Instrain profile (#1822)

* add instrain profile

* module instrain/profile

* add instrain profile

* module instrain/profile

* instrain profile

* Update modules/instrain/profile/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/instrain/profile/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/instrain/profile/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/instrain/profile/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/instrain/profile/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/instrain/profile/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/instrain/profile/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* update

* linting

* Apply suggestions from code review

* Update modules/instrain/profile/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

modules/instrain/profile/main.nf

@@ -0,0 +1,44 @@
+process INSTRAIN_PROFILE {
+    tag "$meta.id"
+    label 'process_high'
+
+    conda (params.enable_conda ? "bioconda::instrain=1.6.1" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/instrain:1.6.1--pyhdfd78af_0':
+        'quay.io/biocontainers/instrain:1.6.1--pyhdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(bam)
+    path genome_fasta
+    path genes_fasta
+    path stb_file
+
+    output:
+    tuple val(meta), path("*.IS") , emit: profile
+    path "versions.yml"           , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def genes_args = genes_fasta ? "-g ${genes_fasta}": ''
+    def stb_args = stb_file ? "-s ${stb_file}": ''
+    """
+    inStrain \\
+        profile \\
+        $bam \\
+        $genome_fasta \\
+        -o ${prefix}.IS \\
+        -p $task.cpus \\
+        $genes_args \\
+        $stb_args \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        instrain: \$(echo \$(inStrain profile --version 2>&1) | awk 'NF{ print \$NF }')
+    END_VERSIONS
+    """
+}

modules/instrain/profile/meta.yml

@@ -0,0 +1,54 @@
+name: "instrain_profile"
+description: inStrain is python program for analysis of co-occurring genome populations from metagenomes that allows highly accurate genome comparisons, analysis of coverage, microdiversity, and linkage, and sensitive SNP detection with gene localization and synonymous non-synonymous identification
+keywords:
+  - instrain
+  - metagenomics
+  - population genomics
+  - profile
+tools:
+  - instrain:
+      description: Calculation of strain-level metrics
+      homepage: https://github.com/MrOlm/instrain
+      documentation: https://instrain.readthedocs.io/en/latest/
+      tool_dev_url: https://github.com/MrOlm/instrain
+      doi: 10.1038/s41587-020-00797-0
+      licence: ["MIT"]
+
+input:
+  - meta:
+      type: map
+      description: Groovy Map containing sample information e.g. [ id:'test']
+  - bam:
+      type: path
+      description: Path to .bam file to be profiled
+      pattern: "*.{bam,sam}"
+  - genome_fasta:
+      type: path
+      description: Path to .fasta file to be profiled; MUST be the .fasta file that was mapped to to create the .bam file
+      pattern: "*.{fasta,fna,fa}"
+  - genes_fasta:
+      type: path
+      description: Path to .fna file of genes to be profiled (OPTIONAL)
+      pattern: "*.{fasta,fna,fa}"
+  - stb_file:
+      type: path
+      description: Path to .stb (scaffold to bin) file to be profiled (OPTIONAL)
+      pattern: "*.stb"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - profile:
+      type: path
+      description: InStrain profile folder
+      pattern: "*.IS/"
+
+authors:
+  - "@mrolm"

tests/config/pytest_modules.yml

@@ -1113,6 +1113,10 @@ imputeme/vcftoprs:
   - modules/imputeme/vcftoprs/**
   - tests/modules/imputeme/vcftoprs/**
 
+instrain/profile:
+  - modules/instrain/profile/**
+  - tests/modules/instrain/profile/**
+
 iqtree:
   - modules/iqtree/**
   - tests/modules/iqtree/**

tests/modules/instrain/profile/main.nf

@@ -0,0 +1,18 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { INSTRAIN_PROFILE } from '../../../../modules/instrain/profile/main.nf'
+
+workflow test_instrain_profile {
+    
+    input = [
+        [ id:'test', single_end:true ], // meta map
+        [
+            file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true)
+        ]
+    ]
+    genome_fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+
+    INSTRAIN_PROFILE ( input , genome_fasta , [] , [] )
+}

tests/modules/instrain/profile/nextflow.config

@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+}

tests/modules/instrain/profile/test.yml

@@ -0,0 +1,9 @@
+- name: "instrain"
+  command: nextflow run ./tests/modules/instrain/profile -entry test_instrain_profile -c ./tests/config/nextflow.config -c ./tests/modules/instrain/profile/nextflow.config
+  tags:
+    - instrain
+    - instrain/profile
+  files:
+    - path: output/instrain/versions.yml
+    - path: output/instrain/test.IS/output/test.IS_SNVs.tsv
+      md5sum: 1d753903af5a25be540dcff255e25a1f