Merge branch 'develop' into enhancement/produce_metafusion_ref

.editorconfig

@@ -12,7 +12,7 @@ indent_style = space
 indent_size = 2
 
 # These files are edited and tested upstream in nf-core/modules
-[/modules/nf-core/**]
+[/modules/{nf-core,msk}/**]
 charset = unset
 end_of_line = unset
 insert_final_newline = unset

conf/modules.config

@@ -492,8 +492,8 @@ process {
         ]
     }
 
-    withName: GETBASECOUNTSMULTISAMPLE {
-        ext.args = "--omaf --maq 20  --baq 20 --filter_improper_pair 0"
+    withName: GBCMS {
+        ext.args = { "--omaf --maq 20  --baq 20 --filter_improper_pair 0 --thread ${task.cpus}" }
         publishDir = [
             enabled:false
         ]

modules.json

@@ -2,6 +2,17 @@
     "name": "mskcc/forte",
     "homePage": "https://github.com/mskcc/forte",
     "repos": {
+        "[email protected]:mskcc-omics-workflows/modules.git": {
+            "modules": {
+                "msk": {
+                    "gbcms": {
+                        "branch": "develop",
+                        "git_sha": "e70681357ccd39e7a10d328ca696c074750d8163",
+                        "installed_by": ["modules"]
+                    }
+                }
+            }
+        },
         "https://github.com/nf-core/modules.git": {
             "modules": {
                 "nf-core": {

modules/msk/gbcms/environment.yml

@@ -0,0 +1,9 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+name: "gbcms"
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - "YOUR-TOOL-HERE"

modules/msk/gbcms/main.nf

@@ -0,0 +1,63 @@
+process GBCMS {
+    tag "$meta.id"
+    label 'process_single'
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'ghcr.io/msk-access/gbcms:1.2.5':
+        'ghcr.io/msk-access/gbcms:1.2.5' }"
+
+    input:
+    tuple val(meta), path(bam), path(bambai), path(variant_file), val(output)
+    path(fasta)
+    path(fastafai)
+
+    output:
+    tuple val(meta), path('*.{vcf,maf}') , emit: variant_file
+    tuple val(meta), path("versions.yml")   , emit: versions
+
+    when:
+        task.ext.when == null || task.ext.when
+    script:
+    if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+        error "GetBaseCountsMultiSample module does not support Conda. Please use Docker / Singularity instead."
+    }
+    def args = task.ext.args ?: ''
+    def sample = meta.sample
+    // determine if input file is a maf of vcf
+
+    def input_ext = variant_file.getExtension()
+    def variant_input = ''
+
+    if(input_ext == 'maf') {
+        variant_input = '--maf ' + variant_file
+    }
+    if(input_ext == 'vcf'){
+            variant_input = '--vcf ' + variant_file
+    }
+    if(variant_input == ''){
+        throw new Exception("Variant file must be maf or vcf, not ${input_ext}")
+    }
+
+    """
+    GetBaseCountsMultiSample --fasta ${fasta} \\
+    ${variant_input} \\
+    --output ${output} \\
+    --bam $sample:${bam} $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        GetBaseCountsMultiSample: \$(echo \$(GetBaseCountsMultiSample --help) | grep -oP '[0-9]\\.[0-9]\\.[0-9]')
+    END_VERSIONS
+    """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+
+    touch variant_file.maf
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        GetBaseCountsMultiSample:  1.2.5
+    END_VERSIONS
+    """
+}

modules/msk/gbcms/meta.yml

@@ -0,0 +1,59 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/yaml-schema.json
+name: "gbcms"
+description: This module wraps GetBaseCountsMultiSample, which calculates the base counts in multiple BAM files for all the sites in a given VCF file or MAF file
+keywords:
+  - basecount
+  - bams
+  - vcf
+tools:
+  - "gbcms":
+      description: "Calculate the base counts in multiple BAM files for all the sites in a given VCF file or MAF file"
+      homepage: "https://github.com/msk-access/GetBaseCountsMultiSample"
+      documentation: "https://github.com/msk-access/GetBaseCountsMultiSample/blob/master/README.md"
+
+input:
+  # Only when we have meta
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+
+  - fasta:
+      type: file
+      description: Input reference sequence file
+      pattern: "*.fasta"
+
+  - fastafai:
+      type: file
+      description: Index of the reference Fasta
+      pattern: "*.fai"
+  - bam:
+      type: file
+      description: Input bam file, in the format of SAMPLE_NAME:BAM_FILE. This paramter need to be specified at least once
+      pattern: "*.bam"
+  - bambai:
+      type: file
+      description: Index of Bam
+      pattern: "*.bai"
+  - variant_file:
+      type: file
+      description: Input variant file in TCGA maf format. --maf or --vcf need to be specified at least once. But --maf and --vcf are mutually exclusive
+      pattern: "*.{maf,vcf}"
+  - output:
+      type: string
+      description: Output file
+
+output:
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - variant_file:
+      type: file
+      description: base counts in multiple BAM files for all the sites in a given VCF file or MAF file
+      pattern: "*.{vcf,maf}"
+
+authors:
+  - "@buehlere"

modules/msk/gbcms/tests/main.nf.test

@@ -0,0 +1,41 @@
+// nf-core modules test gbcms
+nextflow_process {
+
+    name "Test Process GBCMS"
+    script "../main.nf"
+    process "GBCMS"
+    config "./nextflow.config"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "modules_msk"
+    tag "gbcms"
+
+    test("sarscov2 illumina  paired-end [fastq]") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                            [ id:'test', sample:'197' ], // meta map
+                            file(params.test_data['sarscov2']['illumina']['test_single_end_sorted_bam'], checkIfExists: true),
+                            file(params.test_data['sarscov2']['illumina']['test_single_end_sorted_bam_bai'], checkIfExists: true),
+                            file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true),
+                            "variant_file.vcf" 
+                        ]
+                input[1] = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+                input[2] = file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true)
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+}

modules/msk/gbcms/tests/main.nf.test.snap

@@ -0,0 +1,45 @@
+{
+    "sarscov2 illumina  paired-end [fastq]": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test",
+                            "sample": "197"
+                        },
+                        "variant_file.vcf:md5,28c8df33c7ea5ed5d1cf9997d8e00ffa"
+                    ]
+                ],
+                "1": [
+                    [
+                        {
+                            "id": "test",
+                            "sample": "197"
+                        },
+                        "versions.yml:md5,a94265ed3bc4b5631d85b9b9b5d2b7e5"
+                    ]
+                ],
+                "variant_file": [
+                    [
+                        {
+                            "id": "test",
+                            "sample": "197"
+                        },
+                        "variant_file.vcf:md5,28c8df33c7ea5ed5d1cf9997d8e00ffa"
+                    ]
+                ],
+                "versions": [
+                    [
+                        {
+                            "id": "test",
+                            "sample": "197"
+                        },
+                        "versions.yml:md5,a94265ed3bc4b5631d85b9b9b5d2b7e5"
+                    ]
+                ]
+            }
+        ],
+        "timestamp": "2023-12-13T15:05:34.981489"
+    }
+}

modules/msk/gbcms/tests/nextflow.config

@@ -0,0 +1,12 @@
+params {
+    enable_conda = false
+}  
+
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    withName: "GBCMS" {
+        ext.args = ''
+    }
+
+}

modules/msk/gbcms/tests/tags.yml

@@ -0,0 +1,2 @@
+gbcms:
+  - "modules/msk/gbcms/**"