NEW MODULE: PURECLIP

modules/nf-core/pureclip/main.nf

@@ -0,0 +1,69 @@
+process PURECLIP {
+    tag "$meta.id"
+    label 'process_high'
+
+    conda "bioconda::pureclip=1.3.1"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/pureclip:1.3.1--0':
+        'biocontainers/pureclip:1.3.1--0' }"
+
+    input:
+    tuple val(meta), path(ipbam), path(controlbam)
+    tuple val(meta), path(ipbai), path(controlbai)
+    path genome_fasta
+    val input_control
+
+    output:
+    tuple val(meta), path("${crosslinks_output_name}"), emit: crosslinks
+    tuple val(meta), path("${peaks_output_name}")     , emit: peaks
+    path "versions.yml"                               , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    crosslinks_output_name = "${prefix}_pureclip_crosslinks.bed"
+    peaks_output_name      = "${prefix}_pureclip_peaks.bed"
+
+    if(input_control){
+        control_bam   = "-ibam $controlbam"
+        control_bai   = "-ibai $controlbai"
+    } else {
+        control_bam   = ""
+        control_bai   = ""
+    }
+
+    """
+    pureclip \
+        -i $ipbam \
+        -bai $ipbai \
+        -g $genome_fasta \
+        -nt ${task.cpus} \
+        -o $crosslinks_output_name \
+        -or $peaks_output_name \
+        ${control_bam} \
+        ${control_bai} \
+        ${args}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        pureclip: \$(echo \$(pureclip --version 2>&1) | sed 's/^.*pureclip //; s/Using.*\$//; s/version: //; s/ Seq.*//' ))
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    touch ${prefix}_pureclip_crosslinks.bed
+    touch ${prefix}_pureclip_peaks.bed
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        pureclip: \$(echo \$(pureclip --version 2>&1) | sed 's/^.*pureclip //; s/Using.*\$//; s/version: //; s/ Seq.*//' ))
+    END_VERSIONS
+    """
+}

modules/nf-core/pureclip/meta.yml

@@ -0,0 +1,77 @@
+---
+name: "pureclip"
+description: PureCLIP is a tool to detect protein-RNA interaction footprints from single-nucleotide CLIP-seq data, such as iCLIP and eCLIP.
+keywords:
+  - iCLIP
+  - eCLIP
+  - CLIP
+tools:
+  - "pureclip":
+      description: "PureCLIP is a tool to detect protein-RNA interaction footprints from single-nucleotide CLIP-seq data, such as iCLIP and eCLIP."
+      homepage: "https://github.com/skrakau/PureCLIP"
+      documentation: "https://pureclip.readthedocs.io/en/latest/GettingStarted/index.html"
+      tool_dev_url: "https://github.com/skrakau/PureCLIP"
+      doi: "10.1186/s13059-017-1364-2"
+      licence: "['GPL v3']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+
+  - ipbam:
+      type: file
+      description: Sorted BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+  - controlbam:
+      type: file
+      description: Sorted BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+  - ipbai:
+      type: file
+      description: BAM index
+      pattern: "*.{bai}"
+
+  - controlbai:
+      type: file
+      description: BAM index
+      pattern: "*.{bai}"
+
+  - input_control:
+      type: boolean
+      description: Whether to run PureCLIP with an input control
+
+  - genome_fasta:
+      type: file
+      description: FASTA file of reference genome
+      pattern: "*.{fa,fasta,fa.gz,fasta.gz}"
+
+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"
+
+  - crosslinks:
+      type: file
+      description: Bed file of crosslinks
+      pattern: "*.{bed}"
+
+  - peaks:
+      type: file
+      description: Bed file of peaks
+      pattern: "*.{bed}"
+
+authors:
+  - "@charlotteanne"
+  - "@marcjones"

tests/config/pytest_modules.yml

@@ -2879,6 +2879,10 @@ proteus/readproteingroups:
   - modules/nf-core/proteus/readproteingroups/**
   - tests/modules/nf-core/proteus/readproteingroups/**
 
+pureclip:
+  - modules/nf-core/pureclip/**
+  - tests/modules/nf-core/pureclip/**
+
 purecn/coverage:
   - modules/nf-core/purecn/coverage/**
   - tests/modules/nf-core/purecn/coverage/**

tests/modules/nf-core/pureclip/main.nf

@@ -0,0 +1,43 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { PURECLIP } from '../../../../modules/nf-core/pureclip/main.nf'
+
+workflow test_pureclip_no_ctrl {
+    input_bam = [ [ id:'test', single_end:false ], // meta map
+              [ file( params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true) ],
+              [ file( params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam'], checkIfExists: true)  ]] //dummy file
+
+    input_bai = [ [ id:'test', single_end:false ], // meta map
+              [ file( params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true) ],
+              [ file( params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam_bai'], checkIfExists: true) ]] //dummy file
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+
+    PURECLIP (
+        input_bam,
+        input_bai,
+        fasta,
+        false
+        )
+}
+
+workflow test_pureclip_input_ctrl {
+    input_bam = [ [ id:'test', single_end:false ], // meta map
+              [ file( params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true) ],
+              [ file( params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam'], checkIfExists: true)  ]]
+
+    input_bai = [ [ id:'test', single_end:false ], // meta map
+              [ file( params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true) ],
+              [ file( params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam_bai'], checkIfExists: true) ]]
+
+    genome = [file( params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)]
+
+    PURECLIP (
+        input_bam,
+        input_bai,
+        genome,
+        true
+        )
+}

tests/modules/nf-core/pureclip/nextflow.config

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

tests/modules/nf-core/pureclip/test.yml

@@ -0,0 +1,21 @@
+- name: pureclip test_pureclip_no_ctrl
+  command: nextflow run ./tests/modules/nf-core/pureclip -entry test_pureclip_no_ctrl -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/pureclip/nextflow.config
+  tags:
+    - pureclip
+  files:
+    - path: output/pureclip/test_pureclip_crosslinks.bed
+      md5sum: c13c98e4d205025a3a824fbf36aaeef5
+    - path: output/pureclip/test_pureclip_peaks.bed
+      md5sum: 5fa8033fbe5901ac68a5b11e47aa22c6
+    - path: output/pureclip/versions.yml
+
+- name: pureclip test_pureclip_input_ctrl
+  command: nextflow run ./tests/modules/nf-core/pureclip -entry test_pureclip_input_ctrl -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/pureclip/nextflow.config
+  tags:
+    - pureclip
+  files:
+    - path: output/pureclip/test_pureclip_crosslinks.bed
+      md5sum: 511696bc8fab569c562f1cbc8ff01c45
+    - path: output/pureclip/test_pureclip_peaks.bed
+      md5sum: 927987b0ab82b08255caa4b5d143729b
+    - path: output/pureclip/versions.yml