phabox initial commit (Helmholtz-UFZ#68)

* phabox initial commit * review comments and add docs * add doc * add data manager
rmassei · Nov 12, 2024 · 3c8e23e · 3c8e23e
1 parent 129359f
commit 3c8e23e
Show file tree

Hide file tree

Showing 23 changed files with 937 additions and 0 deletions.
diff --git a/data_managers/data_manager_phabox/.lint_skip b/data_managers/data_manager_phabox/.lint_skip
@@ -0,0 +1 @@
+InputsMissing
diff --git a/data_managers/data_manager_phabox/.shed.yml b/data_managers/data_manager_phabox/.shed.yml
@@ -0,0 +1,10 @@
+categories:
+- Data Managers
+- Metagenomics
+description: Data managers for PhaBOX reference data
+homepage_url: https://github.com/KennthShang/PhaBOX
+long_description: Data managers for PhaBOX reference data
+name: phabox_build_database
+owner: ufz
+remote_repository_url: https://github.com/Helmholtz-UFZ/ufz-galaxy-tools/blob/main/data_managers/data_manager_phabox
+type: unrestricted
diff --git a/data_managers/data_manager_phabox/data_manager/phabox_datamanager.xml b/data_managers/data_manager_phabox/data_manager/phabox_datamanager.xml
@@ -0,0 +1,54 @@
+<tool id="phabox_build_database" name="PhaBOX" tool_type="manage_data" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
+    <description>database builder</description>
+    <macros>
+        <token name="@TOOL_VERSION@">2.1.5</token>
+        <token name="@DB_VERSION@">2</token>
+        <token name="@VERSION_SUFFIX@">0</token>
+        <token name="@PROFILE@">22.01</token>
+    </macros>
+    <requirements>
+        <requirement type="package" version="6.0">unzip</requirement>
+        <requirement type="package" version="1.21.4">wget</requirement>
+    </requirements>
+    <command detect_errors="exit_code"><![CDATA[
+        wget https://github.com/KennthShang/PhaBOX/releases/download/v2/phabox_db_v@[email protected] &&
+        unzip phabox_db_v@[email protected] &&
+        cp '$dmjson' '$out_file'
+    ]]></command>
+    <configfiles>
+        <configfile name="dmjson"><![CDATA[
+{
+  "data_tables":{
+    "phabox":[
+      {
+        "path":"phabox_db_v@DB_VERSION@",
+        "name":"Version @DB_VERSION@",
+        "version":"@DB_VERSION@",
+        "value":"@DB_VERSION@"
+      }
+    ]
+  }
+}]]>
+        </configfile>
+    </configfiles>
+    <inputs/>
+    <outputs>
+        <data name="out_file" format="data_manager_json" />
+    </outputs>
+    <tests>
+        <test expect_num_outputs="1">
+            <output name="out_file">
+                <assert_contents>
+                    <has_text text='"value":"@DB_VERSION@"'/>
+                    <has_text text='"name":"Version 2'/>
+                </assert_contents>
+            </output>
+        </test>
+    </tests>
+    <help><![CDATA[
+Download and extract PhaBOX reference data. The current tool version is used for versioning. For the user only the major version is shown
+    ]]></help>
+    <citations>
+        <citation type="doi">10.1093/bioadv/vbad101</citation>
+    </citations>
+</tool>
diff --git a/data_managers/data_manager_phabox/data_manager_conf.xml b/data_managers/data_manager_phabox/data_manager_conf.xml
@@ -0,0 +1,19 @@
+<data_managers>
+    <data_manager tool_file="data_manager/phabox_datamanager.xml" id="phabox_build_database">
+        <data_table name="phabox">
+            <output>
+                <column name="value"/>
+                <column name="name"/>
+                <column name="version"/>
+                <column name="path" output_ref="out_file">
+                    <move type="directory">
+                        <source>${path}</source>
+                        <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">phabox/${path}</target>
+                    </move>
+                    <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/phabox/${path}</value_translation>
+                    <value_translation type="function">abspath</value_translation>
+                </column>
+            </output>
+        </data_table>
+    </data_manager>
+</data_managers>
diff --git a/data_managers/data_manager_phabox/tool-data/phabox.loc.sample b/data_managers/data_manager_phabox/tool-data/phabox.loc.sample
@@ -0,0 +1,2 @@
+# Format:
+# value	name	version	path
diff --git a/data_managers/data_manager_phabox/tool_data_table_conf.xml.sample b/data_managers/data_manager_phabox/tool_data_table_conf.xml.sample
@@ -0,0 +1,7 @@
+<tables>
+    <!-- phabox reference data  -->
+    <table name="phabox" comment_char="#">
+        <columns>value, name, version, path</columns>
+        <file path="tool-data/phabox.loc" />
+    </table>
+</tables>
diff --git a/tools/phabox/.shed.yml b/tools/phabox/.shed.yml
@@ -0,0 +1,24 @@
+categories:
+- Metagenomics
+description: Identify and analyze phage contigs in metagenomic data
+long_description: |
+  PhaBOX can comprehensively identify and analyze phage contigs in metagenomic
+  data. It supports integrated phage analysis, including phage contig
+  identification from the metagenomic assembly, lifestyle prediction, taxonomic
+  classification, and host prediction. 
+name: phabox
+owner: ufz
+homepage_url: https://github.com/KennthShang/PhaBOX
+remote_repository_url: https://github.com/Helmholtz-UFZ/ufz-galaxy-tools/blob/main/tools/phabox
+type: unrestricted
+auto_tool_repositories:
+  name_template: "{{ tool_id }}"
+  description_template: "Wrapper for phabox task: {{ tool_name }}."
+suite:
+  name: "suite_phabox"
+  description: "A suite of tools that brings the phabox project into Galaxy."
+  long_description: |
+    PhaBOX can comprehensively identify and analyze phage contigs in metagenomic
+    data. It supports integrated phage analysis, including phage contig
+    identification from the metagenomic assembly, lifestyle prediction, taxonomic
+    classification, and host prediction. 
diff --git a/tools/phabox/cherry.xml b/tools/phabox/cherry.xml
@@ -0,0 +1,54 @@
+<tool id="phabox_cherry" name="PhaBOX cherry" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="23.1" license="MIT">
+    <description>Host prediction</description>
+    <macros>
+        <import>macros.xml</import>
+    </macros>
+    <xrefs>
+        <xref type="bio.tools">phabox</xref>
+    </xrefs>
+    <requirements>
+        <requirement type="package" version="@TOOL_VERSION@">phabox</requirement>
+    </requirements>
+    <command detect_errors="exit_code"><![CDATA[
+        @CRISPR_PRE@
+        phabox2 --task cherry
+            @GENERAL@
+            @NETWORK@
+            @CRISPR@
+    ]]></command>
+    <inputs>
+        <expand macro="general"/>
+        <expand macro="network"/>
+        <expand macro="crispr"/>
+    </inputs>
+    <outputs>
+        <data name="out" format="tabular" from_work_dir="output/final_prediction/cherry_prediction.tsv"/>
+    </outputs>
+    <tests>
+        <test>
+            <param name="dbdir" value="phaboxdb"/>
+            <param name="contigs" value="example_contigs.fa"/>
+            <output name="out">
+                <assert_contents>
+                    <has_line line="Accession&#9;Length&#9;Host&#9;CHERRYScore&#9;Method&#9;Host_NCBI_lineage&#9;Host_GTDB_lineage"/>
+                    <has_n_lines n="11"/>
+                    <has_n_columns n="7"/>
+                </assert_contents>
+            </output>
+        </test>
+    </tests>
+    <help><![CDATA[
+Predict hosts for viruses.
+
+
+@COMMON_INPUT_DOC@
+
+**Output**
+
+@COMMON_OUTPUT_DOC@
+@CHERRY_OUTPUT_DOC@
+    ]]></help>
+    <expand macro="citations">
+        <citation type="doi">10.1093/bib/bbac182</citation>
+    </expand>
+</tool>
diff --git a/tools/phabox/contamination.xml b/tools/phabox/contamination.xml
@@ -0,0 +1,58 @@
+<tool id="phabox_contamination" name="PhaBOX contamination" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="23.1" license="MIT">
+    <description>Contamination/provirus detection</description>
+    <macros>
+        <import>macros.xml</import>
+    </macros>
+    <xrefs>
+        <xref type="bio.tools">phabox</xref>
+    </xrefs>
+    <requirements>
+        <requirement type="package" version="@TOOL_VERSION@">phabox</requirement>
+    </requirements>
+    <command detect_errors="exit_code"><![CDATA[
+        phabox2 --task contamination
+            @GENERAL@
+            @CONTAMINATION@
+    ]]></command>
+    <inputs>
+        <expand macro="general"/>
+        <expand macro="contamination"/>
+    </inputs>
+    <outputs>
+        <data name="out" format="tabular" from_work_dir="output/final_prediction/contamination_prediction.tsv"/>
+    </outputs>
+    <tests>
+        <test>
+            <param name="dbdir" value="phaboxdb"/>
+            <param name="contigs" value="example_contigs.fa"/>
+            <output name="out">
+                <assert_contents>
+                    <has_line line="Accession&#9;Length&#9;Total_genes&#9;Viral_genes&#9;Prokaryotic_genes&#9;Kmer_freq&#9;Contamination&#9;Provirus&#9;Pure_viral"/>
+                    <has_n_lines n="11"/>
+                    <has_n_columns n="9"/>
+                </assert_contents>
+            </output>
+        </test>
+    </tests>
+    <help><![CDATA[
+
+Check for contaminations / proviruses.
+
+
+@COMMON_INPUT_DOC@
+
+**Output**:
+
+@COMMON_OUTPUT_DOC@
+- Total_genes: number of genes in the contigs (predicted by prodigal-gv)
+- Viral_genes: number of viral marker genes
+- Prokaryotic_genes: number of prokaryotic marker genes
+- Kmer_freq: average frequency of 20-mer. This is a value to estimate the copy number of the genes; usually, the Kmer_freq of 99.9% virus is less than 1.25.
+- Contamination:
+- Provirus: Whether the sequence is a provirus
+- Pure_viral: High quality or Medium quality or Low quality
+
+
+    ]]></help>
+    <expand macro="citations"/>
+</tool>
diff --git a/tools/phabox/end_to_end.xml b/tools/phabox/end_to_end.xml
@@ -0,0 +1,76 @@
+<tool id="phabox_end_to_end" name="PhaBOX end to end" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="23.1" license="MIT">
+    <description></description>
+    <macros>
+        <import>macros.xml</import>
+    </macros>
+    <xrefs>
+        <xref type="bio.tools">phabox</xref>
+    </xrefs>
+    <requirements>
+        <requirement type="package" version="@TOOL_VERSION@">phabox</requirement>
+    </requirements>
+    <command detect_errors="exit_code"><![CDATA[
+        @CRISPR_PRE@
+        phabox2 --task end_to_end
+            @GENERAL@
+            @PHAMER@
+            @NETWORK@
+            @CRISPR@
+    ]]></command>
+    <inputs>
+        <expand macro="general"/>
+        <expand macro="phamer"/>
+        <expand macro="network"/>
+        <expand macro="crispr"/>
+        <param name="supplements" type="select" optional="true" multiple="true" label="Output supplementary collections">
+            <option value="phamer">phamer</option>
+            <option value="phagcn">phagcn</option>
+            <option value="cherry">cherry</option>
+        </param>
+    </inputs>
+    <outputs>
+        <data name="end_to_end_out" format="tabular" from_work_dir="output/final_prediction/final_prediction_summary.tsv"/>
+        <expand macro="supp_out" task="phamer"/>
+        <expand macro="supp_out" task="phagcn"/>
+        <expand macro="supp_out" task="cherry"/>
+    </outputs>
+    <tests>
+        <test expect_num_outputs="1">
+            <param name="dbdir" value="phaboxdb"/>
+            <param name="contigs" value="example_contigs.fa"/>
+            <output name="end_to_end_out">
+                <assert_contents>
+                    <has_n_lines n="11"/>
+                    <has_n_columns n="17"/>
+                    <has_text text="Accession&#9;Length&#9;Pred&#9;Proportion&#9;PhaMerScore&#9;PhaMerConfidence&#9;Lineage&#9;PhaGCNScore&#9;Genus&#9;GenusCluster&#9;TYPE&#9;PhaTYPScore&#9;Host&#9;CHERRYScore&#9;Method&#9;Host_NCBI_lineage&#9;Host_GTDB_lineage"/>
+                </assert_contents>
+            </output>
+        </test>
+    </tests>
+    <help><![CDATA[
+
+.. class:: infomark
+
+**What it does**
+
+Runs the phabox2 pipeline, i.e. it runs
+
+- phamer: Virus identification
+- phagcn: Taxonomic classification
+- phatyp: Lifestyle prediction
+- cherry: Host prediction
+
+The outputs of the separate tools are joined into one big table.
+
+@COMMON_INPUT_DOC@
+
+**Output**
+
+@COMMON_OUTPUT_DOC@
+@PHAMER_OUTPUT_DOC@
+@PHAGCN_OUTPUT_DOC@
+@PHATYP_OUTPUT_DOC@
+@CHERRY_OUTPUT_DOC@
+    ]]></help>
+    <expand macro="citations"/>
+</tool>