view tools/ncbi_blast_plus/ncbi_tblastn_wrapper.xml @ 26:2889433c7ae1 draft

v0.3.3 - fixed legacy dependecy definition

<tool id="ncbi_tblastn_wrapper" name="NCBI BLAST+ tblastn" version="@WRAPPER_VERSION@">
    <description>Search translated nucleotide database with protein query sequence(s)</description>
    <macros>
        <token name="@BINARY@">tblastn</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
tblastn
@QUERY@
@BLAST_DB_SUBJECT@
-task $blast_type
-evalue $evalue_cutoff
@BLAST_OUTPUT@
@THREADS@
#if $adv_opts.adv_opts_selector=="advanced":
-db_gencode $adv_opts.db_gencode
@ADV_FILTER_QUERY@
@ADV_MAX_HITS@
@ADV_WORD_SIZE@
@ADV_MATRIX_GAPCOSTS@
##Ungapped disabled for now - see comments below
##$adv_opts.ungapped
@ADV_ID_LIST_FILTER@
@ADV_QCOV_HSP_PERC@
@ADV_WINDOW_SIZE@
@ADV_THRESHOLD@
@ADV_COMP_BASED_STATS@
## End of advanced options:
#end if
    </command>
    <inputs>
        <expand macro="prot_query" />

        <expand macro="input_conditional_nucleotide_db" />
        <param name="blast_type" argument="-task" type="select" display="radio" label="Type of BLAST">
            <option value="tblastn">tblastn - Traditional TBLASTN to compare protein query to translated nucleotide database</option>
            <option value="tblastn-fast">tblastn-fast - Use longer words for seeding, faster but less accurate</option>
        </param>
        <expand macro="input_evalue" />
        <expand macro="input_out_format" />
        <expand macro="advanced_options">
            <expand macro="input_db_gencode" />

            <!-- Could use a select (yes, no, other) where other allows setting 'window locut hicut' -->
            <expand macro="input_filter_query_default_true" />
            <expand macro="input_matrix_gapcosts" />
            <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 tblastn 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" />
            <expand macro="input_window_size" />
            <expand macro="input_threshold" />
            <expand macro="input_comp_based_stats" />
        </expand>
    </inputs>
    <outputs>
        <data name="output1" format="tabular" label="tblastn $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_nucs.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-10" />
            <param name="out_format" value="5" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM80" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="" />
            <param name="parse_deflines" value="false" />
            <output name="output1" file="tblastn_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_nucs.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-10" />
            <param name="out_format" value="ext" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM80" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="" />
            <param name="parse_deflines" value="false" />
            <output name="output1" file="tblastn_four_human_vs_rhodopsin_ext.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_nucs.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-10" />
            <param name="out_format" value="6" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM80" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="" />
            <param name="parse_deflines" value="false" />
            <output name="output1" file="tblastn_four_human_vs_rhodopsin.tabular" ftype="tabular" />
        </test>
        <test>
            <!-- Same as above, but parse deflines -->
            <!-- Note back on BLAST 2.2.25+ - 2.2.27+ made no difference -->
            <param name="query" value="four_human_proteins.fasta" ftype="fasta" />
            <param name="db_opts_selector" value="file" />
            <param name="subject" value="rhodopsin_nucs.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-10" />
            <param name="out_format" value="6" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM80" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="" />
            <param name="parse_deflines" value="true" />
            <output name="output1" file="tblastn_four_human_vs_rhodopsin_deflines.tabular" ftype="tabular" />
        </test>
        <test>
            <!-- As of BLAST 2.5.0+ the full path appears twice, so get small diff -->
            <param name="query" value="four_human_proteins.fasta" ftype="fasta" />
            <param name="db_opts_selector" value="file" />
            <param name="subject" value="rhodopsin_nucs.fasta" ftype="fasta" />
            <param name="database" value="" />
            <param name="evalue_cutoff" value="1e-10" />
            <param name="out_format" value="0 -html" />
            <param name="adv_opts_selector" value="advanced" />
            <param name="filter_query" value="false" />
            <param name="matrix" value="BLOSUM80" />
            <param name="max_hits" value="0" />
            <param name="word_size" value="" />
            <param name="parse_deflines" value="false" />
            <output name="output1" file="tblastn_four_human_vs_rhodopsin.html" ftype="html" lines_diff="8" />
        </test>
    </tests>
    <help>

@SEARCH_TIME_WARNING@

**What it does**

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

@FASTA_WARNING@

------

@OUTPUT_FORMAT@

------

@CLI_OPTIONS@

------

**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>