<tool id="blast_reciprocal_best_hits" name="BLAST Reciprocal Best Hits (RBH)" version="0.1.0">
    <description>from two FASTA files</description>
    <requirements>
            <requirement type="binary">makeblastdb</requirement>
            <requirement type="binary">blastp</requirement>
            <requirement type="binary">blastn</requirement>
            <requirement type="package" version="2.2.29">blast+</requirement>
    </requirements>
    <version_command interpreter="python">
blast_rbh.py --version
    </version_command>
    <command interpreter="python">
blast_rbh.py "$fasta_a" "$fasta_b" $seq.dbtype
#if $seq.dbtype=="nucl"
    $seq.nucl_type
#else
    $seq.prot_type
#end if
$identity $q_cover "$output"
    </command>
    <inputs>
        <!-- Galaxy does not have sub-types for protein vs nucletide FASTA -->
        <param name="fasta_a" type="data" format="fasta"
	       label="Genes/proteins from species A"
	       description="FASTA file, one sequence per gene/protein." /> 
        <param name="fasta_b" type="data" format="fasta"
	       label="Genes/proteins from species B"
	       description="FASTA file, one sequence per gene/protein." /> 
        <conditional name="seq">
            <param name="dbtype" type="select" label="Molecule type of FASTA inputs">
                <option value="prot">protein</option>
                <option value="nucl">nucleotide</option>
            </param>
            <when value="prot">
                <param name="prot_type" type="select" display="radio" label="Type of BLAST">
                    <option value="blastp">blastp - Traditional BLASTP to compare a protein query to a protein database</option>
                    <option value="blastp-short">blastp-short - BLASTP optimized for queries shorter than 30 residues</option>
                </param>
            </when>
            <when value="nucl">
                <param name="nucl_type" type="select" display="radio" label="Type of BLAST">
                    <option value="megablast">megablast - Traditional megablast used to find very similar (e.g., intraspecies or closely related species) sequences</option>
                    <option value="blastn">blastn - Traditional BLASTN requiring an exact match of 11, for somewhat similar sequences</option>
                    <option value="blastn-short">blastn-short - BLASTN program optimized for sequences shorter than 50 bases</option>
                    <option value="dc-megablast">dc-megablast - Discontiguous megablast used to find more distant (e.g., interspecies) sequences</option>
                </param>
            </when>
        </conditional>
	<param name="identity" type="float" value="70" min="0" max="100"
	       label="Minimum percentage identity for BLAST matches"
	       help="Default is 70%, use 0 for no filtering." />
        <param name="q_cover" type="float" value="50" min="0" max="100"
	       label="Minimum percentage query coverage for BLAST matches"
	       help="Default is 50%, use 0 for no filtering." />
    </inputs>
    <outputs>
        <data name="output" format="tabular" label="BLAST RBH: $fasta_a.name vs $fasta_b.name" />
    </outputs>
    <requirements>
    </requirements>
    <tests>
        <test>
            <param name="fasta_a" value="rhodopsin_nucs.fasta" ftype="fasta"/>
            <param name="fasta_b" value="three_human_mRNA.fasta" ftype="fasta"/>
            <param name="dbtype" value="nucl"/>
            <param name="nucl_type" value="megablast"/>
            <param name="identity" value="0.0"/>
            <param name="q_cover" value="0.0"/>
            <output name="output" file="rbh_megablast_rhodopsin_nucs_vs_three_human_mRNA.tabular" ftype="tabular"/>
        </test>
        <test>
            <param name="fasta_a" value="rhodopsin_nucs.fasta" ftype="fasta"/>
            <param name="fasta_b" value="three_human_mRNA.fasta" ftype="fasta"/>
            <param name="dbtype" value="nucl"/>
            <param name="nucl_type" value="megablast"/>
            <param name="identity" value="92"/>
            <param name="q_cover" value="86"/>
            <output name="output" file="rbh_megablast_rhodopsin_nucs_vs_three_human_mRNA.tabular" ftype="tabular"/>
        </test>
        <!-- push the percentage identity over the 92.07% level -->
        <test>
            <param name="fasta_a" value="rhodopsin_nucs.fasta" ftype="fasta"/>
            <param name="fasta_b" value="three_human_mRNA.fasta" ftype="fasta"/>
            <param name="dbtype" value="nucl"/>
            <param name="nucl_type" value="megablast"/>
            <param name="identity" value="92.5"/>
            <param name="q_cover" value="86"/>
            <output name="output" file="rbh_none.tabular" ftype="tabular"/>
        </test>
	<!-- push the coverage over the 86% level -->
        <test>
            <param name="fasta_a" value="rhodopsin_nucs.fasta" ftype="fasta"/>
            <param name="fasta_b" value="three_human_mRNA.fasta" ftype="fasta"/>
            <param name="dbtype" value="nucl"/>
            <param name="nucl_type" value="megablast"/>
            <param name="identity" value="92"/>
            <param name="q_cover" value="87"/>
            <output name="output" file="rbh_none.tabular" ftype="tabular"/>
        </test>
        <test>
            <param name="fasta_a" value="three_human_mRNA.fasta" ftype="fasta"/>
            <param name="fasta_b" value="rhodopsin_nucs.fasta" ftype="fasta"/>
            <param name="dbtype" value="nucl"/>
            <param name="nucl_type" value="blastn"/>
            <param name="identity" value="0.0"/>
            <param name="q_cover" value="0.0"/>
            <output name="output" file="rbh_blastn_three_human_mRNA_vs_rhodopsin_nucs.tabular" ftype="tabular"/>
        </test>
    </tests>
    <help>
**What it does**

Takes two FASTA files (species *A* and species *B*), builds a BLAST database
for each, runs reciprocal BLAST searchs (*A vs B*, and *B vs A*), optionally
filters these, and then compiles a list of the reciprocal best hits (RBH).

The output from this tool is a tabular file containing four columns, with
the order taken from input file A:

====== ======================
Column Description
------ ----------------------
     1 ID from species *A*
     2 ID from species *B*
     3 Bitscore from *A vs B*
     4 Bitscore from *B vs A*
====== ======================

.. class:: warningmark

**Note**

If you are trying to use BLAST RBH matches to identify candidate orthologues
or transfer annotation, you *must* use a percentage identity and minimum
coverage threshold or similiar. See:

Punta and Ofran (2008) The Rough Guide to In Silico Function Prediction,
or How To Use Sequence and Structure Information To Predict Protein
Function. PLoS Comput Biol 4(10): e1000160.
http://dx.doi.org/10.1371/journal.pcbi.1000160

The defaults are to require 70% sequence identity over the aligned region
(using ``pident`` in the BLAST+ tabular output), and that the HSP alignment
covers at least 50% of the query sequence (using ``qcovhsp`` in the BLAST+
tabular output).


**References**

A specific paper covering this tool is planned, but please also cite:

Christiam Camacho et al. (2009).
BLAST+: architecture and applications.
BMC Bioinformatics. 15;10:421.
http://dx.doi.org/10.1186/1471-2105-10-421

This wrapper is available to install into other Galaxy Instances via the Galaxy
Tool Shed at http://toolshed.g2.bx.psu.edu/view/peterjc/blast_rbh
    </help>
</tool>