Add YARA module (#353)

* Initial work on yara module

* Adding in index basics

* Updated the index stuff

* Adding in proper tests

* Fix editorconfig

* Odd paths

* that should do it

* Fix tests

* Fix tests

* FFS

* Once more

* Mapping is not deterministic

software/yara/index/functions.nf

@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/yara/index/main.nf

@@ -0,0 +1,39 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process YARA_INDEX {
+    tag "$fasta"
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:'index', publish_id:'') }
+
+    conda (params.enable_conda ? "bioconda::yara=1.0.2" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/yara:1.0.2--2"
+    } else {
+        container "quay.io/biocontainers/yara:1.0.2--2"
+    }
+
+    input:
+    path fasta
+
+    output:
+    path "yara", emit: index
+    path "*.version.txt"          , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+
+    """
+    mkdir yara
+    yara_indexer $fasta -o "yara"
+    mv *.{lf,rid,sa,txt}.* yara
+    cp $fasta yara/yara.fasta
+
+    echo \$(yara_indexer --help  2>&1) | grep -e "yara_indexer version:" | sed 's/yara_indexer version: //g' > ${software}.version.txt
+    """
+}

software/yara/index/meta.yml

@@ -0,0 +1,34 @@
+name: yara_index
+description: Builds a YARA index for a reference genome
+keywords:
+  - build
+  - index
+  - fasta
+  - genome
+  - reference
+tools:
+  - yara:
+      description: Yara is an exact tool for aligning DNA sequencing reads to reference genomes.
+      homepage: https://github.com/seqan/seqan
+      documentation: https://github.com/seqan/seqan
+      tool_dev_url: https://github.com/seqan/seqan
+      doi: ""
+      licence: ['https://raw.githubusercontent.com/seqan/seqan/develop/apps/yara/LICENSE']
+
+input:
+  - fasta:
+      type: file
+      description: Input genome fasta file
+
+output:
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - index:
+      type: file
+      description: YARA genome index files
+      pattern: "yara.*"
+
+authors:
+  - "@apeltzer"

software/yara/mapper/functions.nf

@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/yara/mapper/main.nf

@@ -0,0 +1,48 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process YARA_MAPPER {
+    tag "$meta.id"
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:meta.id) }
+
+    conda (params.enable_conda ? "bioconda::yara=1.0.2 bioconda::samtools=1.12" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/mulled-v2-f13549097a0d1ca36f9d4f017636fb3609f6c083:f794a548b8692f29264c8984ff116c2141b90d9e-0"
+    } else {
+        container "quay.io/biocontainers/mulled-v2-f13549097a0d1ca36f9d4f017636fb3609f6c083:f794a548b8692f29264c8984ff116c2141b90d9e-0"
+    }
+
+    input:
+    tuple val(meta), path(reads)
+    path index
+
+    output:
+    tuple val(meta), path("*.mapped.bam"), emit: bam
+    path "*.version.txt"          , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    if(meta.single_end) {
+    """
+    yara_mapper $options.args -t ${task.cpus} -f bam ${index}/yara $reads | samtools view -@ ${task.cpus} -hb -F4 > ${prefix}.mapped.bam
+
+    echo \$(yara_mapper --help  2>&1) > ${software}.version.txt
+    """
+    } else {
+    """
+    yara_mapper $options.args -t ${task.cpus} -f bam ${index}/yara ${reads[0]} ${reads[1]} > output.bam
+    samtools view -@ ${task.cpus} -hF 4 -f 0x40 -b output.bam > ${prefix}_1.mapped.bam
+    samtools view -@ ${task.cpus} -hF 4 -f 0x80 -b output.bam > ${prefix}_2.mapped.bam
+    echo \$(yara_mapper --version  2>&1) | grep -e "yara_mapper version:" | sed 's/yara_mapper version: //g' > ${software}.version.txt
+    """
+    }
+
+}

software/yara/mapper/meta.yml

@@ -0,0 +1,47 @@
+name: yara_mapper
+description: Align reads to a reference genome using YARA
+keywords:
+  - align
+  - genome
+  - reference
+tools:
+  - yara:
+      description: Yara is an exact tool for aligning DNA sequencing reads to reference genomes.
+      homepage: https://github.com/seqan/seqan
+      documentation: https://github.com/seqan/seqan
+      tool_dev_url: https://github.com/seqan/seqan
+      doi: ""
+      licence: ['https://raw.githubusercontent.com/seqan/seqan/develop/apps/yara/LICENSE']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: |
+        List of input FastQ files of size 1 and 2 for single-end and paired-end data,
+        respectively.
+  - index:
+      type: file
+      description: YARA genome index files
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - bam:
+      type: file
+      description: Sorted BAM file
+      pattern: "*.{bam}"
+
+authors:
+  - "@apeltzer"

tests/config/pytest_software.yml

@@ -458,3 +458,11 @@ untar:
 vcftools:
   - software/vcftools/**
   - tests/software/vcftools/**
+
+yara:
+  - software/yara/mapper/**
+  - tests/software/yara/mapper/**
+
+yara_index:
+  - software/yara/index/**
+  - tests/software/yara/index/**

tests/software/yara/index/main.nf

@@ -0,0 +1,12 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { YARA_INDEX } from '../../../../software/yara/index/main.nf' addParams( options: [:] )
+
+workflow test_yara_index {
+
+    def input = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
+
+    YARA_INDEX ( input )
+}

tests/software/yara/index/test.yml

@@ -0,0 +1,30 @@
+- name: yara index test_yara_index
+  command: nextflow run tests/software/yara/index -entry test_yara_index -c tests/config/nextflow.config
+  tags:
+    - yara_index
+    - yara
+  files:
+    - path: output/index/yara/yara.rid.limits
+      md5sum: 8b814661f30a0c9e350bfbcb454930ce
+    - path: output/index/yara/yara.sa.len
+      md5sum: 45677f66c28c79c02250ceb8b58645e8
+    - path: output/index/yara/yara.sa.ind
+      md5sum: 464314583efb5f07260b0efecc29a1ce
+    - path: output/index/yara/yara.lf.drp
+      md5sum: 3ef99a87a4e44513f46d42f4261f7842
+    - path: output/index/yara/yara.txt.size
+      md5sum: 063987b3c3f747be7d2b8043c9d91000
+    - path: output/index/yara/yara.rid.concat
+      md5sum: 1e4e4c88ddeaf907a12f02f0d88367c5
+    - path: output/index/yara/yara.txt.concat
+      md5sum: 6074d1933c9e7e5ab05fa0def5ce28c0
+    - path: output/index/yara/yara.lf.drs
+      md5sum: 55a54008ad1ba589aa210d2629c1df41
+    - path: output/index/yara/yara.txt.limits
+      md5sum: 4480a068db603e4c9a27bc4fa9ceaf14
+    - path: output/index/yara/yara.sa.val
+      md5sum: ce57cc82e2d3ae7b9824210f54168ce9
+    - path: output/index/yara/yara.lf.pst
+      md5sum: e8daba34298e99e42942435286f9b3f0
+    - path: output/index/yara/yara.lf.drv
+      md5sum: cf6408307fe9fd7f99c33f521bf95550

tests/software/yara/mapper/main.nf

@@ -0,0 +1,32 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+
+include { YARA_INDEX } from '../../../../software/yara/index/main.nf' addParams( options: ['args': '-e 3'] )
+include { YARA_MAPPER } from '../../../../software/yara/mapper/main.nf' addParams( options: ['args': '-e 3'] )
+
+workflow test_yara_single_end {
+
+    def fasta = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
+    YARA_INDEX ( fasta )
+
+    def input = []
+    input = [ [ id:'test', single_end:true ], // meta map
+              file("${launchDir}/tests/data/genomics/sarscov2/illumina/fastq/test_1.fastq.gz", checkIfExists: true) ]
+
+    YARA_MAPPER ( input, YARA_INDEX.out.index )
+}
+
+workflow test_yara_paired_end {
+
+    def fasta = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
+    YARA_INDEX ( fasta )
+
+    def input = []
+    input = [ [ id:'test', single_end:false ], // meta map
+              [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/fastq/test_1.fastq.gz", checkIfExists: true),
+                file("${launchDir}/tests/data/genomics/sarscov2/illumina/fastq/test_2.fastq.gz", checkIfExists: true) ] ]
+
+    YARA_MAPPER ( input, YARA_INDEX.out.index )
+}

tests/software/yara/mapper/test.yml

@@ -0,0 +1,70 @@
+- name: yara mapper test_yara_single_end
+  command: nextflow run tests/software/yara/mapper -entry test_yara_single_end -c tests/config/nextflow.config
+  tags:
+    - yara_single_end
+    - yara_mapper
+    - yara
+  files:
+    - path: output/yara/test.mapped.bam
+    - path: output/index/yara/yara.txt.size
+      md5sum: 063987b3c3f747be7d2b8043c9d91000
+    - path: output/index/yara/yara.lf.drs
+      md5sum: 55a54008ad1ba589aa210d2629c1df41
+    - path: output/index/yara/yara.lf.pst
+      md5sum: e8daba34298e99e42942435286f9b3f0
+    - path: output/index/yara/yara.sa.len
+      md5sum: 45677f66c28c79c02250ceb8b58645e8
+    - path: output/index/yara/yara.rid.concat
+      md5sum: 1e4e4c88ddeaf907a12f02f0d88367c5
+    - path: output/index/yara/yara.txt.concat
+      md5sum: 6074d1933c9e7e5ab05fa0def5ce28c0
+    - path: output/index/yara/yara.sa.val
+      md5sum: ce57cc82e2d3ae7b9824210f54168ce9
+    - path: output/index/yara/yara.sa.ind
+      md5sum: 464314583efb5f07260b0efecc29a1ce
+    - path: output/index/yara/yara.rid.limits
+      md5sum: 8b814661f30a0c9e350bfbcb454930ce
+    - path: output/index/yara/yara.lf.drp
+      md5sum: 3ef99a87a4e44513f46d42f4261f7842
+    - path: output/index/yara/yara.txt.limits
+      md5sum: 4480a068db603e4c9a27bc4fa9ceaf14
+    - path: output/index/yara/yara.lf.drv
+      md5sum: cf6408307fe9fd7f99c33f521bf95550
+    - path: output/index/yara/yara.fasta
+      md5sum: 6e9fe4042a72f2345f644f239272b7e6
+
+- name: yara mapper test_yara_paired_end
+  command: nextflow run tests/software/yara/mapper -entry test_yara_paired_end -c tests/config/nextflow.config
+  tags:
+    - yara_mapper
+    - yara_paired_end
+    - yara
+  files:
+    - path: output/yara/test_2.mapped.bam
+    - path: output/yara/test_1.mapped.bam
+    - path: output/index/yara/yara.txt.size
+      md5sum: 063987b3c3f747be7d2b8043c9d91000
+    - path: output/index/yara/yara.lf.drs
+      md5sum: 55a54008ad1ba589aa210d2629c1df41
+    - path: output/index/yara/yara.lf.pst
+      md5sum: e8daba34298e99e42942435286f9b3f0
+    - path: output/index/yara/yara.sa.len
+      md5sum: 45677f66c28c79c02250ceb8b58645e8
+    - path: output/index/yara/yara.rid.concat
+      md5sum: 1e4e4c88ddeaf907a12f02f0d88367c5
+    - path: output/index/yara/yara.txt.concat
+      md5sum: 6074d1933c9e7e5ab05fa0def5ce28c0
+    - path: output/index/yara/yara.sa.val
+      md5sum: ce57cc82e2d3ae7b9824210f54168ce9
+    - path: output/index/yara/yara.sa.ind
+      md5sum: 464314583efb5f07260b0efecc29a1ce
+    - path: output/index/yara/yara.rid.limits
+      md5sum: 8b814661f30a0c9e350bfbcb454930ce
+    - path: output/index/yara/yara.lf.drp
+      md5sum: 3ef99a87a4e44513f46d42f4261f7842
+    - path: output/index/yara/yara.txt.limits
+      md5sum: 4480a068db603e4c9a27bc4fa9ceaf14
+    - path: output/index/yara/yara.lf.drv
+      md5sum: cf6408307fe9fd7f99c33f521bf95550
+    - path: output/index/yara/yara.fasta
+      md5sum: 6e9fe4042a72f2345f644f239272b7e6