Add snpeff (#546)

* feat: add snpeff

* fix: linting

* fix: tests

* fix: add csv output

* fix: add params information

* fix: improve script

* Update software/snpeff/environment.yml

* Update software/snpeff/environment.yml

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Apply suggestions from code review

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Apply suggestions from code review

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Apply suggestions from code review

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Apply suggestions from code review

* fix test

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

software/snpeff/Dockerfile

@@ -0,0 +1,22 @@
+FROM nfcore/base:1.14
+LABEL \
+    author="Maxime Garcia" \
+    description="snpEff image for nf-core pipelines" \
+    maintainer="maxime.garcia@scilifelab.se"
+
+# Install the conda environment
+COPY environment.yml /
+RUN conda env create -f /environment.yml && conda clean -a
+
+# Add conda installation dir to PATH (instead of doing 'conda activate')
+ENV PATH /opt/conda/envs/nf-core-snpeff-5.0/bin:$PATH
+
+# Setup default ARG variables
+ARG GENOME=GRCh38
+ARG SNPEFF_CACHE_VERSION=99
+
+# Download Genome
+RUN snpEff download -v ${GENOME}.${SNPEFF_CACHE_VERSION}
+
+# Dump the details of the installed packages to a file for posterity
+RUN conda env export --name nf-core-snpeff-5.0 > nf-core-snpeff-5.0.yml

software/snpeff/build.sh

@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# Build and push all containers
+
+build_push() {
+    GENOME=$1
+    SNPEFF_CACHE_VERSION=$2
+    SNPEFF_TAG=$3
+
+    docker build \
+        -t nfcore/snpeff:${SNPEFF_TAG}.${GENOME} \
+        software/snpeff/. \
+        --build-arg GENOME=${GENOME} \
+        --build-arg SNPEFF_CACHE_VERSION=${SNPEFF_CACHE_VERSION}
+
+    docker push nfcore/snpeff:${SNPEFF_TAG}.${GENOME}
+}
+
+build_push "GRCh37"    "75" "5.0"
+build_push "GRCh38"    "99" "5.0"
+build_push "GRCm38"    "99" "5.0"
+build_push "CanFam3.1" "99" "5.0"
+build_push "WBcel235"  "99" "5.0"

software/snpeff/environment.yml

@@ -0,0 +1,10 @@
+# You can use this file to create a conda environment for this module:
+#   conda env create -f environment.yml
+name: nf-core-snpeff-5.0
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+
+dependencies:
+  - bioconda::snpeff=5.0

software/snpeff/functions.nf

@@ -0,0 +1,68 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/snpeff/main.nf

@@ -0,0 +1,57 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options = initOptions(params.options)
+params.use_cache = false
+params.snpeff_tag = ""
+
+process SNPEFF {
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::snpeff=5.0" : null)
+    if (params.use_cache) {
+        if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+            container "https://depot.galaxyproject.org/singularity/snpeff:5.0--hdfd78af_1"
+        } else {
+            container "quay.io/biocontainers/snpeff:5.0--hdfd78af_1"
+        }
+    } else {
+        container "nfcore/snpeff:${params.snpeff_tag}"
+    }
+
+    input:
+    tuple val(meta), path(vcf)
+    val   db
+    path  cache
+
+    output:
+    tuple val(meta), path("*.ann.vcf"), emit: vcf
+    path "*.csv"                      , emit: report
+    path "*.version.txt"              , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def avail_mem = 6
+    if (!task.memory) {
+        log.info '[snpEff] Available memory not known - defaulting to 6GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = task.memory.giga
+    }
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    cache        = params.use_cache ? "-dataDir \${PWD}/${snpeff_cache}" : ""
+    """
+    snpEff -Xmx${avail_mem}g \\
+        $db \\
+        $options.args \\
+        -csvStats ${prefix}.csv \\
+        $cache \\
+        $vcf \\
+        > ${prefix}.ann.vcf
+
+    echo \$(snpEff -version 2>&1) > ${software}.version.txt
+    """
+}

software/snpeff/meta.yml

@@ -0,0 +1,57 @@
+name: snpEff
+description: Genetic variant annotation and functional effect prediction toolbox
+keywords:
+    - annotation
+tools:
+    - snpeff:
+        description: |
+            SnpEff is a variant annotation and effect prediction tool.
+            It annotates and predicts the effects of genetic variants on genes and proteins (such as amino acid changes).
+        homepage: https://pcingola.github.io/SnpEff/
+        documentation: https://pcingola.github.io/SnpEff/se_introduction/
+params:
+    - use_cache:
+        type: boolean
+        description: |
+          boolean to enable the usage of containers with cache
+          Enable the usage of containers with cache
+          Does not work with conda
+    - snpeff_tag:
+        type: value
+        description: |
+          Specify the tag for the container
+          https://hub.docker.com/r/nfcore/snpeff/tags
+input:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - vcf:
+        type: file
+        description: |
+            vcf to annotate
+    - db:
+        type: value
+        description: |
+            which db to annotate with
+    - cache:
+        type: file
+        description: |
+            path to snpEff cache (optional)
+output:
+    - vcf:
+        type: file
+        description: |
+            annotated vcf
+        pattern: "*.ann.vcf"
+    - report:
+        type: file
+        description: snpEff report file
+        pattern: "*.html"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@maxulysse"

tests/config/pytest_software.yml

@@ -719,6 +719,10 @@ shovill:
   - software/shovill/**
   - tests/software/shovill/**
 
+snpeff:
+  - software/snpeff/**
+  - tests/software/snpeff/**
+
 snpsites:
   - software/snpsites/**
   - tests/software/snpsites/**

tests/software/snpeff/main.nf

@@ -0,0 +1,12 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SNPEFF } from '../../../software/snpeff/main.nf' addParams( snpeff_tag: '5.0.WBcel235', use_cache: false )
+
+workflow test_snpeff {
+    input = [ [ id:'test' ], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true) ]
+            ]
+    SNPEFF ( input, "WBcel235.99", [] )
+}

tests/software/snpeff/test.yml

@@ -0,0 +1,8 @@
+- name: snpeff test_snpeff
+  command: nextflow run tests/software/snpeff -entry test_snpeff -c tests/config/nextflow.config
+  tags:
+    - snpeff
+  files:
+    - path: output/snpeff/test.ann.vcf
+      md5sum: 42db84cb186e33c6992f9306627201ac
+    - path: output/snpeff/test.csv