Add: ANGSD_DOCOUNTS (#2044)

* Update main.nf

* Update meta.yml

* Re-add logos as not staged in a way that works with MultiQC config files

* Add ANGSD doCounts

* Prettier

* Update main.nf

* Apply suggestions from code review

* Apply suggestions from code review

* Apply suggestions from code review

* Try loosening test due to conda

* Remove md5 for binary file

Co-authored-by: Simon Pearce <24893913+SPPearce@users.noreply.github.com>

modules/angsd/docounts/main.nf

@@ -0,0 +1,45 @@
+process ANGSD_DOCOUNTS {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::angsd=0.939" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/angsd:0.939--h468462d_0':
+        'quay.io/biocontainers/angsd:0.939--h468462d_0' }"
+
+    input:
+    tuple val(meta), path(bam), path(bai), path(minqfile)
+
+    output:
+    tuple val(meta), path("*.depthSample"), optional: true, emit: depth_sample
+    tuple val(meta), path("*.depthGlobal"), optional: true, emit: depth_global
+    tuple val(meta), path("*.qs")         , optional: true, emit: qs
+    tuple val(meta), path("*.pos.gz")     , optional: true, emit: pos
+    tuple val(meta), path("*.counts.gz")  , optional: true, emit: counts
+    tuple val(meta), path("*.icnts.gz")   , optional: true, emit: icounts
+    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 minq = minqfile ? "-minQfile ${minqfile}" : ""
+    """
+    readlink -f *.bam > bamlist.txt
+
+    angsd \\
+        -nThreads ${task.cpus} \\
+        -doCounts 1 \\
+        $args \\
+        -bam bamlist.txt \\
+        -out ${prefix} \\
+        $minq
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        angsd: \$(echo \$(angsd 2>&1) | grep version | head -n 1 | sed 's/.*version: //g;s/ .*//g')
+    END_VERSIONS
+    """
+}

modules/angsd/docounts/meta.yml

@@ -0,0 +1,72 @@
+name: "angsd_docounts"
+description: Calculates base frequency statistics across reference positions from BAM.
+keywords:
+  - angsd
+  - population genetics
+  - allele counts
+  - doCounts
+tools:
+  - "angsd":
+      description: "ANGSD: Analysis of next generation Sequencing Data"
+      homepage: "http://www.popgen.dk/angsd/"
+      documentation: "http://www.popgen.dk/angsd/"
+      tool_dev_url: "https://github.com/ANGSD/angsd"
+      doi: "10.1186/s12859-014-0356-4"
+      licence: "['GPL v3, MIT']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bam:
+      type: file
+      description: A list of BAM or CRAM files
+      pattern: "*.{bam,cram}"
+  - bai:
+      type: file
+      description: List of BAM/CRAM index files
+      pattern: "*.{bai,csi}"
+  - minqfile:
+      type: file
+      description: File with individual quality score thresholds
+      pattern: "*"
+
+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"
+  - depth_sample:
+      type: file
+      description: Distribution of sequencing depths
+      pattern: "*.depthSample"
+  - depth_global:
+      type: file
+      description: Distribution of sequencing depths
+      pattern: "*.depthGlobal"
+  - qs:
+      type: file
+      description: Distribution of scores
+      pattern: "*.qs"
+  - pos:
+      type: file
+      description: Various types of depth statistics (depending on value for -dumpCounts)
+      pattern: "*.pos.gz"
+  - counts:
+      type: file
+      description: Various types of statistics (related to pos.gz)
+      pattern: "*.counts.gz"
+  - icounts:
+      type: file
+      description: Internal format for dumping binary single chrs. Useful for ANGSD contamination
+      pattern: "*.icnts.gz"
+
+authors:
+  - "@jfy133"

tests/config/pytest_modules.yml

@@ -46,6 +46,10 @@ amrfinderplus/update:
   - modules/amrfinderplus/update/**
   - tests/modules/amrfinderplus/update/**
 
+angsd/docounts:
+  - modules/angsd/docounts/**
+  - tests/modules/angsd/docounts/**
+
 antismash/antismashlite:
   - modules/antismash/antismashlite/**
   - tests/modules/antismash/antismashlite/**

tests/modules/angsd/docounts/main.nf

@@ -0,0 +1,17 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { ANGSD_DOCOUNTS } from '../../../../modules/angsd/docounts/main.nf'
+
+workflow test_angsd_docounts {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_markduplicates_sorted_bam'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_markduplicates_sorted_bam_bai'], checkIfExists: true),
+        []
+    ]
+
+    ANGSD_DOCOUNTS ( input )
+}

tests/modules/angsd/docounts/nextflow.config

@@ -0,0 +1,9 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: ANGSD_DOCOUNTS {
+        ext.args = "-iCounts 1 -doDepth -dumpCounts 2"
+    }
+
+}

tests/modules/angsd/docounts/test.yml

@@ -0,0 +1,11 @@
+- name: angsd docounts test_angsd_docounts
+  command: nextflow run ./tests/modules/angsd/docounts -entry test_angsd_docounts -c ./tests/config/nextflow.config  -c ./tests/modules/angsd/docounts/nextflow.config
+  tags:
+    - angsd/docounts
+    - angsd
+  files:
+    - path: output/angsd/test.counts.gz
+      contains: ["ind0TotDepth"]
+    - path: output/angsd/test.icnts.gz
+    - path: output/angsd/test.pos.gz
+      contains: ["chr	pos	totDepth"]