Add sentieon wgsmetrics (#3727)

* wgsmetrics

* fix lint error

* typo

* Update modules/nf-core/sentieon/wgsmetrics/main.nf

Co-authored-by: Anders Sune Pedersen <37172585+asp8200@users.noreply.github.com>

---------

Co-authored-by: Anders Sune Pedersen <37172585+asp8200@users.noreply.github.com>

.github/workflows/test.yml

@@ -234,6 +234,8 @@ jobs:
  tags: sentieon/readwriter
  - profile: "conda"
  tags: sentieon/varcal
+ - profile: "conda"
+ tags: sentieon/wgsmetrics
  - profile: "conda"
  tags: spaceranger/count
  - profile: "conda"

modules/nf-core/sentieon/wgsmetrics/main.nf

@@ -0,0 +1,81 @@
+process SENTIEON_WGSMETRICS {
+ tag "$meta.id"
+ label 'process_medium'
+ label 'sentieon'
+
+ secret 'SENTIEON_LICENSE_BASE64'
+
+ container 'nf-core/sentieon:202112.06'
+
+ input:
+ tuple val(meta), path(bam), path(bai)
+ tuple val(meta2), path(fasta)
+ tuple val(meta3), path(fai)
+ tuple val(meta4), path(intervals_list)
+
+
+ output:
+ tuple val(meta), path('*.txt'), emit: wgs_metrics
+ path "versions.yml" , emit: versions
+
+ when:
+ task.ext.when == null || task.ext.when
+
+ script:
+ // Exit if running this module with -profile conda / -profile mamba
+ if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+ error "Sentieon modules do not support Conda. Please use Docker / Singularity / Podman instead."
+ }
+ def args = task.ext.args ?: ''
+ def input = bam.sort().collect{"-i $it"}.join(' ')
+ def prefix = task.ext.prefix ?: "${meta.id}"
+ def interval = intervals_list ? "--interval ${intervals_list}" : ""
+ def sentieon_auth_mech_base64 = task.ext.sentieon_auth_mech_base64 ?: ''
+ def sentieon_auth_data_base64 = task.ext.sentieon_auth_data_base64 ?: ''
+ """
+ if [ "\${#SENTIEON_LICENSE_BASE64}" -lt "1500" ]; then # If the string SENTIEON_LICENSE_BASE64 is short, then it is an encrypted url.
+ export SENTIEON_LICENSE=\$(echo -e "\$SENTIEON_LICENSE_BASE64" | base64 -d)
+ else # Localhost license file
+ # The license file is stored as a nextflow variable like, for instance, this:
+ # nextflow secrets set SENTIEON_LICENSE_BASE64 \$(cat <sentieon_license_file.lic> | base64 -w 0)
+ export SENTIEON_LICENSE=\$(mktemp)
+ echo -e "\$SENTIEON_LICENSE_BASE64" | base64 -d > \$SENTIEON_LICENSE
+ fi
+
+ if [ ${sentieon_auth_mech_base64} ] && [ ${sentieon_auth_data_base64} ]; then
+ # If sentieon_auth_mech_base64 and sentieon_auth_data_base64 are non-empty strings, then Sentieon is mostly likely being run with some test-license.
+ export SENTIEON_AUTH_MECH=\$(echo -n "${sentieon_auth_mech_base64}" | base64 -d)
+ export SENTIEON_AUTH_DATA=\$(echo -n "${sentieon_auth_data_base64}" | base64 -d)
+ echo "Decoded and exported Sentieon test-license system environment variables"
+ fi
+
+ sentieon \\
+ driver \\
+ -t $task.cpus \\
+ -r $fasta \\
+ $input \\
+ $interval \\
+ $args \\
+ --algo WgsMetricsAlgo ${prefix}.txt
+
+ cat <<-END_VERSIONS > versions.yml
+ "${task.process}":
+ sentieon: \$(echo \$(sentieon driver --version 2>&1) | sed -e "s/sentieon-genomics-//g")
+ END_VERSIONS
+ """
+
+ stub:
+ // Exit if running this module with -profile conda / -profile mamba
+ if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+ error "Sentieon modules do not support Conda. Please use Docker / Singularity / Podman instead."
+ }
+ def prefix = task.ext.prefix ?: "${meta.id}"
+ """
+ touch ${prefix}.txt
+
+ cat <<-END_VERSIONS > versions.yml
+ "${task.process}":
+ sentieon: \$(echo \$(sentieon driver --version 2>&1) | sed -e "s/sentieon-genomics-//g")
+ END_VERSIONS
+ """
+}

modules/nf-core/sentieon/wgsmetrics/meta.yml

@@ -0,0 +1,73 @@
+name: "sentieon_wgsmetrics"
+description: Collects whole genome quality metrics from a bam file
+keywords:
+ - metrics
+ - bam
+ - sentieon
+tools:
+ - sentieon:
+ description: |
+ Sentieon® provides complete solutions for secondary DNA/RNA analysis for a variety of sequencing platforms, including short and long reads.
+ Our software improves upon BWA, STAR, Minimap2, GATK, HaplotypeCaller, Mutect, and Mutect2 based pipelines and is deployable on any generic-CPU-based computing system.
+ homepage: https://www.sentieon.com/
+ documentation: https://www.sentieon.com/
+
+input:
+ - meta:
+ type: map
+ description: |
+ Groovy Map containing sample information
+ e.g. `[ id:'test', single_end:false ]`
+ - meta2:
+ type: map
+ description: |
+ Groovy Map containing reference information
+ e.g. `[ id:'test' ]`
+ - meta3:
+ type: map
+ description: |
+ Groovy Map containing reference information
+ e.g. `[ id:'test' ]`
+ - meta4:
+ type: map
+ description: |
+ Groovy Map containing reference information
+ e.g. `[ id:'test' ]`
+ - bam:
+ type: file
+ description: Sorted BAM/CRAM/SAM file
+ pattern: "*.{bam,cram,sam}"
+ - bai:
+ type: file
+ description: Index of th sorted BAM/CRAM/SAM file
+ pattern: "*.{bai,crai,sai}"
+ - fasta:
+ type: file
+ description: Genome fasta file
+ pattern: "*.{fa,fasta}"
+ - fai:
+ type: file
+ description: Index of the genome fasta file
+ pattern: "*.fai"
+ - interval_list:
+ type: file
+ description: bed or interval_list file containing interval in the reference that will be used in the analysis
+ pattern: "*.{bed,interval_list}"
+
+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"
+ - wgs_metrics:
+ type: file
+ description: File containing the information about mean base quality score for each sequencing cycle
+ pattern: "*.txt"
+
+authors:
+ - "@ramprasadn"

tests/config/pytest_modules.yml

@@ -3299,6 +3299,10 @@ sentieon/readwriter:
  - modules/nf-core/sentieon/readwriter/**
  - tests/modules/nf-core/sentieon/readwriter/**
 
+sentieon/wgsmetrics:
+ - modules/nf-core/sentieon/wgsmetrics/**
+ - tests/modules/nf-core/sentieon/wgsmetrics/**
+
 sentieon/varcal:
  - modules/nf-core/sentieon/varcal/**
  - tests/modules/nf-core/sentieon/varcal/**

tests/modules/nf-core/sentieon/wgsmetrics/main.nf

@@ -0,0 +1,24 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SENTIEON_WGSMETRICS } from '../../../../../modules/nf-core/sentieon/wgsmetrics/main.nf'
+
+workflow test_sentieon_wgsmetrics {
+
+ input = [
+ [ id:'test', single_end:false ], // 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 = [
+ [id:'genome'],
+ file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+ ]
+ fai = [
+ [id:'genome'],
+ file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true)
+ ]
+
+ SENTIEON_WGSMETRICS ( input, fasta, fai, [[:],[]] )
+}

tests/modules/nf-core/sentieon/wgsmetrics/nextflow.config

@@ -0,0 +1,10 @@
+process {
+
+ publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+ withLabel: 'sentieon' {
+ ext.sentieon_auth_mech_base64 = secrets.SENTIEON_AUTH_MECH_BASE64
+ ext.sentieon_auth_data_base64 = secrets.SENTIEON_AUTH_DATA_BASE64
+ }
+
+}

tests/modules/nf-core/sentieon/wgsmetrics/test.yml

@@ -0,0 +1,8 @@
+- name: "sentieon wgsmetrics"
+ command: nextflow run ./tests/modules/nf-core/sentieon/wgsmetrics -entry test_sentieon_wgsmetrics -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/sentieon/wgsmetrics/nextflow.config
+ tags:
+ - "sentieon"
+ - "sentieon/wgsmetrics"
+ files:
+ - path: "output/sentieon/test.txt"
+ - path: "output/sentieon/versions.yml"