From 7d0ddbc8abf45760e92276e6bb377c7b535baa95 Mon Sep 17 00:00:00 2001
From: "James A. Fellows Yates" <jfy133@gmail.com>
Date: Fri, 1 Jul 2022 15:06:07 +0200
Subject: [PATCH] Add atlas/call (#1809)

* Add atlas/call

* Apply suggestions from code review

* Update modules/atlas/call/main.nf

Co-authored-by: Thiseas C. Lamnidis <thisseass@gmail.com>

* Apply suggestions from code review

Co-authored-by: Thiseas C. Lamnidis <thisseass@gmail.com>
---
 modules/atlas/call/main.nf               | 52 +++++++++++++++++
 modules/atlas/call/meta.yml              | 72 ++++++++++++++++++++++++
 tests/config/pytest_modules.yml          |  4 ++
 tests/modules/atlas/call/main.nf         | 22 ++++++++
 tests/modules/atlas/call/nextflow.config |  5 ++
 tests/modules/atlas/call/test.yml        |  9 +++
 6 files changed, 164 insertions(+)
 create mode 100644 modules/atlas/call/main.nf
 create mode 100644 modules/atlas/call/meta.yml
 create mode 100644 tests/modules/atlas/call/main.nf
 create mode 100644 tests/modules/atlas/call/nextflow.config
 create mode 100644 tests/modules/atlas/call/test.yml

diff --git a/modules/atlas/call/main.nf b/modules/atlas/call/main.nf
new file mode 100644
index 00000000..b1354505
--- /dev/null
+++ b/modules/atlas/call/main.nf
@@ -0,0 +1,52 @@
+process ATLAS_CALL {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::atlas=0.9.9" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/atlas:0.9.9--h082e891_0':
+        'quay.io/biocontainers/atlas:0.9.9--h082e891_0' }"
+
+    input:
+    tuple val(meta), path(bam), path(bai)
+    path fasta
+    path fai
+    path recal
+    path pmd
+    path known_alleles
+    val method
+
+    output:
+    tuple val(meta), path("*.vcf.gz"), emit: vcf
+    path "versions.yml"              , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args               = task.ext.args ?: ''
+    def prefix             = task.ext.prefix ?: "${meta.id}"
+    def recal_file         = recal ? "recal=${recal}" : ""
+    def pmd_file           = pmd ? "pmdFile=${pmd}" : ""
+    def known_alleles_file = known_alleles ? "pmdFile=${known_alleles}" : ""
+
+    def valid_method = ['MLE', 'Bayesian', 'allelePresence', 'randomBase', 'majorityBase']
+    if ( !valid_method.contains(method) )  { error "Unrecognised calling method for ATLAS_CALL. Options: MLE, Bayesian, allelePresence, randomBase, majorityBase" }
+
+    """
+    atlas \\
+        task=call \\
+        bam=${bam} \\
+        fasta=${fasta} \\
+        $recal_file \\
+        $pmd_file \\
+        method=${method} \\
+        $args
+
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        atlas: \$((atlas 2>&1) | grep Atlas | head -n 1 | sed -e 's/^[ \t]*Atlas //')
+    END_VERSIONS
+    """
+}
diff --git a/modules/atlas/call/meta.yml b/modules/atlas/call/meta.yml
new file mode 100644
index 00000000..fd6e7436
--- /dev/null
+++ b/modules/atlas/call/meta.yml
@@ -0,0 +1,72 @@
+name: "atlas_call"
+description: generate VCF file from a BAM file using various calling methods
+keywords:
+  - atlas
+  - variant calling
+  - vcf
+  - population genetics
+tools:
+  - "atlas":
+      description: "ATLAS, a suite of methods to accurately genotype and estimate genetic diversity"
+      homepage: "https://bitbucket.org/wegmannlab/atlas/wiki/Home"
+      documentation: "https://bitbucket.org/wegmannlab/atlas/wiki/Home"
+      tool_dev_url: "https://bitbucket.org/wegmannlab/atlas"
+      doi: "10.1101/105346"
+      licence: "['GPL v3']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bam:
+      type: file
+      description: A BAM/ file
+      pattern: "*.bam"
+  - bai:
+      type: file
+      description: The BAI file for the input BAM file
+      pattern: "*.bai"
+  - fasta:
+      type: file
+      description: The reference FASTA file used to generate the BAM file
+      pattern: "*.{fasta,fna,fa}"
+  - fai:
+      type: file
+      description: The index of the FASTA file used for to generate the BAM file
+      pattern: "*.fai"
+  - recal:
+      type: file
+      description: Optional recalibration file from atlas recal function in text format
+      pattern: "*.txt"
+  - pmd:
+      type: file
+      description: Optional PMD file from atlas pmd function in text format
+      pattern: "*.txt"
+  - known_alleles:
+      type: file
+      description: Optional tab separated text file containing 1-based list of known alleles. See atlas call documentation.
+      pattern: "*.{txt.tsv}"
+  - method:
+      type: character
+      description: Which variant calling algorithm to use. Some may require additional parameters supplied via ext.args. Check atlas documentation.
+      pattern: "MLE|Bayesian|allelePresence|randomBase|majorityBase"
+
+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"
+  - bam:
+      type: file
+      description: VCF file with variant calls
+      pattern: "*.vcf.gz"
+
+authors:
+  - "@jfy133"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index f7e49c83..c4d71736 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -86,6 +86,10 @@ ataqv/ataqv:
   - modules/ataqv/ataqv/**
   - tests/modules/ataqv/ataqv/**
 
+atlas/call:
+  - modules/atlas/call/**
+  - tests/modules/atlas/call/**
+
 atlas/pmd:
   - modules/atlas/pmd/**
   - tests/modules/atlas/pmd/**
diff --git a/tests/modules/atlas/call/main.nf b/tests/modules/atlas/call/main.nf
new file mode 100644
index 00000000..945812ef
--- /dev/null
+++ b/tests/modules/atlas/call/main.nf
@@ -0,0 +1,22 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { ATLAS_CALL } from '../../../../modules/atlas/call/main.nf'
+
+workflow test_atlas_call {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true)
+    ]
+    fasta         = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    fai           = file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true)
+    recal         = []
+    pmd           = []
+    known_alleles = []
+    method        = 'randomBase'
+
+    ATLAS_CALL ( input, fasta, fai, recal, pmd, known_alleles, method )
+}
diff --git a/tests/modules/atlas/call/nextflow.config b/tests/modules/atlas/call/nextflow.config
new file mode 100644
index 00000000..50f50a7a
--- /dev/null
+++ b/tests/modules/atlas/call/nextflow.config
@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+}
\ No newline at end of file
diff --git a/tests/modules/atlas/call/test.yml b/tests/modules/atlas/call/test.yml
new file mode 100644
index 00000000..b3e13951
--- /dev/null
+++ b/tests/modules/atlas/call/test.yml
@@ -0,0 +1,9 @@
+- name: atlas call test_atlas_call
+  command: nextflow run ./tests/modules/atlas/call -entry test_atlas_call -c ./tests/config/nextflow.config  -c ./tests/modules/atlas/call/nextflow.config
+  tags:
+    - atlas
+    - atlas/call
+  files:
+    - path: output/atlas/test.paired_end.sorted_randomBase.vcf.gz
+      contains:
+        - "##source=atlas"