Implemented TMB score calculation

assets/tmb/mutect2.yml

@@ -0,0 +1,10 @@
+## Describe the fields
+## For definition, provide the expected key:values
+## Note that several keys/values can be defined
+##
+###############################################
+
+freq: "AF"
+depth: "DP"
+altDepth: "AD"
+maxVaf: "1"

assets/tmb/snpeff.yml

@@ -0,0 +1,119 @@
+## Describe the fields
+## For definition, provide the expected key:values
+## Note that several keys/values can be defined
+
+###############################################
+## SnpEff Parsing
+
+## Tags
+tag: "ANN"
+sep: "&"
+
+## Annotation flags
+
+isCoding:
+  1:
+    - chromosome_number_variation
+    - coding_sequence_variant
+    - conservative_inframe_deletion
+    - conservative_inframe_insertion
+    - disruptive_inframe_deletion
+    - disruptive_inframe_insertion
+    - exon_loss
+    - exon_loss_variant
+    - exon_variant
+    - frameshift_variant
+    - gene_variant
+    - initiator_codon_variant
+    - missense_variant
+    - rare_amino_acid_variant
+    - splice_acceptor_variant
+    - splice_donor_variant
+    - splice_region_variant
+    - start_lost
+    - start_retained
+    - stop_gained
+    - stop_lost
+    - stop_retained_variant
+    - synonymous_variant
+    - transcript_ablation
+    - transcript_amplification
+    - transcript_variant
+
+isNonCoding:
+  1:
+    - 3_prime_UTR_truncation
+    - 3_prime_UTR_variant
+    - 5_prime_UTR_premature_start_codon_gain_variant
+    - 5_prime_UTR_truncation
+    - 5_prime_UTR_variant
+    - conserved_intergenic_variant
+    - conserved_intron_variant
+    - downstream_gene_variant
+    - feature_elongation
+    - feature_truncation
+    - intergenic_region
+    - intragenic_variant
+    - intron_variant
+    - mature_miRNA_variant
+    - miRNA
+    - NMD_transcript_variant
+    - non_coding_transcript_exon_variant
+    - non_coding_transcript_variant
+    - regulatory_region_ablation
+    - regulatory_region_amplification
+    - regulatory_region_variant
+    - TF_binding_site_variant
+    - TFBS_ablation
+    - TFBS_amplification
+    - upstream_gene_variant
+
+isSplicing:
+  1:
+    - splice_donor_variant
+    - splice_acceptor_variant
+    - splice_region_variant
+
+isSynonymous:
+  1:
+    - start_retained_variant
+    - stop_retained_variant
+    - synonymous_variant
+
+isNonSynonymous:
+  1:
+    - frameshift_variant
+    - missense_variant
+    - rare_amino_acid_variant
+    - splice_acceptor_variant
+    - splice_donor_variant
+    - splice_region_variant
+    - start_lost
+    - stop_gained
+    - stop_lost
+
+## Databases
+cancerDb:
+  cosmic:
+    - cosmic_coding_ID
+    - cosmic_noncoding_ID
+
+polymDb:
+  1k:
+    - kg_AMR_AF
+    - kg_AFR_AF
+    - kg_EAS_AF
+    - kg_EUR_AF
+    - kg_SAS_AF
+    - KG_AF_GLOBAL
+
+  gnomad:
+    - GnomAD
+    - gnomAD_genomes_AF
+    - AF
+
+  esp:
+    - ESP_AF_GLOBAL
+
+  dbsnp:
+    - dbSNPBuildID

assets/tmb/strelka.yml

@@ -0,0 +1,10 @@
+## Describe the fields
+## For definition, provide the expected key:values
+## Note that several keys/values can be defined
+##
+###############################################
+
+freq: "AF"
+depth: "DP"
+altDepth: "AD"
+maxVaf: "1"

conf/modules/msisensorpro.config

@@ -22,4 +22,14 @@ process {
             saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
         ]
     }
+
+    withName: 'MSISENSORPRO_MSITUMORONLY' {
+        ext.args   = { params.wes ? '-c 20' : '-c 15' } // default values by MSIsensorpro
+        ext.prefix = { "${meta.id}.tumor_only" }
+        publishDir = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/msisensorpro/${meta.id}/" },
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
 }

conf/modules/tmb.config

@@ -0,0 +1,52 @@
+/*
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Config file for defining DSL2 per module options and publishing paths
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Available keys to override module options:
+        ext.args   = Additional arguments appended to command in module.
+        ext.args2  = Second set of arguments appended to command in module (multi-tool modules).
+        ext.args3  = Third set of arguments appended to command in module (multi-tool modules).
+        ext.prefix = File name prefix for output files.
+        ext.when   = When to run the module.
+----------------------------------------------------------------------------------------
+*/
+
+// TMB
+
+process {
+
+    withName: 'BCFTOOLS_NORM' {
+        ext.args = { "-m- -c w" }
+        ext.prefix = { vcf.baseName - ".vcf" + ".norm" }
+        ext.when = { meta.dbconfig && meta.varconfig }
+        publishDir = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/tmb/${meta.id}/" },
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: 'TMB' {
+        ext.args = { [
+            "--sample ${meta.patient}_${meta.sample}",
+            params.tmb_effgenomesize ? "--effGenomeSize ${params.tmb_effgenomesize}" : '',
+            '--vaf 0.05',
+            '--maf 1',
+            '--minDepth 100',
+            '--minAltDepth 2',
+            '--filterLowQual',
+            '--filterNonCoding',
+            '--filterSyn',
+            '--filterPolym',
+            '--polymDb 1k,gnomad,dbsnp'
+            ].join(' ').trim()
+        }
+        ext.prefix = { vcf.baseName }
+        ext.when = { dbconfig && varconfig }
+        publishDir = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/tmb/${meta.id}/" },
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+}

main.nf

@@ -77,6 +77,7 @@ include { PIPELINE_INITIALISATION          } from './subworkflows/local/utils_nf
 include { PREPARE_GENOME                   } from './subworkflows/local/prepare_genome'
 include { PREPARE_INTERVALS                } from './subworkflows/local/prepare_intervals'
 include { PREPARE_REFERENCE_CNVKIT         } from './subworkflows/local/prepare_reference_cnvkit'
+include { MSISENSORPRO_SCAN                } from './modules/nf-core/msisensorpro/scan/main'
 
 // Initialize fasta file with meta map:
 fasta = params.fasta ? Channel.fromPath(params.fasta).map{ it -> [ [id:it.baseName], it ] }.collect() : Channel.empty()
@@ -91,6 +92,7 @@ germline_resource       = params.germline_resource       ? Channel.fromPath(para
 known_indels            = params.known_indels            ? Channel.fromPath(params.known_indels).collect()              : Channel.value([])
 known_snps              = params.known_snps              ? Channel.fromPath(params.known_snps).collect()                : Channel.value([])
 mappability             = params.mappability             ? Channel.fromPath(params.mappability).collect()               : Channel.value([])
+msisensorpro_baseline   = params.msisensorpro_baseline   ? Channel.fromPath(params.msisensorpro_baseline).collect()     : Channel.empty()
 pon                     = params.pon                     ? Channel.fromPath(params.pon).collect()                       : Channel.value([]) // PON is optional for Mutect2 (but highly recommended)
 sentieon_dnascope_model = params.sentieon_dnascope_model ? Channel.fromPath(params.sentieon_dnascope_model).collect()   : Channel.value([])
 
@@ -166,8 +168,18 @@ workflow NFCORE_SAREK {
         aligner == "bwa-mem2" ? bwamem2 :
         dragmap
 
-    // TODO: add a params for msisensorpro_scan
-    msisensorpro_scan      = PREPARE_GENOME.out.msisensorpro_scan
+    // Reference msi list for MSIsensorpro
+    if (params.tools && params.tools.split(',').contains('msisensorpro')) {
+        if (params.msisensorpro_scan) {
+            msisensorpro_scan = Channel.fromPath(params.msisensorpro_scan).collect()
+        } else {
+            MSISENSORPRO_SCAN(fasta)
+            msisensorpro_scan = MSISENSORPRO_SCAN.out.list.map{ meta, list -> [list] }
+            versions = versions.mix(MSISENSORPRO_SCAN.out.versions)
+        }
+    } else {
+        msisensorpro_scan = Channel.empty()
+    }
 
     // For ASCAT, extracted from zip or tar.gz files
     allele_files           = PREPARE_GENOME.out.allele_files
@@ -177,7 +189,7 @@ workflow NFCORE_SAREK {
     rt_file                = PREPARE_GENOME.out.rt_file
 
     // Tabix indexed vcf files
-    bcftools_annotations_tbi  = params.bcftools_annotations    ? params.bcftools_annotations_tbi ? Channel.fromPath(params.bcftools_annotations_tbi).collect() : PREPARE_GENOME.out.bcftools_annotations_tbi : Channel.empty([])
+    bcftools_annotations_tbi  = params.bcftools_annotations    ? params.bcftools_annotations_tbi ? Channel.fromPath(params.bcftools_annotations_tbi).collect() : PREPARE_GENOME.out.bcftools_annotations_tbi : Channel.value([])
     dbsnp_tbi                 = params.dbsnp                   ? params.dbsnp_tbi                ? Channel.fromPath(params.dbsnp_tbi).collect()                : PREPARE_GENOME.out.dbsnp_tbi                : Channel.value([])
     germline_resource_tbi     = params.germline_resource       ? params.germline_resource_tbi    ? Channel.fromPath(params.germline_resource_tbi).collect()    : PREPARE_GENOME.out.germline_resource_tbi    : [] //do not change to Channel.value([]), the check for its existence then fails for Getpileupsumamries
     known_indels_tbi          = params.known_indels            ? params.known_indels_tbi         ? Channel.fromPath(params.known_indels_tbi).collect()         : PREPARE_GENOME.out.known_indels_tbi         : Channel.value([])
@@ -296,6 +308,7 @@ workflow NFCORE_SAREK {
         loci_files,
         mappability,
         msisensorpro_scan,
+        msisensorpro_baseline,
         ngscheckmate_bed,
         pon,
         pon_tbi,

modules.json

@@ -26,6 +26,11 @@
                         "git_sha": "a5ba4d59c2b248c0379b0f8aeb4e7e754566cd1f",
                         "installed_by": ["bam_ngscheckmate"]
                     },
+                    "bcftools/norm": {
+                        "branch": "master",
+                        "git_sha": "33ef773a7ea36e88323902f63662aa53c9b88988",
+                        "installed_by": ["modules"]
+                    },
                     "bcftools/sort": {
                         "branch": "master",
                         "git_sha": "a5ba4d59c2b248c0379b0f8aeb4e7e754566cd1f",

modules/local/tmb/environment.yml

@@ -0,0 +1,7 @@
+name: tmb
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - bioconda::tmb=1.3.0

modules/local/tmb/main.nf

@@ -0,0 +1,41 @@
+process TMB {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/tmb:1.3.0--pyh5e36f6f_0':
+        'quay.io/biocontainers/tmb:1.3.0--pyh5e36f6f_0' }"
+
+    input:
+    tuple val(meta), path(vcf), path(dbconfig), path(varconfig)
+    path (intervals)
+
+    output:
+    tuple val(meta), path("*.log"), emit: log
+    path("*_export.vcf")          , emit: vcf
+    path "versions.yml"           , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def intervals_bed = intervals ? "--bed ${intervals}" : ''
+    def target_region = args.contains("--effGenomeSize") ? '' : intervals_bed
+    """
+    pyTMB.py -i ${vcf} \\
+        --dbConfig ${dbconfig} \\
+        --varConfig ${varconfig} \\
+        ${target_region} \\
+        ${args} \\
+        --export \\
+        > ${prefix}.tmb.log
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        tmb: \$(echo \$(pyTMB.py --version 2>&1) | sed 's/^.*pyTMB.py //; s/.*\$//' | sed 's|[()]||g')
+    END_VERSIONS
+    """
+}