<tool id="wtdbg" name="WTDBG" version="2.0">
    <description>De novo assembler AND consensuser for long noisy sequences</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <version_command>wtdbg2 -help | grep 'Version:'</version_command>
    <command detect_errors="exit_code"><![CDATA[
    wtdbg2
        -t \${GALAXY_SLOTS:-4} 
        -i '$i'
        -o 'dbg' 
        #if $I:
            -I '$I'
        #end if
        #if $load_alignments:
            --load-alignments '$load_alignments'
        #end if

        -k $k
        -p $p
        -K $K
        -E $E
        $F
        -S $S
        -X $X
        -Y $Y
        -x $x
        -y $y
        -l $l
        -m $m
        -s $s
        --tidy-reads $tidy_reads 
        --edge-min $edge_min 
        $rescue_low_cov_edges
    &&    
    wtdbg-cns 
        -t \${GALAXY_SLOTS:-4} 
        -o dbg.ctg.lay.fa
        -i dbg.ctg.lay
        -j $cns.j
        -k $cns.k
        -Z $cns.Z
        -W $cns.W
        -H $cns.H
        -L $cns.L
        -c $cns.c
        -M $cns.M
        -X $cns.X
        -I $cns.I
        -D $cns.D
        -E $cns.E
        -m $cns.m
        -S $cns.S

    ]]></command>
    <inputs>
        <param type="data" argument="-i" format="fasta,fasta.gz" label="Long reads sequences file"/>
        <param type="data" argument="-I" format="fasta,fasta.gz" optional="True" label="Error-free sequences file"/>
        <param type="data" argument="--load-alignments" name="load_alignments" format="tabular" optional="True" label="Load pre-computed alignments"/>

        <param argument="k" type="integer" value="0" min="0" max="25" label="Kmer fsize" />
        <param argument="p" type="integer" value="21" min="0" max="25" label="Kmer psize" />
        <param argument="K" type="float" value="1000" min="0" max="65535" label="Filter high frequency kmers" />
        <param argument="E" type="integer" value="2" label="Min kmer frequency" />
        <param argument="F" type="boolean" truevalue="-F" falsevalue="" checked="False" label="Filter low frequency kmers by a 4G-bytes array" />
        <param argument="S" type="integer" value="4" label="Subsampling kmers, 1/S kmers are indexed" />
        <param argument="X" type="integer" value="4" label="Max number of bin (256bp) in one gap" />
        <param argument="Y" type="integer" value="4" label="Max number of bin (256bp) in one deviation" />
        <param argument="x" type="integer" value="-7" label="penalty for BIN gap" />
        <param argument="y" type="integer" value="-21" label="penalty for BIN deviation" />
        <param argument="l" type="float" value="2048" min="1" label="Min length of alignment" />
        <param argument="m" type="float" value="200" label="Min matched" />
        <param argument="s" type="float" value="0.2" label="Max length variation of two aligned fragments" />

        <param argument="--tidy-reads" name="tidy_reads" type="integer" value="0" label="Filter reads less than tidy-reads" />
        <param argument="--edge-min" name="edge_min" type="integer" value="3" label="The minimal depth of a valid edge set to" />
        <param argument="--rescue-low-cov-edges" name="rescue_low_cov_edges" type="boolean" truevalue="--rescue-low-cov-edges" 
            falsevalue="" label="Try to rescue low coverage edges" />

        <section name="cns" title="Consensus options">
            <!-- optional inputs -->
            <!-- <param argument="-i" type="data" format="utg.cns" label="Input file(s) *.utg.cns" /> -->

            <param argument="-j" type="integer" value="1000" label="Expected length of node" />
            <param argument="-k" type="integer" value="15" label="Kmer size for long reads" />
            <param argument="-Z" type="integer" value="4" label="Z-cutoff, drop the lower" />
            <param argument="-W" type="integer" value="48" label="W-cutoff, drop the lagger (position)" />
            <param argument="-H" type="integer" value="1" label="High coverage bonus" />
            <param argument="-L" type="integer" value="10" label="High coverage cutoff" />
            <param argument="-c" type="select" label="Candidate strategy">
                <option value="0" selected="true">best-kmers</option>
                <option value="1" >median length</option>
                <option value="2" >first (include)</option>
                <option value="3" >first (exclude)</option>
                <option value="4" >longest</option>
                <option value="5" >shortest</option>
            </param>

            <param argument="-M" type="integer" value="2" label="Match score" />
            <param argument="-X" type="integer" value="-7" label="Mismatch score" />
            <param argument="-I" type="integer" value="-3" label="Insertion score" />
            <param argument="-D" type="integer" value="-4" label="Deletion score" />
            <param argument="-E" type="integer" value="-2" label="Gap extension score" />
            <param argument="-m" type="select" label="Correction mode">
                <option value="1" selected="true">DBG correction</option>
                <option value="2" >DAG correction</option>
            </param>
            <param argument="-S" type="integer" value="1" label="Correct structure before error correction" />
        </section>

    </inputs>
    <outputs>
        <data name="output_alignments" format="fasta" label="${tool.name}  alignments" from_work_dir="dbg.alignments" />
        <data name="output_ctglay" format="txt" label="${tool.name}  contigs layout" from_work_dir="dbg.ctg.lay" />
        <data name="output_consensus" format="fasta" label="${tool.name} consensus" from_work_dir="dbg.ctg.lay.fa" />
    </outputs>
    <tests>
        <test>
            <param name="i" value="ecoli-reads.fa"/>
            <output name="output_alignments" file="result1.alignments"/>
            <output name="output_ctglay" file="result1.ctg.lay"/>
            <output name="output_consensus" file="consensus_result1.fa"/>
        </test>
        <test>
            <param name="i" value="ecoli-reads.fa"/>
            <param name="tidy_reads" value="5000"/>
            <param name="edge_min" value="2"/>
            <param name="rescue_low_cov_edges" value="True"/>
            <output name="output_consensus" file="consensus_result2.fa"/>
        </test>
        <test>
            <param name="i" value="ecoli-reads.fa"/>
            <param name="cns.c" value="1"/>
            <param name="cns.E" value="-3"/>
            <param name="cns.j" value="500"/>
            <param name="cns.m" value="2"/>
            <param name="cns.k" value="5"/>
            <output name="output_consensus" file="consensus_result3.fa"/>
        </test>
    
    </tests>
    
    <help><![CDATA[
**What it does**

WTDBG is a de novo assembler for long noisy sequences, based on fuzzy Bruijn graphs (FBG).

**Alignment**

KBM (Kmer-BIN-Mapping) groups k-mers from each non-overlapped sliding 256 bp fragments in long reads into bins.
Bins of which most k-mers are high frequency, are filtered as highly repetitive ones.
Then, KBM searches synteny of matched bin pairs in sequences in a dynamic programming way.
A matched bin pair in two sequences is defined as two bins different by original but share a set of k-mers.
The result of alignments in KBM have the same features of traditional sequence alignment, excepting the unit of
KBM alignments is 256 bp bin instead of single base.

**Assembly**

FBG (Fuzzy Bruijn Graph) is composed of vertices in length of 1024 bp from reads, and edges connecting vertices
in their order on read paths. Comparing with DBG, the size of vertices in FBG are much bigger, thus won't be
sensitive to small repeat. To tolerate high sequencing errors, FBG's vertices are found using gapped
sequence alignments from KBM or other aligners, comparing with searching identical k-mers in DBG.

  ]]></help>
    <expand macro="citations" />
</tool>