Mercurial > repos > stheil > metavelvet_wrapper

view meta_velvet.xml @ 4:34ea8f113018 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author stheil
date Thu, 24 Sep 2015 10:53:51 -0400
parents 0591d844bd08
children
line wrap: on
line source

<tool id="meta-velvetg" name="meta-velvetg" version="3.0 (2013-4-30)">

	<macros>
		<import>velvet_macro.xml</import>
	</macros>
	<description>Run velvet or meta-velvetg assembly.</description>

  <command interpreter="perl">velvet.pl

    $meta

		#for $i, $library in enumerate( $shortPaired )

			#set $libNumber = $i + 1

			#if $libNumber == 1

				#set $libNumber = ''

			#end if

			#if not $library.ins_length.value is None

			-ins_length$libNumber $library.ins_length

			#end if

			#if not $library.ins_length_sd.value is None

			-ins_length_sd$libNumber $library.ins_length_sd

			#end if

      -shortPaired$libNumber

			@PAIRED_FILES_OPTIONS@

		#end for

		#for $i, $library in enumerate( $short )

			#set $libNumber = $i + 1

			#if $libNumber == 1

				#set $libNumber = ''

			#end if

			-short$libNumber

			@UNPAIRED_FILES_OPTIONS@

		#end for

		#for $i, $library in enumerate( $longPaired )

			#if not $library.ins_length.value is None

			-ins_length_long $library.ins_length

			#end if

			#if not $library.ins_length_sd.value is None

			-ins_length_long_sd $library.ins_length_sd

			#end if

      -longPaired

			@PAIRED_FILES_OPTIONS@

		#end for

		#for $i, $library in enumerate( $long )

			-long

			@UNPAIRED_FILES_OPTIONS@

		#end for

    #if not $hash_length.value == ''

    -hash_length $hash_length

    #end if

    $strand_specific

    #if not $cov_cutoff.value == ''

    -cov_cutoff $cov_cutoff

    #end if

    $read_trkg

    #if not $min_contig_lgth.value is None

		-min_contig_lgth $min_contig_lgth

		#end if

    $amos_file

    #if not $exp_cov.value == ''

		-exp_cov $exp_cov

		#end if

    #if not $long_cov_cutoff.value == ''

		-long_cov_cutoff $long_cov_cutoff

		#end if

    $scaffolding

    #if not $max_branch_length.value is None

		-max_branch_length $max_branch_length

		#end if

    #if not $max_divergence.value is None

		-max_divergence $max_divergence

		#end if

    #if not $max_gap_count.value is None

		-max_gap_count $max_gap_count

		#end if

    #if not $min_pair_count.value is None

		-min_pair_count $min_pair_count

		#end if

    #if not $max_coverage.value is None

		-max_coverage $max_coverage

		#end if

    #if not $coverage_mask.value is None

		-coverage_mask $coverage_mask

		#end if

    #if not $long_mult_cutoff.value is None

    -long_mult_cutoff $long_mult_cutoff

		#end if

    $unused_reads

    $alignments

    $exportFiltered

    #if not $paired_exp_fraction.value is None

    -paired_exp_fraction $paired_exp_fraction

		#end if

    $shortMatePaired

    $conserveLong

    $discard_chimera

    #if not $max_chimera_rate.value is None

    -max_chimera_rate $max_chimera_rate

		#end if

    #if not $repeat_cov_sd.value is None

    -repeat_cov_sd $repeat_cov_sd

		#end if

    #if not $min_split_length.value is None

    -min_split_length $min_split_length

		#end if

    #if not $valid_connections.value is None

    -valid_connections $valid_connections

		#end if

    #if not $noise_connections.value is None

    -noise_connections $noise_connections

		#end if

    $use_connections

    $report_split_detail

    $report_subgraph

    #if not $exp_covs.value == ''

    -exp_covs $exp_covs

		#end if

    #if not $min_peak_cov.value is None

    -min_peak_cov $min_peak_cov

		#end if

    #if not $max_peak_cov.value is None

    -max_peak_cov $max_peak_cov

		#end if

    #if not $histo_bin_width.value is None

    -histo_bin_width $histo_bin_width

		#end if

    #if not $histo_sn_ratio.value is None

    -histo_sn_ratio $histo_sn_ratio

		#end if

  </command>

  <inputs>

		<param name="hash_length" type="text" value="31" label="Hash length : integer &lt;= 150 OR m,M,s with m (min kmer length) &lt; M (max kmer length) &lt;= 150 and s is a step number"/>

    <param name="meta" type="select" label="Perform a metagenomic assembly using meta-velvetg">
			<option value="">Yes</option>
			<option value="-nometa">No</option>
		</param>

		<repeat name="shortPaired" max="2" title="Short paired reads library">
			<expand macro="paired_library_options" />
		</repeat>

		<repeat name="short" max="2" title="Short unpaired reads library">
			<expand macro="unpaired_library_options" />
		</repeat>

		<repeat name="longPaired" max="1" title="Long paired reads library">
			<expand macro="paired_library_options" />
		</repeat>

		<repeat name="long" max="1" title="Long unpaired reads library">
			<expand macro="unpaired_library_options" />
		</repeat>

		<param name="reference" type="data" format="fasta" optional="true" label="VELVETH - Reference file (-reference)"/>

    <param name="strand_specific" type="select" label="VELVETH - Strand specific transcriptome sequencing data. Default: No (-strand_specific)">
      <option value="">No</option>
      <option value="-strand_specific">Yes</option>
    </param>

    <param name="cov_cutoff" type="text" optional="true" label="VELVETG - Coverage cutoff. Float or auto (-cov_cutoff)"/>

    <param name="read_trkg" type="select" label="VELVETG - Tracking of short read positions in assembly. Default: no tracking (-read_trkg)">
      <option value="">No</option>
      <option value="-read_trkg yes">Yes</option>
    </param>

    <param name="min_contig_lgth" type="integer" optional="true" label="VELVETG - Minimum contig length. Default: hash length * 2 (-min_contig_lgth)"/>

    <param name="amos_file" type="select" label="VELVETG - Export assembly to AMOS file. Default: no export (-amos_file)">
      <option value="">No</option>
      <option value="-amos_file yes">Yes</option>
    </param>

    <param name="exp_cov" type="text" optional="true" label="VELVETG - Expected coverage of unique regions (float or auto). In case of metagenomic analysis it will be set to auto (-exp_cov)"/>
    <param name="long_cov_cutoff" type="text" optional="true" label="VELVETG - Cutoff for removal of nodes with low long-read coverage (-long_cov_cutoff)"/>

    <param name="scaffolding" type="select" label="VELVETG - Scaffolding of contigs used paired end information. Default: on (-scaffolding)">
			<option value="">Yes</option>
			<option value="-scaffolding no">No</option>
		</param>

    <param name="max_branch_length" type="integer" optional="true" label="VELVETG - Maximum length in base pair of bubble. Default: 100 (-max_branch_length)"/>
    <param name="max_divergence" type="float" optional="true" label="VELVETG - Maximum divergence rate between two branches in a bubble. Default: 0.2 (-max_divergence)"/>
    <param name="max_gap_count" type="integer" optional="true" label="VELVETG - Maximum number of gaps allowed in the alignment of the two branches of a bubble. Default: 3 (-max_gap_count)"/>
    <param name="min_pair_count" type="integer" optional="true" label="VELVETG -Minimum number of paired end connections to justify the scaffolding of two long contigs. Default: 5 (-min_pair_count)"/>

    <param name="max_coverage" type="float" optional="true" label="VELVETG - Cutoff for removal of high coverage nodes AFTER tour bus. Default: no removal (-max_coverage)"/>
    <param name="coverage_mask" type="integer" optional="true" label="VELVETG - Minimum coverage required for confident regions of contigs. Default: 1 (-coverage_mask)"/>
    <param name="long_mult_cutoff" type="integer" optional="true" label="VELVETG -Minimum number of long reads required to merge contigs. Default: 2 (-long_mult_cutoff)"/>

    <param name="unused_reads" type="select" label="VELVETG - Export unused reads in UnusedReads.fa file. Default: no (-unused_reads)">
      <option value="">No</option>
      <option value="-unused_reads yes">Yes</option>
    </param>

    <param name="alignments" type="select" label="VELVETG - Export a summary of contig alignment to the reference sequences. Default: no (-alignments)">
      <option value="">No</option>
      <option value="-alignments yes">Yes</option>
    </param>

    <param name="exportFiltered" type="select" label="VELVETG - Export the long nodes which were eliminated by the coverage filters. Default: no (-exportFiltered)">
      <option value="">No</option>
      <option value="-exportFiltered yes">Yes</option>
    </param>

    <param name="paired_exp_fraction" type="float" optional="true" label="VELVETG -Remove all the paired end connections which less than the specified fraction of the expected count. Default: 0.1 (-paired_exp_fraction)"/>

    <param name="shortMatePaired" type="select" label="VELVETG - For mate-pair libraries, indicate that the library might be contaminated with paired-end reads. Default no (-shortMatePaired)">
      <option value="">No</option>
      <option value="-shortMatePaired yes">Yes</option>
    </param>

    <param name="conserveLong" type="select" label="VELVETG - Preserve sequences with long reads in them. Default no (-conserveLong)">
      <option value="">No</option>
      <option value="-conserveLong yes">Yes</option>
    </param>

    <param name="discard_chimera" type="select" label="META-VELVETG - Discard chimera sub-graph. Default: no (-discard_chimera)">
      <option value="">No</option>
      <option value="-discard_chimera yes">Yes</option>
    </param>

    <param name="max_chimera_rate" type="float" optional="true" label="META-VELVETG - Maximum allowable chimera rate. Default: 0.0 (-max_chimera_rate)"/>
    <param name="repeat_cov_sd" type="float" optional="true" label="META-VELVETG - Standard deviation of repeat node coverages. Default: 0.1 (-repeat_cov_sd)"/>
    <param name="min_split_length" type="integer" optional="true" label="META-VELVETG - Minimum node length required for repeat resolution. Default: 0 (-min_split_length)"/>
    <param name="valid_connections" type="integer" optional="true" label="META-VELVETG - Minimum allowable number of consistent paired-end connections. Default: 1 (-valid_connections)"/>
    <param name="noise_connections" type="integer" optional="true" label="META-VELVETG - Maximum allowable number of inconsistent paired-end connections. Default: 0 (-noise_connections)"/>

    <param name="use_connections" type="select" label="META-VELVETG - Use paired-end connections for graph splitting. Default: yes (-use_connections)">
      <option value="">Yes</option>
      <option value="-use_connections no">No</option>
    </param>

    <param name="report_split_detail" type="select" label="META-VELVETG - Report sequences around repeat nodes. Default: no (-report_split_detail)">
      <option value="">No</option>
      <option value="-report_split_detail yes">Yes</option>
    </param>

    <param name="report_subgraph" type="select" label="META-VELVETG - Report node sequences for each subgraph. Default: no (-report_subgraph)">
      <option value="">No</option>
      <option value="-report_subgraph yes">Yes</option>
    </param>

    <param name="exp_covs" type="text" optional="true" label="META-VELVETG - Expected coverages for each species in microbiome. Default: auto (-exp_covs)"/>

    <param name="min_peak_cov" type="float" optional="true" label="META-VELVETG - Minimum peak coverage. Default: 0 (-min_peak_cov)"/>
    <param name="max_peak_cov" type="float" optional="true" label="META-VELVETG - Maximum peak coverage. Default: 500 (-max_peak_cov)"/>
    <param name="histo_bin_width" type="float" optional="true" label="META-VELVETG - Bin width of peak coverage histogram. Default: 1 (-histo_bin_width)"/>
    <param name="histo_sn_ratio" type="float" optional="true" label="META-VELVETG - Signal-noise ratio to remove peak noises. Default: 10 (-histo_sn_ratio)"/>

  </inputs>

	<outputs>

  <data format="fasta" name="meta_velvet_contigs" from_work_dir="meta-velvetg.contigs.fa" label="${tool.name} on ${on_string}: meta-velvetg.contigs.fasta">
    <filter>(meta == '')</filter>
  </data>

  <data format="fasta" name="velvet_contigs" from_work_dir="contigs.fa" label="${tool.name} on ${on_string}: contigs.fasta">
    <filter>(meta == '-nometa')</filter>
  </data>

  <data format="fasta" name="velvet_singlets" from_work_dir="Singlets.fasta" label="${tool.name} on ${on_string}: singlets.fasta">
    <filter>(unused_reads == '-unused_reads yes')</filter>
  </data>

  <data format="txt" name="meta-velvetg.asm.afg" from_work_dir="meta-velvetg.asm.afg" label="${tool.name} on ${on_string}: meta-velvetg.asm.afg">
    <filter>((meta == '') and (amos_file == '-amos_file yes'))</filter>
  </data>

  <data format="txt" name="velvet_asm.afg" from_work_dir="velvet_asm.afg" label="${tool.name} on ${on_string}: velvet_asm.afg">
    <filter>((meta == '-nometa') and (amos_file == '-amos_file yes'))</filter>
  </data>

  <data format="tabular" name="stats.txt" from_work_dir="stats.txt" label="${tool.name} on ${on_string}: stats.txt">
    <filter>(meta == '-nometa')</filter>
  </data>

  <data format="tabular" name="meta-velvetg.split-stats.txt" from_work_dir="meta-velvetg.split-stats.txt" label="${tool.name} on ${on_string}: meta-velvetg.split-stats.txt">
    <filter>(meta == '')</filter>
  </data>

  <data format="txt" name="LastGraph" from_work_dir="LastGraph" label="${tool.name} on ${on_string}: LastGraph">
    <filter>(meta == '-nometa')</filter>
  </data>

  <data format="txt" name="meta-velvetg.LastGraph" from_work_dir="meta-velvetg.LastGraph" label="${tool.name} on ${on_string}: meta-velvetg.LastGraph">
    <filter>(meta == '')</filter>
  </data>

  <data format="tabular" name="meta-velvetg.Graph2-stats.txt" from_work_dir="meta-velvetg.Graph2-stats.txt" label="${tool.name} on ${on_string}: meta-velvetg.Graph2-stats.txt">
    <filter>(meta == '')</filter>
  </data>

  </outputs>

</tool>