view phylogenomics_analysis.xml @ 25:61c90d75c7cd draft

Uploaded

<tool id="plant_tribes_phylogenomics_analysis" name="PhylogenomicsAnalysis" version="1.0.0">
    <description>pipeline</description>
    <requirements>
        <requirement type="package" version="0.2">plant_tribes_phylogenomics_analysis</requirement>
    </requirements>
    <stdio>
        <!-- Anything other than zero is an error -->
        <exit_code range="1:" />
        <!-- In case the return code has not been set properly check stderr too -->
        <regex match="Error:" />
        <regex match="Exception:" />
    </stdio>
    <command>
        <![CDATA[
            #import os
            #import sys

            #set invalid_input_msg = "The selected input is not a valid list of sequences classified into gene family clusters.  Select an input with the label 'Sequences classified into gene family clusters on...'."
            #set no_coding_sequences_msg = "No corresponding coding sequences were generated for the selected input.  You must not have selected 'Yes' for the 'Create corresponding coding sequences?' option in the GeneFamilyClassifier tool for the selected input."

            #set input_extra_files_path = $orthogroup_faa.extra_files_path
            #if not $os.path.exists($input_extra_files_path):
                #raise Exception($invalid_input_msg)
            #end if

            #set num_orthogroup_faa = $len([f for f in $os.listdir($input_extra_files_path) if f.endswith('.faa')])
            #if $num_orthogroup_faa == 0:
                #raise Exception($invalid_input_msg)
            #end if

            #set num_orthogroup_fna = $len([f for f in $os.listdir($input_extra_files_path) if f.endswith('.fna')])

            #set src_fasta_dir = $os.path.join('phylogenomicsAnalysis_dir', 'orthogroups_fasta')
            #set dest_fasta_dir = $output_fasta.extra_files_path

            #if str($options_type.options_type_selector) == 'advanced' and str($options_type.multiple_sequence_alignments_cond.multiple_sequence_alignments) == 'yes':
                #set generate_output_aln = True
                #set src_aln_dir = $os.path.join('phylogenomicsAnalysis_dir', 'orthogroups_aln')
                #set dest_aln_dir = $output_aln.extra_files_path
                mkdir -p $dest_aln_dir &&
            #else:
                #set generate_output_aln = False
            #end if

            #if str($options_type.options_type_selector) == 'advanced' and str($options_type.phylogenetic_trees_cond.phylogenetic_trees) == 'yes':
                #set generate_output_tree = True
                #set src_tree_dir = $os.path.join('phylogenomicsAnalysis_dir', 'orthogroups_tree')
                #set dest_tree_dir = $output_tree.extra_files_path
                mkdir -p $dest_tree_dir &&
            #else:
                #set generate_output_tree = False
            #end if

            mkdir -p $dest_fasta_dir &&

            PhylogenomicsAnalysis
            --num_threads "\${GALAXY_SLOTS:-4}"
            --orthogroup_faa "$orthogroup_faa.extra_files_path"
            --scaffold_dir "${GALAXY_DATA_INDEX_DIR}/plant_tribes/scaffolds"
            --scaffold "$scaffold"
            --method $method

            #if str($options_type.options_type_selector) == 'advanced':
                #if str($options_type.multiple_sequence_alignments_cond.multiple_sequence_alignments) == 'yes':
                    #set multiple_sequence_alignments_option_cond = $options_type.multiple_sequence_alignments_cond.multiple_sequence_alignments_option_cond
                    #set multiple_sequence_alignments_option = $multiple_sequence_alignments_option_cond.multiple_sequence_alignments_option
                    #set input_includes_coding_sequences_cond = $options_type.multiple_sequence_alignments_cond.input_includes_coding_sequences_cond
                    #set input_includes_coding_sequences = $input_includes_coding_sequences_cond.input_includes_coding_sequences
                    #if str($input_includes_coding_sequences) == 'yes':
                        #set multiple_codon_alignments_cond = $input_includes_coding_sequences_cond.multiple_codon_alignments_cond
                        #set multiple_codon_alignments = $multiple_codon_alignments_cond.multiple_codon_alignments
                        #set use_corresponding_coding_sequences_cond = $multiple_codon_alignments_cond.use_corresponding_coding_sequences_cond
                        #set use_corresponding_coding_sequences = $use_corresponding_coding_sequences_cond.use_corresponding_coding_sequences
                    #end if
                    #if str($multiple_sequence_alignments_option) == 'create_alignments':
                        --create_alignments
                    #else if str($multiple_sequence_alignments_option) == 'add_alignments':
                        --add_alignments
                    #else if str($multiple_sequence_alignments_option) == 'pasta_alignments':
                        --pasta_alignments
                        --pasta_iter_limit $multiple_sequence_alignments_option_cond.pasta_iter_limit
                    #end if
                    #if str($input_includes_coding_sequences) == 'yes':
                        #if str($multiple_codon_alignments) == 'yes':
                            --codon_alignments
                            #if str($use_corresponding_coding_sequences) == 'yes':
                                #if $num_orthogroup_fna == 0:
                                    #raise Exception($no_coding_sequences_msg)
                                #end if
                                --orthogroup_fna
                            #end if
                            --sequence_type $multiple_codon_alignments_cond.sequence_type
                        #end if
                    #end if
                #end if

                #if str($options_type.phylogenetic_trees_cond.phylogenetic_trees) == 'yes':
                    #set tree_inference_cond = $options_type.phylogenetic_trees_cond.tree_inference_cond
                    #set tree_inference_method = $tree_inference_cond.tree_inference
                    --tree_inference $tree_inference_method
                    #if str($tree_inference_method) == 'raxml':
                        #set rooting_order_file_cond = $tree_inference_cond.rooting_order_file_cond
                        #set rooting_order_file = $rooting_order_file_cond.rooting_order_file
                        #if str($rooting_order_file) == 'yes':
                            --rooting_order "$rooting_order_file.rooting_order"
                        #end if
                        --bootstrap_replicates $tree_inference_cond.bootstrap_replicates
                    #end if
                    --max_orthogroup_size $options_type.phylogenetic_trees_cond.max_orthogroup_size
                    --min_orthogroup_size $options_type.phylogenetic_trees_cond.min_orthogroup_size
                #end if

                #if str($options_type.msa_quality_control_options_cond.msa_quality_control_options) == 'yes':
                    #set msa_quality_control_options_cond = $options_type.msa_quality_control_options_cond
                    #set remove_gappy_sequences_cond = $msa_quality_control_options_cond.remove_gappy_sequences_cond
                    #if str($remove_gappy_sequences_cond) == 'yes':
                        #set trim_type_cond = $remove_gappy_sequences_cond.trim_type_cond
                        --remove_sequences $remove_gappy_sequences_cond.remove_sequences
                        #if str($trim_type_cond.trim_type) == 'automated_trimming':
                            --automated_trimming
                        else:
                            --gap_trimming $trim_type_cond.gap_trimming
                        #end if
                    #end if
                #end if
            #end if

            && echo "Sequence alignments and corresponding phylogenies: `ls $src_fasta_dir | wc -l` files" > $output_fasta
            && ls $src_fasta_dir >> $output_fasta
            && mv $src_fasta_dir/* $dest_fasta_dir

            #if $generate_output_aln:
                && echo "Multiple sequence alignments using method '$multiple_sequence_alignments_option': `ls $src_aln_dir | wc -l` files" > $output_aln
                && ls $src_aln_dir >> $output_aln
                && mv $src_aln_dir/* $dest_aln_dir
            #end if

            #if $generate_output_tree:
                && echo "Phylogenetic trees: `ls $src_tree_dir | wc -l` files" > $output_tree
                && ls $src_tree_dir >> $output_tree
                && mv $src_tree_dir/* $dest_tree_dir
            #end if
        ]]>
    </command>
    <inputs>
        <param name="orthogroup_faa" format="txt" type="data" label="Select sequences classified into gene family clusters"/>
        <param name="scaffold" type="select" label="Orthogroups or gene families proteins scaffold">
            <options from_data_table="plant_tribes_scaffolds" />
            <validator type="no_options" message="No PlantTribes scaffolds are available.  Use the PlantTribes Scaffolds Download Data Manager tool to install and populate the PlantTribes scaffolds data table."/>
        </param>
        <param name="method" type="select" label="Protein clustering method">
            <option value="gfam" selected="true">GFam</option>
            <option value="orthofinder">OrthoFinder</option>
            <option value="orthomcl">OrthoMCL</option>
        </param>
        <conditional name="options_type">
            <param name="options_type_selector" type="select" label="Options Configuration">
                <option value="basic" selected="true">Basic</option>
                <option value="advanced">Advanced</option>
            </param>
            <when value="basic" />
            <when value="advanced">
                <conditional name="multiple_sequence_alignments_cond">
                    <param name="multiple_sequence_alignments" type="select" label="Set multiple sequence alignments options?">
                        <option value="no" selected="true">No</option>
                        <option value="yes">Yes</option>
                    </param>
                    <when value="no" />
                    <when value="yes">
                        <conditional name="multiple_sequence_alignments_option_cond">
                            <param name="multiple_sequence_alignments_option" label="Select process for multiple sequence alignments" type="select" force_select="True">
                                <option value="create_alignments">Create orthogroup protein multiple sequence alignments including scaffold backbone proteins (MAFFT algorithm)</option>
                                <option value="add_alignments">Add unaligned orthogroup proteins to scaffold backbone multiple sequence alignments (MAFFT algorithm)</option>
                                <option value="pasta_alignments">Create orthogroup protein multiple sequence alignments including scaffold backbone proteins (PASTA algorithm)</option>
                            </param>
                            <when value="create_alignments"/>
                            <when value="add_alignments"/>
                            <when value="pasta_alignments">
                                <param name="pasta_iter_limit" type="integer" value="3" min="0" label="Maximum number of iteration that the PASTA algorithm will run"/>
                            </when>
                        </conditional>
                        <conditional name="input_includes_coding_sequences_cond">
                            <param name="input_includes_coding_sequences" type="select" label="Input sequences include corresponding coding sequences?" help="Selected 'Yes' for the 'Create corresponding coding sequences?' option in the GeneFamilyClassifier tool for the selected input">
                                <option value="no" selected="true">No</option>
                                <option value="yes">Yes</option>
                            </param>
                            <when value="no"/>
                            <when value="yes">
                                <conditional name="multiple_codon_alignments_cond">
                                    <param name="multiple_codon_alignments" label="Construct orthogroup multiple codon alignments?" type="select">
                                        <option value="no" selected="True">No</option>
                                        <option value="yes">Yes</option>
                                    </param>
                                    <when value="no"/>
                                    <when value="yes">
                                        <param name="sequence_type" type="select" label="Sequence type used in the phylogenetic inference - (dna)">
                                            <option value="protein" selected="true">Amino acid based</option>
                                            <option value="dna">Nucleotide based</option>
                                        </param>
                                        <conditional name="use_corresponding_coding_sequences_cond">
                                            <param name="use_corresponding_coding_sequences" type="select" label="Use corresponding coding sequences?">
                                                <option value="no" selected="true">No</option>
                                                <option value="yes">Yes</option>
                                            </param>
                                            <when value="no" />
                                            <when value="yes" />
                                        </conditional>
                                    </when>
                                </conditional>
                            </when>
                        </conditional>
                    </when>
                </conditional>
                <conditional name="phylogenetic_trees_cond">
                    <param name="phylogenetic_trees" type="select" label="Set phylogenetic trees options?">
                        <option value="no" selected="true">No</option>
                        <option value="yes">Yes</option>
                    </param>
                    <when value="no" />
                    <when value="yes">
                        <conditional name="tree_inference_cond">
                            <param name="tree_inference" type="select" label="Phylogenetic trees inference method">
                                <option value="raxml" selected="true">RAxML</option>
                                <option value="fasttree">FastTree</option>
                            </param>
                            <when value="raxml">
                                <conditional name="rooting_order_file_cond">
                                    <param name="rooting_order_file" type="select" label="Select history item containing strings matching sequence identifiers of species for determining the most basal taxa in the orthogroups for rooting trees?">
                                        <option value="no" selected="true">No</option>
                                        <option value="yes">Yes</option>
                                    </param>
                                    <when value="no"/>
                                    <when value="yes">
                                        <param name="rooting_order" format="txt" type="data" label="History item containing strings matching sequence identifiers of species for determining the most basal taxa in the orthogroups for rooting trees"/>
                                    </when>
                                </conditional>
                                <param name="bootstrap_replicates" type="integer" value="100" min="0" label="Number of replicates for rapid bootstrap analysis and search for the best-scoring ML tree"/>
                            </when>
                            <when value="fasttree"/>
                        </conditional>
                        <param name="max_orthogroup_size" type="integer" value="100" min="0" label="Maximum number of sequences in orthogroup alignments"/>
                        <param name="min_orthogroup_size" type="integer" value="4" min="0" label="Minimum number of sequences in orthogroup alignments"/>
                    </when>
                </conditional>
                <conditional name="msa_quality_control_options_cond">
                    <param name="msa_quality_control_options" type="select" label="Set MSA quality control options?">
                        <option value="no" selected="true">No</option>
                        <option value="yes">Yes</option>
                    </param>
                    <when value="no" />
                    <when value="yes">
                        <conditional name="remove_gappy_sequences_cond">
                            <param name="remove_gappy_sequences" type="select" label="Remove gappy sequences in alignments?">
                                <option value="no" selected="true">No</option>
                                <option value="yes">Yes</option>
                            </param>
                            <when value="no"/>
                            <when value="yes">
                                <param name="remove_sequences" type="float" value="0.5" min="0" max="1" label="Remove sequences with gaps of" help="0.5 removes sequences with 50% gaps"/>
                                <conditional name="trim_type_cond">
                                    <param name="trim_type" type="select" label="Select process used for gap trimming">
                                        <option value="automated_trimming" selected="true">Trim alignments using trimAl's ML heuristic trimming approach</option>
                                        <option value="gap_trimming">Nucleotide based</option>
                                    </param>
                                    <when value="automated_trimming"/>
                                    <when value="gap_trimming">
                                        <param name="gap_trimming" type="float" value="0.1" min="0" max="1.0" label="Remove sites in alignments with gaps of" help="0.1 removes sites with 90% gaps"/>
                                    </when>
                                </conditional>
                            </when>
                        </conditional>
                    </when>
                </conditional>
            </when>
        </conditional>
    </inputs>
    <outputs>
        <data name="output_fasta" format="txt" label="Sequence alignments and corresponding phylogenies on ${on_string}"/>
        <data name="output_aln" format="txt" label="Sequence alignments and corresponding phylogenies using PASTA on ${on_string}">
            <filter>options_type['options_type_selector'] == 'advanced' and options_type['multiple_sequence_alignments_cond']['multiple_sequence_alignments'] == 'yes'</filter>
        </data>
        <data name="output_tree" format="txt" label="Phylogenetic trees on ${on_string}">
            <filter>options_type['options_type_selector'] == 'advanced' and options_type['phylogenetic_trees_cond']['phylogenetic_trees'] == 'yes'</filter>
        </data>
    </outputs>
    <tests>
        <test>
        </test>
    </tests>
    <help>
This tool is one of the PlantTribes' collection of automated modular analysis pipelines that utilize objective classifications of
complete protein sequences from sequenced plant genomes to perform comparative evolutionary studies. It performs phylogenomic
analyses, comparing and analyzing the sequences of single genes, or a small number of genes, as well as many other types of data.

-----

**Options**

* **Required options**

 - **Select sequences classified into gene family clusters** - A history item with the label 'Sequences classified into gene family clusters on...'.
 - **Orthogroups or gene families proteins scaffold** - PlantTribes scaffolds data.
 - **Protein clustering method** - One of GFam, OrthoFinder or OrthoMCL.

* **Multiple sequence alignments options**

 - **Select process for multiple sequence alignments** - Process used for setting multiple sequence alignments.
 - **Input sequences include corresponding coding sequences?** - Selecting 'Yes' for this option requires that 'Yes' was selected for the 'Create orthogroup fasta files?' option on the GeneFamilyClassifier tool when generating the selected input.  This tool will produce an error if 'No' was selected.
 - **Construct orthogroup multiple codon alignments?** - Construct orthogroup multiple codon alignments.
 - **Sequence type used in the phylogenetic inference** - Sequence type used in the phylogenetic inference.
 - **Use corresponding coding sequences?** - Selecting 'Yes' for this option requires that 'Yes' was selected for the 'Create corresponding coding sequences?' option on the GeneFamilyClassifier tool when generating the selected input.  This tool will produce an error if 'No' was selected.

* **Phylogenetic trees options**

 - **Phylogenetic trees inference method** - Phylogenetic trees inference method.
 - **Select history item containing strings matching sequence identifiers of species for determining the most basal taxa in the orthogroups for rooting trees?** - If 'No' is selected, trees will be rooted using the most distant taxon present in the orthogroup.
 - **History item containing strings matching sequence identifiers of species for determining the most basal taxa in the orthogroups for rooting trees** -History item containing a set of string fragments matching sequences identifiers of species in the classification (including scaffold taxa) to be used for determining the most basal taxa in the orthogroups for rooting trees. The set of string fragments must be listed in decreasing order from older to younger lineages.
 - **Number of replicates for rapid bootstrap analysis and search for the best-scoring ML tree** - Number of replicates for rapid bootstrap analysis and search for the best-scoring ML tree.
 - **Maximum number of sequences in orthogroup alignments** - Maximum number of sequences in orthogroup alignments.
 - **Minimum number of sequences in orthogroup alignments** - Minimum number of sequences in orthogroup alignments.

* **MSA quality control options**

 - **Remove sequences with gaps of** - Removes gappy sequences in alignments (i.e., 0.5 removes sequences with 50% gaps).
 - **Select process used for gap trimming** - Either nucleotide based trimming or alignments are trimed using using trimAl's ML heuristic trimming approach.
 - **Remove sites in alignments with gaps of** - If the process used for gap trimming is nucleotide based, this is the gap value used when removing gappy sites in alignments (i.e., 0.1 removes sites with 90% gaps).

    </help>
    <citations>
            <citation type="bibtex">
                @unpublished{None,
                author = {None},
                title = {None},
                year = {None},
                eprint = {None},
                url = {None}
            }</citation>
    </citations>
</tool>