 #end if
36 #if $asm.asm_select == 'true': 36 #if $asm.asm_select == 'true':
37 --asm-coverage $asm.asm_coverage 37 --asm-coverage $asm.asm_coverage
38 -g '${asm.genome_size}' 38 -g '${asm.genome_size}'
39 #end if 39 #end if
40 $plasmids
41 $meta 40 $meta
42 $trestle 41 $scaffold
43 ]]></command> 42 ]]></command>
44 <inputs> 43 <inputs>
45 <param name="inputs" type="data" format="fasta,fasta.gz,fastq,fastq.gz,fastqsanger.gz,fastqsanger" multiple="true" label="Input reads" /> 44 <param name="inputs" type="data" format="fasta,fasta.gz,fastq,fastq.gz,fastqsanger.gz,fastqsanger" multiple="true" label="Input reads" />
46 <param name="mode" type="select" label="Mode"> 45 <conditional name="mode_conditional">
47 <option value="--nano-raw">Nanopore raw</option> 46 <param name="mode" type="select" label="Mode">
48 <option value="--nano-corr">Nanopore corrected</option> 47 <option value="--nano-raw">Nanopore raw (--nano-raw)</option>
49 <option value="--pacbio-hifi">PacBio HiFi</option> 48 <option value="--nano-corr">Nanopore corrected (--nano-corr)</option>
50 <option value="--pacbio-raw">PacBio raw</option> 49 <option value="--nano-hq">Nanopore HQ (--nano-hq)</option>
51 <option value="--pacbio-corr">PacBio corrected</option> 50 <option value="--pacbio-raw">PacBio raw (--pacbio-raw)</option>
52 <option value="--subassemblies">High-quality contig-like input</option> 51 <option value="--pacbio-corr">PacBio corrected (--pacbio-corr)</option>
53 </param> 52 <option value="--pacbio-hifi">PacBio HiFi (--pacbio-hifi)</option>
54 <param argument="--iterations" type="integer" value="0" label="Number of polishing iterations" 53 </param>
54 <when value="--nano-raw"/>
55 <when value="--nano-corr"/>
56 <when value="--nano-hq"/>
57 <when value="--pacbio-raw"/>
58 <when value="--pacbio-corr"/>
59 <when value="--pacbio-hifi">
60 <param argument="--hifi-error" type="float" min="0" max="1" optional="true" label="Expected HiFi reads error rate" help="Default: 0.01"/>
61 </when>
62 </conditional>
63 <param argument="--iterations" type="integer" value="1" label="Number of polishing iterations"
55 help="Polishing is performed as the final assembly stage. By default, Flye runs one polishing iteration. Additional iterations 64 help="Polishing is performed as the final assembly stage. By default, Flye runs one polishing iteration. Additional iterations
56 might correct a small number of extra errors (due to improvements on how reads may align to the corrected assembly). If the 65 might correct a small number of extra errors (due to improvements on how reads may align to the corrected assembly). If the
57 parameter is set to 0, the polishing is not performed."/> 66 parameter is set to 0, the polishing is not performed"/>
58 <param argument="--min-overlap" type="integer" optional="true" label="Minimum overlap between reads" 67 <param argument="--min-overlap" type="integer" min="1000" max="10000" optional="true" label="Minimum overlap between reads"
59 help="This sets a minimum overlap length for two reads to be considered overlapping. By default it is chosen 68 help="This sets a minimum overlap length for two reads to be considered overlapping. By default it is chosen
60 automatically based on the read length distribution (reads N90) and does not require manual setting. Typical 69 automatically based on the read length distribution (reads N90) and does not require manual setting. Typical
61 value is 3k-5k (and down to 1k for datasets with shorter read length). Intuitively, we want to set this 70 value is 3k-5k (and down to 1k for datasets with shorter read length). Intuitively, we want to set this
62 parameter as high as possible, so the repeat graph is less tangled. However, higher values might lead to assembly gaps. 71 parameter as high as possible, so the repeat graph is less tangled. However, higher values might lead to assembly gaps.
63 In some rare cases it makes sense to manually increase minimum overlap for assemblies of big genomes with long reads and high coverage." /> 72 In some rare cases it makes sense to manually increase minimum overlap for assemblies of big genomes with long reads and high coverage." />
64 <param argument="--hifi-error" type="float" min="0" max="1" optional="true" label="Expected HiFi reads error rate" help="Default: 0.01"/>
65 <param argument="--plasmids" type="boolean" truevalue="--plasmids" falsevalue="" checked="False" label="Rescue short unassembled plasmids" />
66 <param argument="--keep-haplotypes" type="boolean" truevalue="--keep-haplotypes" falsevalue="" checked="False" label="Keep haplotypes" 73 <param argument="--keep-haplotypes" type="boolean" truevalue="--keep-haplotypes" falsevalue="" checked="False" label="Keep haplotypes"
67 help="By default, Flye collapses graph structures caused by alternative haplotypes (bubbles, superbubbles, roundabouts) to produce longer 74 help="By default, Flye collapses graph structures caused by alternative haplotypes (bubbles, superbubbles, roundabouts) to produce longer
68 consensus contigs. This option retains the alternative paths on the graph, producing less contigouos, but more detailed assembly."/> 75 consensus contigs. This option retains the alternative paths on the graph, producing less contigouos, but more detailed assembly."/>
69 <param argument="--trestle" type="boolean" truevalue="--trestle" falsevalue="" 76 <param argument="--scaffold" type="boolean" truevalue="--scaffold" falsevalue="" label="Enable scaffolding using graph"
70 checked="False" label="Enable Trestle" 77 help="Starting from the version 2.9 Flye does not perform scaffolding by default, which guarantees that all assembled sequences do not have any gaps" />
71 help="Trestle is an extra module that resolves simple repeats of multipicity 2 that were not bridged by reads. Depending on the datasets, it might
72 resolve a few extra repeats, which is helpful for small (bacterial genomes). On large genomes, the contiguity improvements are usually minimal,
73 but the computation might take a lot of time" />
74 <param argument="--meta" type="boolean" truevalue="--meta" falsevalue="" checked="False" label="Perform metagenomic assembly" 78 <param argument="--meta" type="boolean" truevalue="--meta" falsevalue="" checked="False" label="Perform metagenomic assembly"
75 help="It is designed for highly non-uniform coverage and is sensitive to underrepresented sequence at low coverage (as low as 2x). 79 help="It is designed for highly non-uniform coverage and is sensitive to underrepresented sequence at low coverage (as low as 2x).
76 In some examples of simple metagenomes, we observed that the normal mode assembled more contigious bacterial 80 In some examples of simple metagenomes, we observed that the normal mode assembled more contigious bacterial
77 consensus sequence, while the metagenome mode was slightly more fragmented, but revealed strain mixtures"/> 81 consensus sequence, while the metagenome mode was slightly more fragmented, but revealed strain mixtures"/>
78 <conditional name="asm"> 82 <conditional name="asm">
79 <param name="asm_select" type="select" label="Reduced contig assembly coverage"> 83 <param name="asm_select" type="select" label="Reduced contig assembly coverage" help="Typically, assemblies of large genomes at high coverage require a hundreds of RAM. For high coverage assemblies,
84 you can reduce memory usage by using only a subset of longest reads for initial contig extension stage (usually, the memory bottleneck)">
80 <option value="true">Enable reduced coverage for initial disjointing assembly</option> 85 <option value="true">Enable reduced coverage for initial disjointing assembly</option>
81 <option value="false" selected="true">Disable reduced coverage for initial disjointing assembly</option> 86 <option value="false" selected="true">Disable reduced coverage for initial disjointing assembly</option>
82 </param> 87 </param>
83 <when value="true"> 88 <when value="true">
84 <param argument="--asm-coverage" type="integer" min="0" value="30" 89 <param argument="--asm-coverage" type="integer" min="0" value="30"
85 label="Reduced coverage for initial disjointing assembly" 90 label="Reduced coverage for initial disjointing assembly"
86 help="Typically, assemblies of large genomes at high coverage require a hundreds of RAM. For high coverage assemblies, 91 help="This parameter specifies the target coverage of the longest reads. For a typical assembly, 30x is enough to produce good
87 you can reduce memory usage by using only a subset of longest reads for initial contig extension stage (usually, the memory bottleneck).
88 The parameter --asm-coverage specifies the target coverage of the longest reads. For a typical assembly, 30x is enough to produce good
89 initial contigs. Regardless of this parameter, all reads will be used at the later pipeline stages."/> 92 initial contigs. Regardless of this parameter, all reads will be used at the later pipeline stages."/>
90 <param argument="--genome-size" type="text" optional="true" label="Estimated genome size" 93 <param argument="--genome-size" type="text" optional="true" label="Estimated genome size"
91 help="For example, 5m or 2.6g. No longer required as input. However, it must be used in conjunction with --asm-coverage option."> 94 help="For example, 5m or 2.6g. No longer required as input. However, it must be used in conjunction with --asm-coverage option.">
92 <validator type="regex" message="Genome size must be a float or integer, optionally followed by the a unit prefix (kmg)">^([0-9]*[.])?[0-9]+[kmg]?$</validator> 95 <validator type="regex" message="Genome size must be a float or integer, optionally followed by the a unit prefix (kmg)">^([0-9]*[.])?[0-9]+[kmg]?$</validator>
93 </param> 96 </param>
104 <data name="flye_log" format="txt" from_work_dir="out_dir/flye.log" label="${tool.name} on ${on_string}: log"> 107 <data name="flye_log" format="txt" from_work_dir="out_dir/flye.log" label="${tool.name} on ${on_string}: log">
105 <filter>generate_log</filter> 108 <filter>generate_log</filter>
106 </data> 109 </data>
107 </outputs> 110 </outputs>
108 <tests> 111 <tests>
109 <!--Test 01--> 112 <!--Test 01: pacbio-raw-->
110 <test expect_num_outputs="5"> 113 <test expect_num_outputs="5">
111 <param name="inputs" ftype="fastq.gz" value="ecoli_01.fastq.gz,ecoli_02.fastq.gz,ecoli_03.fastq.gz,ecoli_04.fastq.gz,ecoli_05.fastq.gz,ecoli_06.fastq.gz,ecoli_07.fastq.gz"/> 114 <param name="inputs" ftype="fastq.gz" value="ecoli_01.fastq.gz,ecoli_02.fastq.gz,ecoli_03.fastq.gz,ecoli_04.fastq.gz,ecoli_05.fastq.gz,ecoli_06.fastq.gz,ecoli_07.fastq.gz"/>
112 <param name="mode" value="--pacbio-raw"/> 115 <param name="mode" value="--pacbio-raw"/>
116 <param name="iterations" value="0"/>
113 <param name="generate_log" value="true"/> 117 <param name="generate_log" value="true"/>
114 <output name="assembly_info" file="result1_assembly_info.txt" ftype="tabular" compare="sim_size"/> 118 <output name="assembly_info" file="result1_assembly_info.txt" ftype="tabular" compare="sim_size"/>
115 <output name="assembly_graph" file="result1_assembly_graph.dot" ftype="graph_dot" compare="sim_size"/> 119 <output name="assembly_graph" file="result1_assembly_graph.dot" ftype="graph_dot" compare="sim_size"/>
116 <output name="assembly_gfa" file="result1_assembly_graph.gfa" ftype="txt" compare="sim_size"/> 120 <output name="assembly_gfa" file="result1_assembly_graph.gfa" ftype="txt" compare="sim_size"/>
117 <output name="consensus" file="result1_assembly.fasta" ftype="fasta" compare="sim_size"/> 121 <output name="consensus" file="result1_assembly.fasta" ftype="fasta" compare="sim_size"/>
118 <output name="flye_log" file="result1.log" ftype="txt" compare="sim_size"/> 122 <output name="flye_log" file="result1.log" ftype="txt" compare="sim_size"/>
119 </test> 123 </test>
120 <!--Test 02--> 124 <!--Test 02: nano raw-->
121 <test expect_num_outputs="4"> 125 <test expect_num_outputs="4">
122 <param name="inputs" ftype="fasta.gz" value="nanopore.fasta.gz"/> 126 <param name="inputs" ftype="fasta.gz" value="nanopore.fasta.gz"/>
123 <param name="mode" value="--nano-raw"/> 127 <param name="mode" value="--nano-raw"/>
128 <param name="iterations" value="0"/>
124 <output name="assembly_info" ftype="tabular"> 129 <output name="assembly_info" ftype="tabular">
125 <assert_contents> 130 <assert_contents>
126 <has_size value="95" delta="100"/> 131 <has_size value="95" delta="100"/>
127 </assert_contents> 132 </assert_contents>
128 </output> 133 </output>
140 <assert_contents> 145 <assert_contents>
141 <has_size value="35573" delta="100"/> 146 <has_size value="35573" delta="100"/>
142 </assert_contents> 147 </assert_contents>
143 </output> 148 </output>
144 </test> 149 </test>
145 <!--Test 03--> 150 <!--Test 03: reduce coverage-->
146 <test expect_num_outputs="4"> 151 <test expect_num_outputs="4">
147 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/> 152 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/>
148 <param name="mode" value="--pacbio-hifi"/> 153 <conditional name="mode_conditional">
149 <param name="iterations" value="1"/> 154 <param name="mode" value="--nano-raw"/>
155 </conditional>
150 <conditional name="asm"> 156 <conditional name="asm">
151 <param name="asm_select" value="true" /> 157 <param name="asm_select" value="true" />
152 <param name="asm" value="100"/> 158 <param name="asm" value="30"/>
153 <param name="genome_size" value="3980000"/> 159 <param name="genome_size" value="3980000"/>
154 </conditional> 160 </conditional>
155 <output name="assembly_info" ftype="tabular"> 161 <output name="assembly_info" ftype="tabular">
156 <assert_contents> 162 <assert_contents>
157 <has_size value="286" delta="100"/> 163 <has_size value="286" delta="100"/>
158 </assert_contents> 164 </assert_contents>
159 </output> 165 </output>
160 <output name="assembly_graph" ftype="graph_dot"> 166 <output name="assembly_graph" ftype="graph_dot">
161 <assert_contents> 167 <assert_contents>
162 <has_size value="2135" delta="100"/> 168 <has_size value="1840" delta="100"/>
163 </assert_contents> 169 </assert_contents>
164 </output> 170 </output>
165 <output name="assembly_gfa" ftype="txt"> 171 <output name="assembly_gfa" ftype="txt">
166 <assert_contents> 172 <assert_contents>
167 <has_size value="114351" delta="100"/> 173 <has_size value="420752" delta="100"/>
168 </assert_contents> 174 </assert_contents>
169 </output> 175 </output>
170 <output name="consensus" ftype="fasta"> 176 <output name="consensus" ftype="fasta">
171 <assert_contents> 177 <assert_contents>
172 <has_size value="116191" delta="100"/> 178 <has_size value="427580" delta="100"/>
173 </assert_contents> 179 </assert_contents>
174 </output> 180 </output>
175 </test> 181 </test>
176 <!--Test 04--> 182 <!--Test 04: metagenomic mode-->
177 <test expect_num_outputs="4"> 183 <test expect_num_outputs="4">
178 <param name="inputs" ftype="fastq.gz" value="ecoli_01.fastq.gz,ecoli_02.fastq.gz,ecoli_03.fastq.gz,ecoli_04.fastq.gz,ecoli_05.fastq.gz,ecoli_06.fastq.gz,ecoli_07.fastq.gz"/> 184 <param name="inputs" ftype="fastq.gz" value="ecoli_01.fastq.gz,ecoli_02.fastq.gz,ecoli_03.fastq.gz,ecoli_04.fastq.gz,ecoli_05.fastq.gz,ecoli_06.fastq.gz,ecoli_07.fastq.gz"/>
179 <param name="mode" value="--pacbio-raw"/> 185 <conditional name="mode_conditional">
180 <param name="iterations" value="1"/> 186 <param name="mode" value="--pacbio-raw"/>
187 </conditional>
181 <param name="meta" value="true"/> 188 <param name="meta" value="true"/>
182 <param name="plasmids" value="true"/>
183 <output name="assembly_info" ftype="tabular"> 189 <output name="assembly_info" ftype="tabular">
184 <assert_contents> 190 <assert_contents>
185 <has_size value="95" delta="100"/> 191 <has_size value="95" delta="100"/>
186 </assert_contents> 192 </assert_contents>
187 </output> 193 </output>
190 <has_size value="367" delta="100"/> 196 <has_size value="367" delta="100"/>
191 </assert_contents> 197 </assert_contents>
192 </output> 198 </output>
193 <output name="assembly_gfa" ftype="txt"> 199 <output name="assembly_gfa" ftype="txt">
194 <assert_contents> 200 <assert_contents>
195 <has_size value="418051" delta="100"/> 201 <has_size value="418729" delta="100"/>
196 </assert_contents> 202 </assert_contents>
197 </output> 203 </output>
198 <output name="consensus" ftype="fasta"> 204 <output name="consensus" ftype="fasta">
199 <assert_contents> 205 <assert_contents>
200 <has_size value="425000" delta="100"/> 206 <has_size value="425667" delta="100"/>
201 </assert_contents> 207 </assert_contents>
202 </output> 208 </output>
203 </test> 209 </test>
204 <!--Test 05--> 210 <!--Test 05: nanopore HQ mode-->
205 <test expect_num_outputs="4"> 211 <test expect_num_outputs="4">
206 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/> 212 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/>
207 <param name="mode" value="--pacbio-hifi"/> 213 <conditional name="mode_conditional">
208 <param name="iterations" value="1"/> 214 <param name="mode" value="--nano-hq"/>
215 </conditional>
216 <param name="min_overlap" value="1000"/>
209 <output name="assembly_info" ftype="tabular"> 217 <output name="assembly_info" ftype="tabular">
210 <assert_contents> 218 <assert_contents>
211 <has_size value="286" delta="100"/> 219 <has_size value="286" delta="100"/>
212 </assert_contents> 220 </assert_contents>
213 </output> 221 </output>
214 <output name="assembly_graph" ftype="graph_dot"> 222 <output name="assembly_graph" ftype="graph_dot">
215 <assert_contents> 223 <assert_contents>
216 <has_size value="2135" delta="100"/> 224 <has_size value="1248" delta="100"/>
217 </assert_contents> 225 </assert_contents>
218 </output> 226 </output>
219 <output name="assembly_gfa" ftype="txt"> 227 <output name="assembly_gfa" ftype="txt">
220 <assert_contents> 228 <assert_contents>
221 <has_size value="114351" delta="100"/> 229 <has_size value="420252" delta="100"/>
222 </assert_contents> 230 </assert_contents>
223 </output> 231 </output>
224 <output name="consensus" ftype="fasta"> 232 <output name="consensus" ftype="fasta">
225 <assert_contents> 233 <assert_contents>
226 <has_size value="116191" delta="100"/> 234 <has_size value="427129" delta="100"/>
227 </assert_contents> 235 </assert_contents>
228 </output> 236 </output>
229 </test> 237 </test>
230 <!--Test 06--> 238 <!--Test 06: hifi error option-->
231 <test expect_num_outputs="4"> 239 <test expect_num_outputs="4">
232 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/> 240 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/>
233 <param name="mode" value="--pacbio-hifi"/> 241 <conditional name="mode_conditional">
234 <param name="iterations" value="1"/> 242 <param name="mode" value="--pacbio-hifi"/>
235 <param name="hifi-error" value="0.02"/> 243 <param name="hifi_error" value="0.21"/>
244 </conditional>
245 <param name="min_overlap" value="1000"/>
236 <output name="assembly_info" ftype="tabular"> 246 <output name="assembly_info" ftype="tabular">
237 <assert_contents> 247 <assert_contents>
238 <has_size value="286" delta="100"/> 248 <has_size value="286" delta="100"/>
239 </assert_contents> 249 </assert_contents>
240 </output> 250 </output>
241 <output name="assembly_graph" ftype="graph_dot"> 251 <output name="assembly_graph" ftype="graph_dot">
242 <assert_contents> 252 <assert_contents>
243 <has_size value="2135" delta="100"/> 253 <has_size value="1273" delta="100"/>
244 </assert_contents> 254 </assert_contents>
245 </output> 255 </output>
246 <output name="assembly_gfa" ftype="txt"> 256 <output name="assembly_gfa" ftype="txt">
247 <assert_contents> 257 <assert_contents>
248 <has_size value="114351" delta="100"/> 258 <has_size value="420252" delta="100"/>
249 </assert_contents> 259 </assert_contents>
250 </output> 260 </output>
251 <output name="consensus" ftype="fasta"> 261 <output name="consensus" ftype="fasta">
252 <assert_contents> 262 <assert_contents>
253 <has_size value="116191" delta="100"/> 263 <has_size value="427129" delta="100"/>
254 </assert_contents> 264 </assert_contents>
255 </output> 265 </output>
256 </test> 266 </test>
257 <!--Test 07--> 267 <!--Test 07: keep haplotypes-->
258 <test expect_num_outputs="4"> 268 <test expect_num_outputs="4">
259 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/> 269 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/>
260 <param name="mode" value="--pacbio-hifi"/> 270 <conditional name="mode_conditional">
261 <param name="iterations" value="1"/> 271 <param name="mode" value="--pacbio-corr"/>
272 <param name="hifi_error" value="0.21"/>
273 </conditional>
274 <param name="min_overlap" value="1000"/>
262 <param name="keep-haplotypes" value="true"/> 275 <param name="keep-haplotypes" value="true"/>
263 <output name="assembly_info" ftype="tabular"> 276 <output name="assembly_info" ftype="tabular">
264 <assert_contents> 277 <assert_contents>
265 <has_size value="286" delta="100"/> 278 <has_size value="286" delta="100"/>
266 </assert_contents> 279 </assert_contents>
267 </output> 280 </output>
268 <output name="assembly_graph" ftype="graph_dot"> 281 <output name="assembly_graph" ftype="graph_dot">
269 <assert_contents> 282 <assert_contents>
270 <has_size value="2135" delta="100"/> 283 <has_size value="1273" delta="100"/>
271 </assert_contents> 284 </assert_contents>
272 </output> 285 </output>
273 <output name="assembly_gfa" ftype="txt"> 286 <output name="assembly_gfa" ftype="txt">
274 <assert_contents> 287 <assert_contents>
275 <has_size value="114351" delta="100"/> 288 <has_size value="420252" delta="100"/>
276 </assert_contents> 289 </assert_contents>
277 </output> 290 </output>
278 <output name="consensus" ftype="fasta"> 291 <output name="consensus" ftype="fasta">
279 <assert_contents> 292 <assert_contents>
280 <has_size value="116191" delta="100"/> 293 <has_size value="427129" delta="100"/>
294 </assert_contents>
295 </output>
296 </test>
297 <!--Test 08: scaffolding mode-->
298 <test expect_num_outputs="4">
299 <param name="inputs" ftype="fastq.gz" value="ecoli_hifi_01.fastq.gz,ecoli_hifi_02.fastq.gz,ecoli_hifi_03.fastq.gz,ecoli_hifi_04.fastq.gz,ecoli_hifi_05.fastq.gz,ecoli_hifi_06.fastq.gz,ecoli_hifi_07.fastq.gz,ecoli_hifi_08.fastq.gz,ecoli_hifi_09.fastq.gz"/>
300 <param name="mode" value="--nano-hq"/>
301 <param name="min_overlap" value="1000"/>
302 <param name="scaffolding" value="true"/>
303 <output name="assembly_info" ftype="tabular">
304 <assert_contents>
305 <has_size value="286" delta="100"/>
306 </assert_contents>
307 </output>
308 <output name="assembly_graph" ftype="graph_dot">
309 <assert_contents>
310 <has_size value="1248" delta="100"/>
311 </assert_contents>
312 </output>
313 <output name="assembly_gfa" ftype="txt">
314 <assert_contents>
315 <has_size value="420252" delta="100"/>
316 </assert_contents>
317 </output>
318 <output name="consensus" ftype="fasta">
319 <assert_contents>
320 <has_size value="427129" delta="100"/>
281 </assert_contents> 321 </assert_contents>
282 </output> 322 </output>
283 </test> 323 </test>
284 </tests> 324 </tests>
285 <help><![CDATA[ 325 <help><![CDATA[
299 339
300 **Quick usage** 340 **Quick usage**
301 341
302 Input reads can be in FASTA or FASTQ format, uncompressed or compressed with gz. Currently, PacBio (raw, corrected, HiFi) and ONT reads 342 Input reads can be in FASTA or FASTQ format, uncompressed or compressed with gz. Currently, PacBio (raw, corrected, HiFi) and ONT reads
303 (raw, corrected) are supported. Expected error rates are <30% for raw, <3% for corrected, and <1% for HiFi. Note that Flye was primarily 343 (raw, corrected) are supported. Expected error rates are <30% for raw, <3% for corrected, and <1% for HiFi. Note that Flye was primarily
304 developed to run on raw reads. Additionally, the *--subassemblies* option performs a consensus assembly of multiple sets of high-quality 344 developed to run on raw reads. You may specify multiple files with reads (separated by spaces). Mixing different read types is not yet supported. The *--meta* o
305 contigs. You may specify multiple files with reads (separated by spaces). Mixing different read types is not yet supported. The *--meta* o
306 ption enables the mode for metagenome/uneven coverage assembly. 345 ption enables the mode for metagenome/uneven coverage assembly.
307 346
308 Genome size estimate is no longer a required option. You need to provide an estimate if using *--asm-coverage* option. 347 Genome size estimate is no longer a required option. You need to provide an estimate if using *--asm-coverage* option.
309 348
310 To reduce memory consumption for large genome assemblies, you can use a subset of the longest reads for initial disjointig assembly by 349 To reduce memory consumption for large genome assemblies, you can use a subset of the longest reads for initial disjointig assembly by