<tool id="rbc_rnacode" name="RNAcode" version="0.3.1">
    <description>Analyze the protein coding potential in MSA.</description>
    <requirements>
        <requirement type="package" version="0.3">rnacode</requirement>
    </requirements>
    <version_command>RNAcode --version</version_command>
    <command detect_errors="exit_code">
    <![CDATA[
        #if $cond_breakmaf.select_breakmaf == 'breakmaf'
            '$__tool_directory__/breakMAF.pl'
            --maxlength $cond_breakmaf.maxlength
            --desiredLength $cond_breakmaf.desiredlength
            < '$alignment' |  
        #else
            cat '$alignment' | 
        #end if

        #if $cond_breakmaf.select_breakmaf == 'breakmaf' and $cond_breakmaf.processseparately == 'yes'
            '$__tool_directory__/processMAF.sh'
        #else
            RNAcode
        #end if

        $outputFormat
        
        #if $cutoff and $cutoff is not None
            --cutoff $cutoff
        #end if
        
        #if $num_samples and $num_samples is not None
            --num-samples $num_samples
        #end if
        
        $stop_early
        $best_region
        $best_only
        
        #if $cond_generateEPS.generateEPS == 'create'
            --eps
            --eps-dir eps/ 
            #if $cond_generateEPS.eps_cutoff and $cond_generateEPS.eps_cutoff is not None
                --eps-cutoff $cond_generateEPS.eps_cutoff
            #end if
        #end if
        
        #if $cond_scoringParameters.scoringParameters == 'custom'
            --pars "$pars"
        #end if
        
        #if $outputFormat.value == '--tabular'
            --outfile $outFileDefault
        #elif $outputFormat.value == '--gtf'
            --outfile $outFileGTF
        #end if
        
    ]]>
    </command>
    <inputs>
        <param name="alignment" type="data" format="clustal,maf" label="Multiple Alignment" help="Alignment needs to be formatted in ClustalW or MAF format"/>
        <conditional name="cond_breakmaf">
            <param name="select_breakmaf" type="select" label="Break long alignment blocks" help="If your alignments contain blocks of long genomic regions it is usually not reasonable to score these long regions as a whole.">
                <option value="keepmaf" selected="true">Process original alignment</option>
                <option value="breakmaf">Break long alignment blocks</option>
            </param>
            <when value="breakmaf">
                <param argument="--maxLength" name="maxlength" type="integer" optional="true" value="400" label="maxLength" help="Break all blocks longer than that."/>
                <param argument="--desiredLength" name="desiredlength" type="integer" optional="true" value="200" label="desiredLength" help="Try to create blocks of this size."/>
                <param name="processseparately" type="boolean" value="false" truevalue="yes" falsevalue="no" label="process blocks separately" help="If enables RNAcode is called separately for each block. This might be a reasonable strategy if RNAcode fails for some blocks."/>
            </when>
            <when value="keepmaf"/>
        </conditional>
        <param argument="--cutoff" name="cutoff" type="float" optional="true" value="1.0" label="Cutoff" help="Show only regions that have a p-value below the given number. By default all hits are shown."/>
        <param argument="--num_samples" name="num_samples" type="integer" optional="true" value="100" label="Number of samples" help="Number of random alignments that are sampled to calculate the p-value. RNAcode estimates the significance of a coding prediction by sampling a given number of random alignments. Default is 100 which gives reasonably stable p-values that are useful for assessing the relevance of a prediction."/>
        <param argument="--stop_early" name="stop_early" type="boolean" truevalue="--stop-early" falsevalue="" checked="false" label="Stop early" help="Setting this option stops the sampling process as soon as it is clear that the best hit will not fall below the given p-value cutoff. For example, assume a p-value cutoff of 0.05 (see --cutoff) and a sample size of 1000 is given (see --num-samples). As soon as 50 random samples score better than the original alignment, the process is stopped and all hits in the original alignment are reported as p>0.05 (or by convention as 1.0 in gtf and tabular output)."/>
        <param argument="--best_region" name="best_region" type="boolean" truevalue="--best-region" falsevalue="" checked="false" label="Show only best non-overlapping hits" help="By default all positive scoring segments are shown in the output if they fall below the given p-value cutoff. If two hits overlap (different frame or different strand) and --best_region is given only the hit with the highest score is shown. Strong coding regions often lead to statistically significant signals also in other frames. These hits are suppressed by this option and only the correct reading frame is reported."/>
        <param argument="--best_only" name="best_only" type="boolean" truevalue="--best-only" falsevalue="" checked="false" label="Show only best hit" help="This options shows only the one single best hit for each alignment."/>
        <conditional name="cond_scoringParameters">
            <param name="scoringParameters" type="select" label="Scoring parameters" help="">
                <option value="default" selected="true">Default</option>
                <option value="custom">Custom</option>
            </param>
            <when value="default"/>
            <when value="custom">
                <param argument="--pars" name="pars" type="text" label="Scoring parameters as comma separated string:'DELTA,OMEGA,omega,stop_penalty'" help="See the appendix of the Paper for an explanation for the meaning of these parameters. Default: '-10.0,-4.0,-2.0,-8.0'"/>
            </when>
        </conditional>
        <conditional name="cond_generateEPS">
            <param name="generateEPS" type="select" label="Create colored plots in EPS format" help="The generated plots are resolution independent vector graphics that can be included in any graphics software. For each high scoring segment below a given cutoff (see --eps-cutoff) a file named hss-N.eps is created (N is the running number of the high scoring segment)">
                <option value="create" selected="true">Create Plots</option>
                <option value="nocreate">Do not generate EPS plots</option>
            </param>
            <when value="create">
                <param argument="--eps_cutoff" name="eps_cutoff" type="float" optional="true" value="0.05" label="Create plots only for high scoring segments with p better than:" help=""/>
            </when>
            <when value="nocreate"/>
        </conditional>

        <param name="outputFormat" type="select" label="Output format">
            <option value="--tabular" selected="true">Default</option>
            <option value="--gtf">GTF genome annotation file</option>
        </param>
    </inputs>
    <outputs>
        <data name="outFileDefault" format="tabular" label="${tool.name} on ${on_string}">
            <filter>outputFormat == '--tabular'</filter>
        </data>
        <data name="outFileGTF" format="gtf" label="${tool.name} on ${on_string}">
            <filter>outputFormat == '--gtf'</filter>
        </data>
        <collection name="output_eps" type="list" label="Plots for ${tool.name} on ${on_string}">
            <filter>cond_generateEPS['generateEPS'] == "create"</filter>
            <discover_datasets pattern="(?P&lt;designation&gt;.*)\.eps" directory="eps" ext="eps"/>
        </collection>
    </outputs>
    <tests>
        <test>
            <param name="alignment" value="coding.aln"/>
            <param name="generateEPS" value="nocreate"/>
            <param name="outputFormat" value="--tabular"/>
            <output name="outFileDefault" ftype="tabular" file="rnacode_result1.tabular" compare="sim_size"/>
            <!-- sim_size is needed due to rnacode using random sampling: result files differ, better tests should be implemented -->
        </test>
        <test>
            <param name="alignment" value="coding.aln"/>
            <param name="generateEPS" value="nocreate"/>
            <param name="outputFormat" value="--tabular"/>
            <output name="outFileDefault" ftype="tabular" file="rnacode_result1.tabular" compare="sim_size"/>
            <!-- sim_size is needed due to rnacode using random sampling: result files differ, better tests should be implemented -->
        </test>
        <test>
            <param name="alignment" value="coding.maf"/>
            <conditional name="cond_breakmaf">
                <param name="select_breakmaf" value="breakmaf"/>
            </conditional>
            <param name="generateEPS" value="nocreate"/>
            <param name="outputFormat" value="--gtf"/>
            <output name="outFileGTF" ftype="gtf" file="rnacode_result2.gtf" compare="sim_size"/>
            <!-- sim_size is needed due to rnacode using random sampling: result files differ, better tests should be implemented -->
        </test>
        <test>
            <param name="alignment" value="coding.maf"/>
            <conditional name="cond_breakmaf">
                <param name="select_breakmaf" value="breakmaf"/>
                <param name="processseparately" value="yes"/>
            </conditional>
            <param name="generateEPS" value="nocreate"/>
            <param name="outputFormat" value="--gtf"/>
            <output name="outFileGTF" ftype="gtf" file="rnacode_result2.gtf" compare="sim_size"/>
            <!-- sim_size is needed due to rnacode using random sampling: result files differ, better tests should be implemented -->
        </test>
    </tests>
    <help>
<![CDATA[
**RNAcode** 
Predicts protein coding regions in an alignment of homologous
nucleotide sequences. The prediction is based on evolutionary
signatures typical for protein genese, i.e. the presence of
synonyomous/conservative nucleotide mutations, conservation of the
reading frame and absence of stop codons.

RNAcode does not rely on any species specific sequence characteristics
whatsoever and does not use any machine learning techniques. The only
input required for RNAcode is a multiple sequence alignment either in
MAF or Clustal W format. RNAcode reports local regions of unusual high
coding potential together with an associated p-value.

**Input alignment**

The input alignment needs to be formatted in ClustalW format or MAF
format (http://genome.ucsc.edu/FAQ/FAQformat#format5). The latter
format allows to include genomic coordinates which can be used to
produce annotation files. 

Important: RNAcode uses the first sequence as reference sequence,
i.e. all results and reported coding regions apply to this reference
sequence.

Currently the alignments has to contain at least 3 sequences. Gaps
have to be given as dash ('-'). Unspecified letters given as 'N' are
allowed and treated neutrally during all calculations. No difference is
made between uppercase or lowercase input, i.e. 'softly'-repeat masked
sequences which use lowercase letters for masked regions are treated
the same way as unmasked sequences.

**Output format**

In the default output each prediction is reported on one line by 10 fields.

1. HSS id    
            Unique running number for each high scoring segment
            predicted in one RNAcode call

2. Frame    
            The reading frame phasing relative to the starting
            nucleotide position in the reference sequence. +1 means
            that the first nucleotide in the reference sequence is in
            the same frame as the predicted coding region. Negative
            frames indicate that the predicted regions are on the
            reverse complement strand.

3. Length   
            The length of the predicted region in amino acids

4. From  5. To   
            The position of the first/last amino acid in the translated
            nucleotide sequence of the reference sequence starting
            with 1.

6. Name      
            The name of the reference sequence as given in the input alignment.
  
7. Start 8. End   
            The nucleotide position in the reference sequence of the
            predicted coding region. If no genomic coordinates are given
            (if you provide a CLUSTAL W as input) the first nucleotide position in
            the references sequence is set to 1, otherwise the positions are the
            1-based genomic coordinates as given in the input MAF file.

9.  Score    
            The coding potential score. High scores indicate high coding potential. 

10. P        
            The p-value associated with the score. This is the probability
            that a random alignment with same properties contains an equally good
            or better hit.
]]>
    </help>
    <citations>
        <citation type="doi">10.1261/rna.2536111</citation>
    </citations>
</tool>