Merge branch 'master' into bump/picard

modules/ensemblvep/Dockerfile

@@ -11,8 +11,8 @@ RUN conda env create -f /environment.yml && conda clean -a
 # Setup default ARG variables
 ARG GENOME=GRCh38
 ARG SPECIES=homo_sapiens
-ARG VEP_VERSION=104
-ARG VEP_TAG=104.3
+ARG VEP_VERSION=105
+ARG VEP_TAG=105.0
 
 # Add conda installation dir to PATH (instead of doing 'conda activate')
 ENV PATH /opt/conda/envs/nf-core-vep-${VEP_TAG}/bin:$PATH

modules/ensemblvep/build.sh

@@ -20,9 +20,9 @@ build_push() {
     docker push nfcore/vep:${VEP_TAG}.${GENOME}
 }
 
-build_push "GRCh37"    "homo_sapiens"           "104" "104.3"
-build_push "GRCh38"    "homo_sapiens"           "104" "104.3"
-build_push "GRCm38"    "mus_musculus"           "102" "104.3"
-build_push "GRCm39"    "mus_musculus"           "104" "104.3"
-build_push "CanFam3.1" "canis_lupus_familiaris" "104" "104.3"
-build_push "WBcel235"  "caenorhabditis_elegans" "104" "104.3"
+build_push "GRCh37"    "homo_sapiens"           "105" "105.0"
+build_push "GRCh38"    "homo_sapiens"           "105" "105.0"
+build_push "GRCm38"    "mus_musculus"           "102" "105.0"
+build_push "GRCm39"    "mus_musculus"           "105" "105.0"
+build_push "CanFam3.1" "canis_lupus_familiaris" "104" "105.0"
+build_push "WBcel235"  "caenorhabditis_elegans" "105" "105.0"

modules/ensemblvep/environment.yml

@@ -1,10 +1,10 @@
 # You can use this file to create a conda environment for this module:
 #   conda env create -f environment.yml
-name: nf-core-vep-104.3
+name: nf-core-vep-105.0
 channels:
   - conda-forge
   - bioconda
   - defaults
 
 dependencies:
-  - bioconda::ensembl-vep=104.3
+  - bioconda::ensembl-vep=105.0

modules/gatk4/calibratedragstrmodel/main.nf

@@ -0,0 +1,50 @@
+process GATK4_CALIBRATEDRAGSTRMODEL {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::gatk4=4.2.6.1" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gatk4:4.2.6.1--hdfd78af_0':
+        'quay.io/biocontainers/gatk4:4.2.6.1--hdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(bam), path(bam_index), path(intervals)
+    path  fasta
+    path  fasta_fai
+    path  dict
+    path  strtablefile
+
+    output:
+    tuple val(meta), path("*.txt")   , emit: dragstr_model
+    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 intervals_command = intervals ? "--intervals $intervals" : ""
+
+    def avail_mem = 3
+    if (!task.memory) {
+        log.info '[GATK CalibrateDragstrModel] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = task.memory.giga
+    }
+    """
+    gatk --java-options "-Xmx${avail_mem}g" CalibrateDragstrModel \\
+        --input $bam \\
+        --output ${prefix}.txt \\
+        --reference $fasta \\
+        --str-table-path $strtablefile \\
+        $intervals_command \\
+        --tmp-dir . \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+}

modules/gatk4/calibratedragstrmodel/meta.yml

@@ -0,0 +1,74 @@
+name: gatk4_calibratedragstrmodel
+description: estimates the parameters for the DRAGstr model
+keywords:
+  - gatk4
+  - bam
+  - cram
+  - sam
+  - calibratedragstrmodel
+tools:
+  - gatk4:
+      description:
+        Genome Analysis Toolkit (GATK4). Developed in the Data Sciences Platform at the Broad Institute, the toolkit offers a wide variety of tools
+        with a primary focus on variant discovery and genotyping. Its powerful processing engine
+        and high-performance computing features make it capable of taking on projects of any size.
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us/articles/360057441571-CalibrateDragstrModel-BETA-
+      tool_dev_url: https://github.com/broadinstitute/gatk
+      doi: 10.1158/1538-7445.AM2017-3590
+      licence: ["Apache-2.0"]
+
+input:
+  # Only when we have meta
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bam:
+      type: file
+      description: BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+  - bam_index:
+      type: file
+      description: index of the BAM/CRAM/SAM file
+      pattern: "*.{bai,crai,sai}"
+  - intervals:
+      type: file
+      description: BED file or interval list containing regions (optional)
+      pattern: "*.{bed,interval_list}"
+  - fasta:
+      type: file
+      description: The reference FASTA file
+      pattern: "*.{fasta,fa}"
+  - fasta_fai:
+      type: file
+      description: The index of the reference FASTA file
+      pattern: "*.fai"
+  - dict:
+      type: file
+      description: The sequence dictionary of the reference FASTA file
+      pattern: "*.dict"
+  - strtablefile:
+      type: file
+      description: The StrTableFile zip folder of the reference FASTA file
+      pattern: "*.zip"
+
+output:
+  #Only when we have meta
+  - 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"
+  - dragstr_model:
+      type: file
+      description: The DragSTR model
+      pattern: "*.txt"
+
+authors:
+  - "@nvnieuwk"

modules/gatk4/haplotypecaller/main.nf

@@ -8,7 +8,7 @@ process GATK4_HAPLOTYPECALLER {
         'quay.io/biocontainers/gatk4:4.2.6.1--hdfd78af_0' }"
 
     input:
-    tuple val(meta), path(input), path(input_index), path(intervals)
+    tuple val(meta), path(input), path(input_index), path(intervals), path(dragstr_model)
     path  fasta
     path  fai
     path  dict
@@ -28,6 +28,7 @@ process GATK4_HAPLOTYPECALLER {
     def prefix = task.ext.prefix ?: "${meta.id}"
     def dbsnp_command = dbsnp ? "--dbsnp $dbsnp" : ""
     def interval_command = intervals ? "--intervals $intervals" : ""
+    def dragstr_command = dragstr_model ? "--dragstr-params-path $dragstr_model" : ""
 
     def avail_mem = 3
     if (!task.memory) {
@@ -42,6 +43,7 @@ process GATK4_HAPLOTYPECALLER {
         --reference $fasta \\
         $dbsnp_command \\
         $interval_command \\
+        $dragstr_command \\
         --tmp-dir . \\
         $args
 

modules/gatk4/haplotypecaller/meta.yml

@@ -32,6 +32,10 @@ input:
   - intervals:
       type: file
       description: Bed file with the genomic regions included in the library (optional)
+  - dragstr_model:
+      type: file
+      description: Text file containing the DragSTR model of the used BAM/CRAM file (optional)
+      pattern: "*.txt"
   - fasta:
       type: file
       description: The reference fasta file

tests/config/pytest_modules.yml

@@ -743,6 +743,10 @@ gatk4/calculatecontamination:
   - modules/gatk4/calculatecontamination/**
   - tests/modules/gatk4/calculatecontamination/**
 
+gatk4/calibratedragstrmodel:
+  - modules/gatk4/calibratedragstrmodel/**
+  - tests/modules/gatk4/calibratedragstrmodel/**
+
 gatk4/cnnscorevariants:
   - modules/gatk4/cnnscorevariants/**
   - tests/modules/gatk4/cnnscorevariants/**

tests/config/test_data.config

@@ -123,6 +123,7 @@ params {
                 genome_elfasta                                 = "${test_data_dir}/genomics/homo_sapiens/genome/genome.elfasta"
                 genome_fasta                                   = "${test_data_dir}/genomics/homo_sapiens/genome/genome.fasta"
                 genome_fasta_fai                               = "${test_data_dir}/genomics/homo_sapiens/genome/genome.fasta.fai"
+                genome_strtablefile                            = "${test_data_dir}/genomics/homo_sapiens/genome/genome_strtablefile.zip"
                 genome_dict                                    = "${test_data_dir}/genomics/homo_sapiens/genome/genome.dict"
                 genome_gff3                                    = "${test_data_dir}/genomics/homo_sapiens/genome/genome.gff3"
                 genome_gtf                                     = "${test_data_dir}/genomics/homo_sapiens/genome/genome.gtf"
@@ -265,6 +266,8 @@ params {
                 test_pileups_table                             = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/test.pileups.table"
                 test2_pileups_table                            = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/test2.pileups.table"
 
+                test_paired_end_sorted_dragstrmodel            = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/test_paired_end_sorted_dragstrmodel.txt"
+
                 test_genomicsdb_tar_gz                         = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/test_genomicsdb.tar.gz"
                 test_pon_genomicsdb_tar_gz                     = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/test_pon_genomicsdb.tar.gz"
 

tests/modules/gatk4/calibratedragstrmodel/main.nf

@@ -0,0 +1,66 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_CALIBRATEDRAGSTRMODEL } from '../../../../modules/gatk4/calibratedragstrmodel/main.nf'
+
+workflow test_gatk4_calibratedragstrmodel_bam {
+    
+    input = [
+        [ 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']['test_paired_end_sorted_bam_bai'], checkIfExists: true),
+        []
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    strtablefile = file(params.test_data['homo_sapiens']['genome']['genome_strtablefile'], checkIfExists: true)
+
+    GATK4_CALIBRATEDRAGSTRMODEL ( input, fasta, fasta_fai, dict, strtablefile )
+}
+
+workflow test_gatk4_calibratedragstrmodel_cram {
+    
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
+        []
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    strtablefile = file(params.test_data['homo_sapiens']['genome']['genome_strtablefile'], checkIfExists: true)
+
+    GATK4_CALIBRATEDRAGSTRMODEL ( input, fasta, fasta_fai, dict, strtablefile )
+}
+
+workflow test_gatk4_calibratedragstrmodel_beds {
+    
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['genome']['genome_bed'], checkIfExists: true)
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    strtablefile = file(params.test_data['homo_sapiens']['genome']['genome_strtablefile'], checkIfExists: true)
+
+    GATK4_CALIBRATEDRAGSTRMODEL ( input, fasta, fasta_fai, dict, strtablefile )
+}
+

tests/modules/gatk4/calibratedragstrmodel/nextflow.config

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

tests/modules/gatk4/calibratedragstrmodel/test.yml

@@ -0,0 +1,26 @@
+- name: gatk4 calibratedragstrmodel test_gatk4_calibratedragstrmodel_bam
+  command: nextflow run ./tests/modules/gatk4/calibratedragstrmodel -entry test_gatk4_calibratedragstrmodel_bam -c ./tests/config/nextflow.config  -c ./tests/modules/gatk4/calibratedragstrmodel/nextflow.config
+  tags:
+    - gatk4/calibratedragstrmodel
+    - gatk4
+  files:
+    - path: output/gatk4/test.txt
+      md5sum: 0a1a1583b157fa2251dd931ed165da4f
+
+- name: gatk4 calibratedragstrmodel test_gatk4_calibratedragstrmodel_cram
+  command: nextflow run ./tests/modules/gatk4/calibratedragstrmodel -entry test_gatk4_calibratedragstrmodel_cram -c ./tests/config/nextflow.config  -c ./tests/modules/gatk4/calibratedragstrmodel/nextflow.config
+  tags:
+    - gatk4/calibratedragstrmodel
+    - gatk4
+  files:
+    - path: output/gatk4/test.txt
+      md5sum: 1aa7ab38023f724877b3323c5e6b9a4e
+
+- name: gatk4 calibratedragstrmodel test_gatk4_calibratedragstrmodel_beds
+  command: nextflow run ./tests/modules/gatk4/calibratedragstrmodel -entry test_gatk4_calibratedragstrmodel_beds -c ./tests/config/nextflow.config  -c ./tests/modules/gatk4/calibratedragstrmodel/nextflow.config
+  tags:
+    - gatk4/calibratedragstrmodel
+    - gatk4
+  files:
+    - path: output/gatk4/test.txt
+      md5sum: def8baccad7bd59006f08fcb0a6721bf

tests/modules/gatk4/haplotypecaller/main.nf

@@ -8,6 +8,7 @@ workflow test_gatk4_haplotypecaller {
     input     = [ [ id:'test' ], // 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)
@@ -21,6 +22,7 @@ workflow test_gatk4_haplotypecaller_cram {
     input = [ [ id:'test' ], // meta map
                 file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
                 file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
+                [],
                 []
               ]
     fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
@@ -34,7 +36,8 @@ workflow test_gatk4_haplotypecaller_intervals_dbsnp {
    input = [ [ id:'test' ], // meta map
                 file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
                 file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
-                file(params.test_data['homo_sapiens']['genome']['genome_bed'], checkIfExists: true)
+                file(params.test_data['homo_sapiens']['genome']['genome_bed'], checkIfExists: true),
+                []
             ]
 
     fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
@@ -45,3 +48,20 @@ workflow test_gatk4_haplotypecaller_intervals_dbsnp {
 
     GATK4_HAPLOTYPECALLER ( input, fasta, fai, dict, sites, sites_tbi )
 }
+
+workflow test_gatk4_haplotypecaller_dragstr_model {
+   input = [ [ id:'test' ], // meta map
+                file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
+                file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
+                [],
+                file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_dragstrmodel'], checkIfExists: true)
+            ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+    sites = []
+    sites_tbi = []
+
+    GATK4_HAPLOTYPECALLER ( input, fasta, fai, dict, sites, sites_tbi )
+}

tests/modules/gatk4/haplotypecaller/test.yml

@@ -6,7 +6,6 @@
   files:
     - path: output/gatk4/test.vcf.gz
     - path: output/gatk4/test.vcf.gz.tbi
-    - path: output/gatk4/versions.yml
 
 - name: gatk4 haplotypecaller test_gatk4_haplotypecaller_cram
   command: nextflow run ./tests/modules/gatk4/haplotypecaller -entry test_gatk4_haplotypecaller_cram -c ./tests/config/nextflow.config -c ./tests/modules/gatk4/haplotypecaller/nextflow.config
@@ -16,7 +15,6 @@
   files:
     - path: output/gatk4/test.vcf.gz
     - path: output/gatk4/test.vcf.gz.tbi
-    - path: output/gatk4/versions.yml
 
 - name: gatk4 haplotypecaller test_gatk4_haplotypecaller_intervals_dbsnp
   command: nextflow run ./tests/modules/gatk4/haplotypecaller -entry test_gatk4_haplotypecaller_intervals_dbsnp -c ./tests/config/nextflow.config -c ./tests/modules/gatk4/haplotypecaller/nextflow.config
@@ -26,4 +24,12 @@
   files:
     - path: output/gatk4/test.vcf.gz
     - path: output/gatk4/test.vcf.gz.tbi
-    - path: output/gatk4/versions.yml
+
+- name: gatk4 haplotypecaller test_gatk4_haplotypecaller_dragstr_model
+  command: nextflow run ./tests/modules/gatk4/haplotypecaller -entry test_gatk4_haplotypecaller_dragstr_model -c ./tests/config/nextflow.config -c ./tests/modules/gatk4/haplotypecaller/nextflow.config
+  tags:
+    - gatk4/haplotypecaller
+    - gatk4
+  files:
+    - path: output/gatk4/test.vcf.gz
+    - path: output/gatk4/test.vcf.gz.tbi