<tool id="hyphy_fubar" name="HyPhy-FUBAR" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
    <description>Fast Unconstrained Bayesian AppRoximation</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="bio_tools"/>
    <expand macro="requirements"/>
    <command detect_errors="exit_code"><![CDATA[
        @SYMLINK_FILES@
        ln -s '$fubar_output' ${input_file}.FUBAR.json &&
        hyphy fubar
            --alignment ./$input_file
            @INPUT_TREE@
            --code '$gencodeid'
            --method '$posteriorEstimationMethod.method'
            @posteriorEstimationMethod_cmd@
            --grid '$advanced_options.grid_points'
            --concentration_parameter '$advanced_options.concentration'
            --non-zero $advanced_options.non_zero
            --kill-zero-lengths $advanced_options.kill_zero_lengths
            > fubar_stdout.md 
        @ERRORS@
    ]]></command>
    <inputs>
        <expand macro="inputs"/>
        <expand macro="gencode"/>
        <expand macro="conditional_posteriorEstimationMethod" />
        <section name="advanced_options" title="Advanced Options" expanded="false">
            <param argument="--grid" name="grid_points" type="integer" value="20" min="5" max="50" label="Grid points" help="The number of grid points used to approximate the posterior distribution of dN and dS." />
            <param argument="--concentration_parameter" name="concentration" type="float" value="0.5" min="0.001" max="1" label="Concentration parameter of the Dirichlet prior" help="The concentration parameter of the Dirichlet prior on the grid weights." />
            <param argument="--non-zero" type="boolean" truevalue="Yes" falsevalue="No" label="Enforce non-zero synonymous rates" help="Enforce non-zero synonymous rates on the grid. This is useful for calculating dN/dS ratios, as it prevents division by zero."/>
            <expand macro="kill_zero_lengths_param"/>
        </section>

    </inputs>
    <outputs>
        <data name="fubar_output" format="hyphy_results.json" />
        <data name="fubar_md_report" format="markdown" from_work_dir="fubar_stdout.md" label="FUBAR Report (Markdown) for ${tool.name} on ${on_string}" />
    </outputs>
    <tests>
        <test expect_num_outputs="2">
            <param name="input_file" ftype="fasta.gz" value="fubar-in1.fa.gz"/>
            <param name="input_nhx" ftype="nhx" value="fubar-in1.nhx"/>
            <conditional name="posteriorEstimationMethod">
                <param name="method" value="Variational-Bayes"/>
            </conditional>
            <output name="fubar_output">
                <assert_contents>
                    <has_text text="Empiricial Bayes Factor for positive selection at a site"/>
                </assert_contents>
            </output>
            <output name="fubar_md_report">
                <assert_contents>
                    <has_text text="Running an iterative zeroth order variational Bayes procedure to estimate the posterior mean of rate weights"/>
                    <has_text text="### Tabulating site-level results"/>
                </assert_contents>
            </output>
        </test>
    </tests>
    <help><![CDATA[

FUBAR : Faste Unbiased Bayesian AppRoximation
=============================================

What question does this method answer?
--------------------------------------

Which site(s) in a gene are subject to pervasive, i.e. consistently across the entire phylogeny, diversifying selection?

Recommended Applications
------------------------

The phenomenon of pervasive selection is generally most prevalent in pathogen evolution and any biological system influenced by evolutionary arms race dynamics
(or balancing selection), including adaptive immune escape by viruses. As such, FUBAR is ideally suited to identify sites under positive selection which
represent candidate sites subject to strong selective pressures across the entire phylogeny.

FUBAR is our recommended method for detecting pervasive selection at individual sites on large (> 500 sequences) datasets for which other methods have prohibitive runtimes, unless you have access to a computer cluster.

Brief description
-----------------

FUBAR (Fast, Unconstrained Bayesian AppRoximation) is a Bayesian method for detecting site-specific positive and negative selection. It is designed to be fast and efficient, making it suitable for large datasets.

The core idea behind FUBAR is to model the non-synonymous (dN) and synonymous (dS) substitution rates at each site in a codon alignment. The ratio of these rates (dN/dS, or omega) is a measure of the selective pressure acting on a site. An omega value greater than 1 indicates positive (diversifying) selection, a value less than 1 indicates negative (purifying) selection, and a value of 1 indicates neutral evolution.

FUBAR uses a Bayesian approach to infer the posterior distribution of dN and dS at each site. It does this by discretizing the dN and dS rates into a grid of points and then using a Bayesian graphical model to infer the posterior probability of each grid point for each site. This approach is much faster than traditional MCMC-based methods, which require long run times to converge.

FUBAR offers three different methods for estimating the posterior distribution:

*   **Variational-Bayes:** A fast approximation method that is the recommended default.
*   **Collapsed-Gibbs:** A faster MCMC method.
*   **Metropolis-Hastings:** The original, slowest MCMC method.

Input
-----

1. A *FASTA* sequence alignment.
2. A phylogenetic tree in the *Newick* format

Note: the names of sequences in the alignment must match the names of the sequences in the tree.


Output
------

A JSON file with analysis results (http://hyphy.org/resources/json-fields.pdf).

A custom visualization module for viewing these results is available (see http://vision.hyphy.org/FUBAR for an example)

Further reading
---------------

http://hyphy.org/methods/selection-methods/#FUBAR


Tool options
------------
::


    --code             Which genetic code to use

    --grid             The number of grid points used to approximate the posterior distribution of dN and dS. A larger grid will provide a more accurate approximation but will also be slower. The default value of 20 is a good compromise between speed and accuracy.

    --method           The inference method to use for estimating the posterior distribution.
                            Variational-Bayes : 0-th order Variational Bayes approximation; fastest [default]
                            Metropolis-Hastings : Full Metropolis-Hastings MCMC algorithm; orignal method [slowest]
                            Collapsed-Gibbs  : Collapsed Gibbs sampler [intermediate speed]


    --chains           The number of MCMC chains to run. This is only applicable to the Metropolis-Hastings and Collapsed-Gibbs methods. A larger number of chains will provide a better exploration of the posterior distribution but will also be slower.
                            default value: 5

    --chain-length     The length of each MCMC chain. This is only applicable to the Metropolis-Hastings and Collapsed-Gibbs methods. A longer chain will provide a better exploration of the posterior distribution but will also be slower.
                            default value: 2,000,000

    --burn-in          The number of samples to discard from the beginning of each MCMC chain. This is done to ensure that the chain has converged to the posterior distribution. This is only applicable to the Metropolis-Hastings and Collapsed-Gibbs methods.
                            default value: 1,000,000

    --samples          The number of samples to draw from each MCMC chain after the burn-in period. These samples are used to estimate the posterior distribution. This is only applicable to the Metropolis-Hastings and Collapsed-Gibbs methods.
                            default value: 1,000

    --concentration_parameter
                        The concentration parameter of the Dirichlet prior on the grid weights.
                        default value: 0.5

    --non-zero          Enforce non-zero synonymous rates on the grid. This is useful for calculating dN/dS ratios, as it prevents division by zero.

    --kill-zero-lengths Automatically delete internal zero-length branches for computational efficiency. This will not affect the results.

    ]]></help>
    <expand macro="citations">
        <citation type="doi">10.1093/molbev/mst030</citation>
    </expand>
</tool>