diff pfamscan.xml @ 0:e7541ea105da draft default tip

planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/pfamscan commit c27a05e92e6e03545903dd2ff90976f8cab5ebf7
author bgruening
date Sat, 04 Feb 2023 16:33:07 +0000
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/pfamscan.xml	Sat Feb 04 16:33:07 2023 +0000
@@ -0,0 +1,158 @@
+<tool id="pfamscan" name="PfamScan" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@">
+    <description>search a FASTA sequence against a library of Pfam HMM</description>
+    <macros>
+        <token name="@TOOL_VERSION@">1.6</token>
+        <token name="@VERSION_SUFFIX@">0</token>
+    </macros>
+    <xrefs>
+        <xref type="bio.tools">pfamscan</xref>
+    </xrefs>
+    <requirements>
+        <requirement type="package" version="@TOOL_VERSION@">pfam_scan</requirement>
+        <requirement type="package" version="3.3.2">hmmer</requirement>
+    </requirements>
+    <version_command>pfam_scan.pl --version</version_command>
+    <command detect_errors="exit_code"><![CDATA[
+        mkdir -p './pfam_files' &&
+        ln -s '${pfam_library}' './pfam_files/Pfam-A.hmm' &&
+        ln -s '${pfam_data}' './pfam_files/Pfam-A.hmm.dat' &&
+        #if $active_sites.selector == 'true'
+            ln -s '${active_sites.active_file}' './pfam_files/active_site.dat' &&
+        #end if
+        hmmpress './pfam_files/Pfam-A.hmm' &&
+        pfam_scan.pl 
+            -cpu \${GALAXY_SLOTS:-4}
+            -fasta '${fasta}'
+            -dir './pfam_files/'
+            -outfile './output.tab'
+            #if $advanced_options.e_seq
+                -e_seq $advanced_options.e_seq
+            #end if
+            #if $advanced_options.e_dom
+                -e_dom $advanced_options.e_dom
+            #end if
+            #if $advanced_options.b_seq
+                -b_seq $advanced_options.b_seq
+            #end if
+            #if $advanced_options.b_dom
+                -b_dom $advanced_options.b_dom
+            #end if
+            #if $active_sites.selector == 'true'
+             -as
+            #end if
+            && tail -n +28 "./output.tab" > "./output_fixed.tab"
+        ]]>
+    </command>
+    <inputs>
+        <param argument="-fasta" type="data" format="fasta" label="Protein sequences FASTA file"/>
+        <param name="pfam_library" type="data" format="hmm3" label="Pfam-A HMM library" help="Pfam-A HMMs in an HMM library searchable with the hmmscan program." />
+        <param name="pfam_data" type="data" format="stockholm" label="Pfam-A HMM Stockholm file" help="Stockholm format is a multiple sequence alignment format used 
+            by Pfam, Rfam and Dfam, to disseminate protein, RNA and DNA sequence alignments." />
+        <conditional name="active_sites">
+            <param name="selector" type="select" label="Predict active site residues" help="Predict active site residues for Pfam-A matches">
+                <option value="false">Disabled</option>
+                <option value="true" selected="true">Enabled</option>
+            </param>
+            <when value="false"/>
+            <when value="true">
+                <param name="active_file" type="data" format="txt" label="Active sites file" help="This file is required for predicting the active site residues." />
+            </when>
+        </conditional>
+        <section name="advanced_options" title="Advanced options">
+            <param argument="-e_seq" type="float" min="0" value=""  optional="true" label="Hmmscan evalue sequence cutoff" help="Specify hmmscan evalue sequence cutoff 
+                for Pfam-A searches (default Pfam defined)" />
+            <param argument="-e_dom" type="float" min="0" value="" optional="true" label="Hmmscan evalue domain cutoff" help="Specify hmmscan evalue domain cutoff for 
+                Pfam-A searches (default Pfam defined)" />
+            <param argument="-b_seq" type="float" min="0" value=""  optional="true" label="Hmmscan bit score sequence cutoff" help="Specify hmmscan bit score sequence 
+                cutoff for Pfam-A searches (default Pfam defined)" />
+            <param argument="-b_dom" type="float" min="0" value=""  optional="true" label="Hmmscan bit score domain cutoff" help="Specify hmmscan bit score domain cutoff 
+                for Pfam-A searches (default Pfam defined)" />
+            <param argument="-clan_overlap" type="boolean" truevalue="-clan_overlap" falsevalue="" checked="false" label="Clan overlap" help="Show overlapping hits within 
+                clan member families (applies to Pfam-A families only)" />
+            <param argument="-align" type="boolean" truevalue="-align" falsevalue="" checked="false" label="Show alignment" help="Show the HMM-sequence alignment for each match" />
+        </section>
+    </inputs>
+    <outputs>
+        <data name="output" format="tabular" from_work_dir="output_fixed.tab" label="${tool.name} on ${on_string}"/>
+    </outputs>
+    <tests>
+        <!-- Test 01: Default parameters not active sites -->
+        <test expect_num_outputs="1">
+            <param name="fasta" value="sequences.fasta.gz"/>
+            <param name="pfam_library" value="Pfam-A.hmm.gz"/>
+            <param name="pfam_data" value="Pfam-A.hmm.dat.gz"/>
+            <conditional name="active_sites">
+                <param name="selector" value="false"/>
+            </conditional>
+            <output name="output" file="test01.tab" ftype="tabular"/>
+        </test>
+        <!-- Test 02: Default parameters: active sites -->
+        <test expect_num_outputs="1">
+            <param name="fasta" value="sequences.fasta.gz"/>
+            <param name="pfam_library" value="Pfam-A.hmm.gz"/>
+            <param name="pfam_data" value="Pfam-A.hmm.dat.gz"/>
+            <conditional name="active_sites">
+                <param name="selector" value="true"/>
+                <param name="active_file" value="active_site.dat.gz"/>
+            </conditional>
+            <output name="output" file="test02.tab" ftype="tabular"/>
+        </test>
+        <!-- Test 03: Non default parameters: active sites -->
+        <test expect_num_outputs="1">
+            <param name="fasta" value="sequences.fasta.gz"/>
+            <param name="pfam_library" value="Pfam-A.hmm.gz"/>
+            <param name="pfam_data" value="Pfam-A.hmm.dat.gz"/>
+            <section name="advanced_options">
+                <param name="e_seq" value="1.2"/>
+                <param name="e_dom" value="1.3"/>
+                <param name="clan_overlap" value="true"/>
+            </section>
+            <conditional name="active_sites">
+                <param name="selector" value="true"/>
+                <param name="active_file" value="active_site.dat.gz"/>
+            </conditional>
+            <output name="output" file="test03.tab" ftype="tabular"/>
+        </test>
+        <!-- Test 04: Non default parameters: active sites -->
+        <test expect_num_outputs="1">
+            <param name="fasta" value="sequences.fasta.gz"/>
+            <param name="pfam_library" value="Pfam-A.hmm.gz"/>
+            <param name="pfam_data" value="Pfam-A.hmm.dat.gz"/>
+            <section name="advanced_options">
+                <param name="b_seq" value="1.4"/>
+                <param name="b_dom" value="1.2"/>
+                <param name="align" value="true"/>
+            </section>
+            <conditional name="active_sites">
+                <param name="selector" value="true"/>
+                <param name="active_file" value="active_site.dat.gz"/>
+            </conditional>
+            <output name="output" file="test04.tab" ftype="tabular"/>
+        </test>
+    </tests>
+    <help><![CDATA[
+
+.. class:: infomark
+
+**Purpose**
+
+Search one or more sequences for matching Pfam domains. Depending on the user options, the script can also process the results such that overlaps between families belonging to the 
+same clan are resolved and can predict active sites.
+
+----
+
+.. class:: infomark
+
+**Required files**
+
+To run PfamScan you will need to download the following files from the Pfam ftp site:
+
+- Pfam-A HMMs in an HMM library searchable with the hmmscan program: `Pfam-A.hmm.gz <https://ftp.ebi.ac.uk/pub/databases/Pfam/current_release/Pfam-A.hmm.gz>`_
+- Pfam-A HMM Stockholm file associated with each HMM required for PfamScan: `Pfam-A.hmm.dat.gz <https://ftp.ebi.ac.uk/pub/databases/Pfam/current_release/Pfam-A.hmm.dat.gz>`_
+- Active sites: `active_sites.dat.gz <ftp://ftp.ebi.ac.uk/pub/databases/Pfam/current_release/active_site.dat.gz>`_
+
+]]></help>
+    <citations>
+        <citation type="doi">10.1093/nar/gkt006</citation>
+    </citations>
+</tool>