annotate eggnog_mapper.xml @ 12:9d1fbff733cf draft

planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/eggnog_mapper/eggnog_mapper commit 07877ba6d3fb6e28e94500f2392db6393cf325fd

<tool id="eggnog_mapper" name="eggNOG Mapper" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@">
    <description>functional sequence annotation by orthology</description>
    <macros>
        <import>eggnog_macros.xml</import>
    </macros>
    <expand macro="requirements"/>
    <expand macro="version_command"/>
    <command detect_errors="aggressive"><![CDATA[
        #if $ortho_method.m == "no_search"
            cat 
            #for aht in $ortho_method.annotate_hits_table
                $aht
            #end for
            > annotate_hits_table.tsv
            &&
        #end if

        emapper.py
        --data_dir '$eggnog_data.fields.path'
        -m '$ortho_method.m'

        #if $ortho_method.m in ['diamond', 'mmseqs', 'cache']:
            -i '$ortho_method.input'
            --itype '$ortho_method.input_trans.itype'
            #if $ortho_method.input_trans.itype in ['CDS', 'genome', 'metagenome']:
                $ortho_method.input_trans.translate
            #end if
            #if $ortho_method.input_trans.itype in ['genome', 'metagenome']:
                --genepred $ortho_method.input_trans.genepred
            #end if
        #elif $ortho_method.m == "no_search"
            --annotate_hits_table annotate_hits_table.tsv
        #end if
        
        #if $ortho_method.m == 'cache'
            --cache '$ortho_method.cache'
        #end if

        #if $ortho_method.m in ['diamond', 'mmseqs']:
            ## Diamond option
            #if $ortho_method.m == "diamond":
                --matrix '$ortho_method.matrix_gapcosts.matrix'
                $ortho_method.matrix_gapcosts.gap_costs
                --sensmode $ortho_method.sensmode
                $ortho_method.dmnd_iterate
                $ortho_method.dmnd_ignore_warnings
            #elif $ortho_method.m == "mmseqs":
                --start_sens $ortho_method.start_sens
                --sens_steps $ortho_method.sens_steps
                --final_sens $ortho_method.final_sens
            #end if

            ## Common options for search filtering (applies to diamond and mmseqs only)
            #if str($ortho_method.query_cover):
                --query_cover $ortho_method.query_cover
            #end if
            #if str($ortho_method.subject_cover):
                --subject_cover $ortho_method.subject_cover
            #end if
            #if str($ortho_method.pident):
                --pident $ortho_method.pident
            #end if
            #if str($ortho_method.evalue):
                --evalue $ortho_method.evalue
            #end if
            #if str($ortho_method.score):
                --score $ortho_method.score
            #end if
        #end if

        #if $annotation_options.no_annot == "--no_annot"
            --no_annot
        #else
            #if str($annotation_options.seed_ortholog_evalue):
                --seed_ortholog_evalue $annotation_options.seed_ortholog_evalue
            #end if
            #if str($annotation_options.seed_ortholog_score):
                --seed_ortholog_score $annotation_options.seed_ortholog_score
            #end if
            #if $annotation_options.tax_scope:
                --tax_scope=$annotation_options.tax_scope
            #end if
            #if $annotation_options.target_orthologs:
                --target_orthologs=$annotation_options.target_orthologs
            #end if
            #if $annotation_options.go_evidence:
                --go_evidence=$annotation_options.go_evidence
            #end if
        #end if
        $output_options.no_file_comments
        $output_options.report_orthologs
        $output_options.md5
        --output='results'
        --cpu "\${GALAXY_SLOTS:-4}"
        --scratch_dir \${TEMP:-\$_GALAXY_JOB_TMP_DIR}
        --temp_dir \${TEMP:-\$_GALAXY_JOB_TMP_DIR}
    ]]></command>
    <inputs>
        <param name="eggnog_data" type="select" label="Version of eggNOG Database">
            <options from_data_table="eggnog_mapper_db_versioned">
                <filter type="static_value" column="3" value="@IDX_VERSION@" />
            </options>
        </param>

        <conditional name="ortho_method">
            <param argument="-m" type="select" label="Method to search seed orthologs">
                <option value="diamond" selected="true">Diamond</option>
                <option value="mmseqs">MMseqs2</option>
                <option value="no_search">Skip search stage: annotate an existing seed orthologs data set</option>
                <option value="cache">Skip search stage: annotate based on cached annotations (see also --md5)</option>
            </param>
            <when value="diamond">
                <expand macro="fasta_input"/>
                <conditional name="matrix_gapcosts">
                    <param argument="--matrix" type="select" label="Scoring matrix and gap costs">
                        <option value="BLOSUM90">BLOSUM90</option>
                        <option value="BLOSUM80">BLOSUM80</option>
                        <option value="BLOSUM62" selected="true">BLOSUM62</option>
                        <option value="BLOSUM50">BLOSUM50</option>
                        <option value="BLOSUM45">BLOSUM45</option>
                        <option value="PAM250">PAM250</option>
                        <option value="PAM70">PAM70</option>
                        <option value="PAM30">PAM30</option>
                    </param>
                    <when value="BLOSUM90">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 9 --gapextend 2">Existence: 9  Extension: 2</option>
                            <option value="--gapopen 8 --gapextend 2">Existence: 8  Extension: 2</option>
                            <option value="--gapopen 7 --gapextend 2">Existence: 7  Extension: 2</option>
                            <option value="--gapopen 6 --gapextend 2">Existence: 6  Extension: 2</option>
                            <option value="--gapopen 11 --gapextend 1">Existence: 11  Extension: 1</option>
                            <option value="--gapopen 10 --gapextend 1" selected="true">Existence: 10  Extension: 1</option>
                            <option value="--gapopen 9 --gapextend 1">Existence: 9  Extension: 1</option>
                        </param>
                    </when>
                    <when value="BLOSUM80">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 8 --gapextend 2">Existence: 8  Extension: 2</option>
                            <option value="--gapopen 7 --gapextend 2">Existence: 7  Extension: 2</option>
                            <option value="--gapopen 6 --gapextend 2">Existence: 6  Extension: 2</option>
                            <option value="--gapopen 11 --gapextend 1">Existence: 11  Extension: 1</option>
                            <option value="--gapopen 10 --gapextend 1" selected="true">Existence: 10  Extension: 1</option>
                            <option value="--gapopen 9 --gapextend 1">Existence: 9  Extension: 1</option>
                        </param>
                    </when>
                    <when value="BLOSUM62">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 11 --gapextend 2">Existence: 11  Extension: 2</option>
                            <option value="--gapopen 10 --gapextend 2">Existence: 10  Extension: 2</option>
                            <option value="--gapopen 9 --gapextend 2">Existence: 9  Extension: 2</option>
                            <option value="--gapopen 8 --gapextend 2">Existence: 8  Extension: 2</option>
                            <option value="--gapopen 7 --gapextend 2">Existence: 7  Extension: 2</option>
                            <option value="--gapopen 6 --gapextend 2">Existence: 6  Extension: 2</option>
                            <option value="--gapopen 13 --gapextend 1">Existence: 13  Extension: 1</option>
                            <option value="--gapopen 12 --gapextend 1">Existence: 12  Extension: 1</option>
                            <option value="--gapopen 11 --gapextend 1" selected="true">Existence: 11  Extension: 1</option>
                            <option value="--gapopen 10 --gapextend 1">Existence: 10  Extension: 1</option>
                            <option value="--gapopen 9 --gapextend 1">Existence: 9  Extension: 1</option>
                        </param>
                    </when>
                    <when value="BLOSUM50">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 13 --gapextend 3">Existence: 13  Extension: 3</option>
                            <option value="--gapopen 12 --gapextend 3">Existence: 12  Extension: 3</option>
                            <option value="--gapopen 11 --gapextend 3">Existence: 11  Extension: 3</option>
                            <option value="--gapopen 10 --gapextend 3">Existence: 10  Extension: 3</option>
                            <option value="--gapopen 9 --gapextend 3">Existence: 9  Extension: 3</option>
                            <option value="--gapopen 16 --gapextend 2">Existence: 16  Extension: 2</option>
                            <option value="--gapopen 15 --gapextend 2">Existence: 15  Extension: 2</option>
                            <option value="--gapopen 14 --gapextend 2">Existence: 14  Extension: 2</option>
                            <option value="--gapopen 13 --gapextend 2" selected="true">Existence: 13  Extension: 2</option>
                            <option value="--gapopen 12 --gapextend 2">Existence: 12  Extension: 2</option>
                            <option value="--gapopen 19 --gapextend 1">Existence: 19  Extension: 1</option>
                            <option value="--gapopen 18 --gapextend 1">Existence: 18  Extension: 1</option>
                            <option value="--gapopen 17 --gapextend 1">Existence: 17  Extension: 1</option>
                            <option value="--gapopen 16 --gapextend 1">Existence: 16  Extension: 1</option>
                            <option value="--gapopen 15 --gapextend 1">Existence: 15  Extension: 1</option>
                        </param>
                    </when>
                    <when value="BLOSUM45">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 13 --gapextend 3">Existence: 13  Extension: 3</option>
                            <option value="--gapopen 12 --gapextend 3">Existence: 12  Extension: 3</option>
                            <option value="--gapopen 11 --gapextend 3">Existence: 11  Extension: 3</option>
                            <option value="--gapopen 10 --gapextend 3">Existence: 10  Extension: 3</option>
                            <option value="--gapopen 15 --gapextend 2" selected="true">Existence: 15  Extension: 2</option>
                            <option value="--gapopen 14 --gapextend 2">Existence: 14  Extension: 2</option>
                            <option value="--gapopen 13 --gapextend 2">Existence: 13  Extension: 2</option>
                            <option value="--gapopen 12 --gapextend 2">Existence: 12  Extension: 2</option>
                            <option value="--gapopen 19 --gapextend 1">Existence: 19  Extension: 1</option>
                            <option value="--gapopen 18 --gapextend 1">Existence: 18  Extension: 1</option>
                            <option value="--gapopen 17 --gapextend 1">Existence: 17  Extension: 1</option>
                            <option value="--gapopen 16 --gapextend 1">Existence: 16  Extension: 1</option>
                        </param>
                    </when>
                    <when value="PAM250">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 15 --gapextend 3">Existence: 15  Extension: 3</option>
                            <option value="--gapopen 14 --gapextend 3">Existence: 14  Extension: 3</option>
                            <option value="--gapopen 13 --gapextend 3">Existence: 13  Extension: 3</option>
                            <option value="--gapopen 12 --gapextend 3">Existence: 12  Extension: 3</option>
                            <option value="--gapopen 17 --gapextend 2">Existence: 17  Extension: 2</option>
                            <option value="--gapopen 16 --gapextend 2">Existence: 16  Extension: 2</option>
                            <option value="--gapopen 15 --gapextend 2">Existence: 15  Extension: 2</option>
                            <option value="--gapopen 14 --gapextend 2" selected="true">Existence: 14  Extension: 2</option>
                            <option value="--gapopen 13 --gapextend 2">Existence: 13  Extension: 2</option>
                            <option value="--gapopen 21 --gapextend 1">Existence: 21  Extension: 1</option>
                            <option value="--gapopen 20 --gapextend 1">Existence: 20  Extension: 1</option>
                            <option value="--gapopen 19 --gapextend 1">Existence: 19  Extension: 1</option>
                            <option value="--gapopen 18 --gapextend 1">Existence: 18  Extension: 1</option>
                            <option value="--gapopen 17 --gapextend 1">Existence: 17  Extension: 1</option>
                        </param>
                    </when>
                    <when value="PAM70">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 12 --gapextend 3">Existence: 12 Extension: 3</option>
                            <option value="--gapopen 11 --gapextend 2">Existence: 11 Extension: 2</option>
                            <option value="--gapopen 8 --gapextend 2">Existence: 8  Extension: 2</option>
                            <option value="--gapopen 7 --gapextend 2">Existence: 7  Extension: 2</option>
                            <option value="--gapopen 6 --gapextend 2">Existence: 6  Extension: 2</option>
                            <option value="--gapopen 11 --gapextend 1">Existence: 11  Extension: 1</option>
                            <option value="--gapopen 10 --gapextend 1" selected="true">Existence: 10  Extension: 1</option>
                            <option value="--gapopen 9 --gapextend 1">Existence: 9  Extension: 1</option>
                        </param>
                    </when>
                    <when value="PAM30">
                        <param name="gap_costs" type="select" label="Gap Costs">
                            <option value="--gapopen 15 --gapextend 3">Existence: 15 Extension: 3</option>
                            <option value="--gapopen 13 --gapextend 3">Existence: 13 Extension: 3</option>
                            <option value="--gapopen 14 --gapextend 2">Existence: 14 Extension: 2</option>
                            <option value="--gapopen 7 --gapextend 2">Existence: 7  Extension: 2</option>
                            <option value="--gapopen 6 --gapextend 2">Existence: 6  Extension: 2</option>
                            <option value="--gapopen 5 --gapextend 2">Existence: 5  Extension: 2</option>
                            <option value="--gapopen 14 --gapextend 1">Existence: 14 Extension: 1</option>
                            <option value="--gapopen 10 --gapextend 1">Existence: 10  Extension: 1</option>
                            <option value="--gapopen 9 --gapextend 1" selected="true">Existence: 9  Extension: 1</option>
                            <option value="--gapopen 8 --gapextend 1">Existence: 8  Extension: 1</option>
                        </param>
                    </when>
                </conditional>
                <param argument="--sensmode" type="select" label="Diamond's sensitivity mode">
                    <option value="default">default</option>
                    <option value="fast">fast</option>
                    <option value="mid-sensitive">mid-sensitive</option>
                    <option value="sensitive" selected="true">sensitive</option>
                    <option value="more-sensitive">more-sensitive</option>
                    <option value="very-sensitive">very-sensitive</option>
                    <option value="ultra-sensitive">ultra-sensitive</option>
                </param>
                <param argument="--dmnd_iterate" type="boolean" truevalue="--dmnd_iterate yes" falsevalue="--dmnd_iterate no" checked="false"
                    label="Run diamond in iterative mode, up to the sensitivity level"/>
                <param argument="--dmnd_ignore_warnings" type="boolean" truevalue="--dmnd_ignore_warnings" falsevalue="" checked="false"
                    label="Ignore Diamond warnings on sequence content (e.g. when a protein contains only ATGC symbols)"/>
                <expand macro="common_search_options"/>
            </when>
            <when value="mmseqs">
                <expand macro="fasta_input"/>
                <param argument="--start_sens" type="integer" value="3" min="0" max="100" label="Starting sensitivity" />
                <param argument="--sens_steps" type="integer" value="3" min="0" max="100" label="Number of sensitivity steps" />
                <param argument="--final_sens" type="integer" value="7" min="0" max="100" label="Final sensititivy step" />
                <expand macro="common_search_options"/>
            </when>
            <when value="no_search">
                <param argument="--annotate_hits_table" type="data" multiple="true" format="tabular" label="Seed orthologs">
                    <validator type="expression" message="No seed orthologs">value.metadata.columns == 11</validator>
                    <!-- would be cool to replace with this validator: 
                        <validator type="expression" message="No seed orthologs">value.metadata.column_names == '@SEED_ORTHOLOG_COLUMNS@'.split()</validator>
                        but this does not work (yet) in tool tests since column_names can not be set in uploads -->
                </param>
            </when>
            <when value="cache">
                <expand macro="fasta_input"/>
                <param argument="--cache"  type="data" multiple="true" format="tabular" label="EggNOG Annotations with md5 hashes" help="Annotations computed with EggNOG mapper with enabled --md5 option">
                    <validator type="expression" message="No seed orthologs">value.metadata.columns == 22</validator>
                </param>
                <param name="output_no_annotations" type="boolean" checked="true" label="Output sequences without annotation" help="Produce an additional FASTA file with the sequences of queries for which an existing annotation was not found using cache mode. This file can be used as input of another eggNOG-mapper run without using the cache, trying to annotate the sequences."/>
            </when>
        </conditional>

        <conditional name="annotation_options">
            <param argument="--no_annot" type="select" label="Annotate seed orthologs">
                <option value="">Yes</option>
                <option value="--no_annot">No</option>
            </param>
            <when value="">
                <param argument="--seed_ortholog_evalue" type="float" value="0.001" min="0" label="Min E-value threshold">
                    <help>
                    Min E-value expected when searching for seed eggNOG ortholog. Applies to phmmer/diamond searches.
                    Queries not having a significant seed orthologs (E-value less than threshold) will not be annotated.
                    </help>
                </param>
                <param argument="--seed_ortholog_score" type="float" optional="true" min="0" label="Minimum bit score threshold">
                    <help>
                    Min bit score expected when searching for seed eggNOG ortholog.
                    Queries not having a significant seed orthologs will not be annotated.
                    </help>
                </param>
                <param argument="--tax_scope" type="integer" optional="true" label="Set taxonomic scope" help="NCBI taxonomy id" />
                <param argument="--target_orthologs" type="select" label="target orthologs for functional transfer">
                    <option value="one2one">one2one</option>
                    <option value="many2one">many2one</option>
                    <option value="one2many">one2many</option>
                    <option value="many2many">many2many</option>
                    <option value="all" selected="true">all</option>
                </param>
                <param argument="--go_evidence" type="select"
                        label="Select the set of GO terms that should be used for annotation">
                    <option value="experimental">experimental = Use only terms inferred from experimental evidence</option>
                    <option value="non-electronic" selected="true">non-electronic = Use only non-electronically curated terms</option>
                    <option value="all" selected="true">All (experimental + non-electronic)</option>
                </param>
            </when>
            <when value="--no_annot"/>
        </conditional>
        
        <section name="output_options" expanded="false" title="Output Options">
            <param argument="--no_file_comments" type="boolean" truevalue="--no_file_comments" falsevalue="" checked="true"
                label="Exclude header lines and stats from output files"/>
            <param argument="--report_orthologs" type="boolean" truevalue="--report_orthologs" falsevalue="" checked="false"
                label="Output a file with the list of orthologs for each hits"/>
            <param argument="--md5" type="boolean" truevalue="--md5" falsevalue="" checked="false"
                label="Add md5 hash of each query to annotations"/>
        </section>
    </inputs>
    <outputs>
        <data name="seed_orthologs" format="tabular" label="${tool.name} on ${on_string}: seed_orthologs" from_work_dir="results.emapper.seed_orthologs">
            <filter>ortho_method['m'] not in ['no_search', 'cache']</filter>
            <actions>
                <action name="column_names" type="metadata" default="@SEED_ORTHOLOG_COLUMNS@"/>
            </actions>
        </data>
        <data name="annotations" format="tabular" label="${tool.name} on ${on_string}: annotations" from_work_dir="results.emapper.annotations">
            <filter>annotation_options['no_annot'] == ''</filter>
            <actions>
                <conditional name="output_options.md5">
                    <when value="True">
                        <action name="column_names" type="metadata" default="query,seed_ortholog,evalue,score,max_annot_lvl,COG_category,Description,Preferred_name,GOs,EC,KEGG_ko,KEGG_Pathway,KEGG_Module,KEGG_Reaction,KEGG_rclass,BRITE,KEGG_TC,CAZy,BiGG_Reaction,PFAMseggNOG_OGs,md5"/>
                    </when>
                    <when value="False">
                        <action name="column_names" type="metadata" default="query,seed_ortholog,evalue,score,max_annot_lvl,COG_category,Description,Preferred_name,GOs,EC,KEGG_ko,KEGG_Pathway,KEGG_Module,KEGG_Reaction,KEGG_rclass,BRITE,KEGG_TC,CAZy,BiGG_Reaction,PFAMseggNOG_OGs"/>
                    </when>
                </conditional>
            </actions>
        </data>
        <data name="annotations_orthologs" format="tabular" label="${tool.name} on ${on_string}: orthologs"  from_work_dir="results.emapper.orthologs">
            <filter>ortho_method['m'] != 'cache' and output_options['report_orthologs']</filter>
            <actions>
                <action name="column_names" type="metadata" default="query,orth_type,species,orthologs"/>
            </actions>
        </data>
        <data name="no_annotations" format="fasta" label="${tool.name} on ${on_string}: Sequences wo annotation"  from_work_dir="results.emapper.no_annotations.fasta">
            <filter>ortho_method['m'] == 'cache' and output_options['output_no_annotations']</filter>
        </data>
    </outputs>
    <tests>
        <!-- test producing only seed orthologs-->
        <test expect_num_outputs="1">
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/>
            <conditional name="ortho_method">
                <param name="input" value="Nmar_0135.fa" ftype="fasta"/>
            </conditional>
            <conditional name="annotation_options">
                <param name="no_annot" value="--no_annot"/>
            </conditional>
            <section name="output_options">
                <param name="no_file_comments" value="true"/>
            </section>
            <expand macro="seed_orthologs_assertion"/>
            <expand macro="stdout_assertion"/>
        </test>
        
        <!-- test producing annotations form seed orthologs -->
        <test expect_num_outputs="2">
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/> <!-- not passed in test, but required for test to work -->
            <conditional name="ortho_method">
                <param name="m" value="no_search"/>
                <param name="annotate_hits_table" value="DIA_nlim.emapper.seed_orthologs" ftype="tabular">
                    <!-- this has no effect at the moment since column_names can not be set in uploads <metadata name="column_names" value="@SEED_ORTHOLOG_COLUMNS@"/> -->
                </param>
            </conditional>
            <conditional name="annotation_options">
            </conditional>
            <section name="output_options">
                <param name="report_orthologs" value="true"/>
                <param name="no_file_comments" value="true"/>
            </section>
            <expand macro="annotations_assertion"/>
            <expand macro="annotations_orthologs_assertion"/>
            <expand macro="stdout_assertion"/>
        </test>

        <!-- test producing seed orthologs and annotations at once (ie the same as the previous two tests but in a single tool run) -->
        <test expect_num_outputs="3">
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/> <!-- not passed in test, but required for test to work -->
            <conditional name="ortho_method">
                <param name="input" value="Nmar_0135.fa" ftype="fasta"/>
            </conditional>
            <section name="output_options">
                <param name="report_orthologs" value="true"/>
                <param name="no_file_comments" value="true"/>
                <param name="md5" value="true"/>
            </section>
            <expand macro="seed_orthologs_assertion"/>
            <expand macro="annotations_assertion" columns="22" add_metadata_columm_names=",md5" add_column_names="&#009;md5" add_column_re="\t[\d\w]+"/>
            <expand macro="annotations_orthologs_assertion"/>
            <expand macro="stdout_assertion"/>
        </test>
        
        <!-- test using chached annotations from previous run -->
        <test expect_num_outputs="2">
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/> <!-- not passed in test, but required for test to work -->
            <conditional name="ortho_method">
                <param name="m" value="cache"/>
                <param name="input" value="Nmar_0135.fa" ftype="fasta"/>
                <param name="cache" value="DIA_nlim.emapper.annotations_cached" ftype="tabular"/>
            </conditional>
            <section name="output_options">
                <param name="report_orthologs" value="true"/>
                <param name="no_file_comments" value="true"/>
                <param name="md5" value="true"/>
            </section>
            <expand macro="annotations_assertion" columns="22" add_metadata_columm_names=",md5" add_column_names="&#009;md5" add_column_re="\t[\d\w]+"/>
            <output name="no_annotations" ftype="fasta">
                <assert_contents>
                    <has_n_lines n="0"/>
                </assert_contents>
            </output>
            <expand macro="stdout_assertion"/>
        </test>

        <!-- test setting tax scope-->
        <test expect_num_outputs="3">
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/> <!-- not passed in test, but required for test to work -->
            <conditional name="ortho_method">
                <param name="input" value="Nmar_0135.fa" ftype="fasta"/>
            </conditional>
            <section name="annotation_options">
                <param name="tax_scope" value="651137" />
            </section>
            <section name="output_options">
                <param name="report_orthologs" value="true"/>
                <param name="no_file_comments" value="true"/>
            </section>
            <expand macro="seed_orthologs_assertion"/>
            <expand macro="annotations_assertion"/>
            <expand macro="annotations_orthologs_assertion"/>
            <expand macro="stdout_assertion">
                <has_text text="--tax_scope=651137"/>
            </expand>
        </test>

        <!-- test setting a diamond option-->
        <test expect_num_outputs="3">
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/> <!-- not passed in test, but required for test to work -->
            <conditional name="ortho_method">
                <param name="m" value="diamond" />
                <param name="input" value="Nmar_0135.fa" ftype="fasta"/>
                <param name="sensmode" value="fast" />
            </conditional>
            <section name="output_options">
                <param name="report_orthologs" value="true"/>
                <param name="no_file_comments" value="true"/>
            </section>
            <expand macro="seed_orthologs_assertion"/>
            <expand macro="annotations_assertion"/>
            <expand macro="annotations_orthologs_assertion"/>
            <expand macro="stdout_assertion">
                <has_text text="--sensmode fast"/>
            </expand>
        </test>
        <!-- not enabled as it requires a specific .db file, hard to minimize -->
        <!--test>
            <param name="eggnog_data" value="@EGGNOG_DB_VERSION@"/>
            <param name="input" value="Nmar_0135.fa" ftype="fasta"/>
            <section name="seed_ortho_options">
                <conditional name="ortho_method">
                    <param name="m" value="mmseqs" />
                    <param name="start_sens" value="4" />
                </conditional>
            </section>
            <param name="report_orthologs" value="true"/>
            <param name="no_file_comments" value="true"/>
            <output name="seed_orthologs" file="DIA_nlim.emapper.seed_orthologs" ftype="tabular" compare="sim_size"/>
            <output name="annotations" file="DIA_nlim.emapper.annotations" ftype="tabular" compare="sim_size"/>
            <output name="annotations_orthologs" file="DIA_nlim.emapper.annotations_orthologs" ftype="tabular"/>
        </test-->
    </tests>
    <help><![CDATA[

eggnog-mapper
=============
Overview
--------

``eggnog-mapper`` is a tool for fast functional annotation of novel sequences (genes or proteins) using precomputed eggNOG-based orthology assignments.
Obvious examples include the annotation of novel genomes, transcriptomes or even metagenomic gene catalogs.
The use of orthology predictions for functional annotation is considered more precise than traditional homology searches,
as it avoids transferring annotations from paralogs (duplicate genes with a higher chance of being involved in functional divergence).

EggNOG-mapper is also available as a public online resource:  `<http://beta-eggnogdb.embl.de/#/app/emapper>`_.

Outputs
-------

**seed_orthologs**

each line in the file provides the best match of each query within the best Orthologous Group (OG)
reported in the [project].hmm_hits file, obtained running PHMMER against all sequences within the best OG.
The seed ortholog is used to fetch fine-grained orthology relationships from eggNOG.
If using the diamond search mode, seed orthologs are directly
obtained from the best matching sequences by running DIAMOND against the whole eggNOG protein space.

**annotations**

This file provides final annotations of each query. Tab-delimited columns in the file are:

- ``query_name``: query sequence name
- ``seed_eggNOG_ortholog``: best protein match in eggNOG
- ``seed_ortholog_evalue``: best protein match (e-value)
- ``seed_ortholog_score``: best protein match (bit-score)
- ``predicted_taxonomic_group``
- ``predicted_protein_name``: Predicted protein name for query sequences
- ``GO_terms``: Comma delimited list of predicted Gene Ontology terms
- ``EC_number``
- ``KEGG_KO``
- ``KEGG_Pathway``: Comma delimited list of predicted KEGG pathways
- ``KEGG_Module``
- ``KEGG_Reaction``
- ``KEGG_rclass``
- ``BRITE``
- ``KEGG_TC``
- ``CAZy``
- ``BiGG_Reactions``
- ``Annotation_tax_scope``: The taxonomic scope used to annotate this query sequence
- ``Matching_OGs``: Comma delimited list of matching eggNOG Orthologous Groups
- ``best_OG|evalue|score``: Best matching Orthologous Groups (deprecated, use smallest from eggnog OGs)
- ``COG_functional_categories``: COG functional category inferred from best matching OG
- ``eggNOG_free_text_description``

**Recommentation for large input data**

EggNOG-mapper consists of two phases

1. finding seed orthologous sequences (compute intensive)
2. expanding annotations (IO intensive)

by default (i.e. if *Method to search seed orthologs* is not *Skip search stage...* and *Annotate seed orthologs* is *Yes*)
both phases are executed within one tool run. 

For large input FASTA datasets in can be favourable to split this in two separate
tool runs as follows:

1. Split the FASTA (e.g. 1M seqs per data set)
2. Run the search phase only (set *Annotate seed orthologs* to *No*) on the separate FASTA files.
3. Run the annotation phase (set *Method to search seed orthologs* to *Skip search stage...*)

See [also](https://github.com/eggnogdb/eggnog-mapper/wiki/eggNOG-mapper-v2.1.5-to-v2.1.8#Setting_up_large_annotation_jobs)

Another alternative is to use cached annotations (produced in a run with --md5 enabled).


    ]]></help>
    <expand macro="citations"/>
</tool>