annotate tools/ncbi_blast_plus/ncbi_blastp_wrapper.xml @ 22:6f386c5dc4fb draft

v0.2.01 add -max_hsps, -use_sw_tback; lists args; internal updates

<tool id="ncbi_blastp_wrapper" name="NCBI BLAST+ blastp" version="@WRAPPER_VERSION@">
    <description>Search protein database with protein query sequence(s)</description>
    <macros>
        <token name="@BINARY@">blastp</token>
        <import>ncbi_macros.xml</import>
    </macros>
    <expand macro="parallelism" />
    <expand macro="preamble" />
    <command detect_errors="aggressive">
## The command is a Cheetah template which allows some Python based syntax.
## Lines starting hash hash are comments. Galaxy will turn newlines into spaces
blastp
-query '$query'
@BLAST_DB_SUBJECT@
-task $blast_type
-evalue $evalue_cutoff
@BLAST_OUTPUT@
@THREADS@
#if $adv_opts.adv_opts_selector=="advanced":
    -matrix $adv_opts.matrix
    @ADV_FILTER_QUERY@
    @ADV_MAX_HITS@
    @ADV_WORD_SIZE@
    ##Ungapped disabled for now - see comments below
    ##$adv_opts.ungapped
    @ADV_ID_LIST_FILTER@
    @ADV_QCOV_HSP_PERC@
    $adv_opts.use_sw_tback
#end if
    </command>
    <inputs>
        <param argument="-query" type="data" format="fasta" label="Protein query sequence(s)"/>

        <expand macro="input_conditional_protein_db" />

        <param name="blast_type" argument="-task" 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-fast">blastp-fast - Use longer words for seeding, faster but less accurate</option>
            <option value="blastp-short">blastp-short - BLASTP optimized for queries shorter than 30 residues</option>
        </param>
        <expand macro="input_evalue" />
        <expand macro="input_out_format" />
        <expand macro="advanced_options">
            <!-- Could use a select (yes, no, other) where other allows setting 'window locut hicut' -->
            <expand macro="input_filter_query_default_false" />
            <expand macro="input_scoring_matrix" />
            <expand macro="input_max_hits" />
            <expand macro="input_word_size" />
            <!--
            Can't use '-ungapped' on its own, error back is:
            Composition-adjusted searched are not supported with an ungapped search, please add -comp_based_stats F or do a gapped search
            Tried using '-ungapped -comp_based_stats F' and blastp crashed with 'Attempt to access NULL pointer.'
            <param argument="-ungapped" type="boolean" label="Perform ungapped alignment only?" truevalue="-ungapped -comp_based_stats F" falsevalue="" checked="false" />
            -->
            <expand macro="input_parse_deflines" />
            <expand macro="advanced_optional_id_files" />
            <expand macro="input_qcov_hsp_perc" />
            <param argument="-use_sw_tback" type="boolean" truevalue="-use_sw_tback" falsevalue="" label="Compute locally optimal Smith-Waterman alignments" />
        </expand>
    </inputs>
    <outputs>
        <data name="output1" format="tabular" label="${blast_type.value} $query.name vs @ON_DB_SUBJECT@">
            <expand macro="output_change_format" />
        </data>
    </outputs>
    <tests>
        <test>
            <param name="query" value="four_human_proteins.fasta" ftype="fasta" />
            <param name="db_opts_selector" value="file" />
            <param name="subject" value="rhodopsin_proteins.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-8" />
            <param name="blast_type" value="blastp" />
            <param name="out_format" value="5" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM62" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="0" />
            <param name="parse_deflines" value="true" />
            <param name="qcov_hsp_perc" value="25" />
            <output name="output1" file="blastp_four_human_vs_rhodopsin.xml" ftype="blastxml" />
        </test>
        <test>
            <param name="query" value="four_human_proteins.fasta" ftype="fasta" />
            <param name="db_opts_selector" value="file" />
            <param name="subject" value="rhodopsin_proteins.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-8" />
            <param name="blast_type" value="blastp" />
            <param name="out_format" value="6" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM62" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="0" />
            <param name="parse_deflines" value="true" />
            <param name="qcov_hsp_perc" value="25" />
            <output name="output1" file="blastp_four_human_vs_rhodopsin.tabular" ftype="tabular" />
        </test>
        <test>
            <param name="query" value="four_human_proteins.fasta" ftype="fasta" />
            <param name="db_opts_selector" value="file" />
            <param name="subject" value="rhodopsin_proteins.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-8" />
            <param name="blast_type" value="blastp" />
            <param name="out_format" value="ext" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM62" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="0" />
            <param name="parse_deflines" value="true" />
            <param name="qcov_hsp_perc" value="25" />
            <output name="output1" file="blastp_four_human_vs_rhodopsin_ext.tabular" ftype="tabular" />
        </test>
        <test>
            <param name="query" value="rhodopsin_proteins.fasta" ftype="fasta" />
            <param name="db_opts_selector" value="file" />
            <param name="subject" value="four_human_proteins.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-8" />
            <param name="blast_type" value="blastp" />
            <param name="out_format" value="6" />
            <param name="adv_opts_selector" value="basic" />
            <output name="output1" file="blastp_rhodopsin_vs_four_human.tabular" ftype="tabular" />
        </test>
    </tests>
    <help>

@SEARCH_TIME_WARNING@

**What it does**

Search a *protein database* using a *protein query*,
using the NCBI BLAST+ blastp command line tool.

@FASTA_WARNING@

-----

@OUTPUT_FORMAT@

-------

**References**

If you use this Galaxy tool in work leading to a scientific publication please
cite the following papers:

@REFERENCES@
    </help>
    <expand macro="blast_citations" />
</tool>