Qc on speed: FastP and Mosdepth as drop-in replacement for Trimgalore/Seqkit & Deeptools/qualimap

.github/workflows/ci.yml

@@ -55,6 +55,7 @@ jobs:
           - "strelkabp"
           - "targeted"
           - "tiddit"
+          - "trimming"
           - "tumor_normal_pair"
           - "variantcalling_channel"
 

assets/multiqc_config.yml

@@ -18,10 +18,10 @@ export_plots: true
 run_modules:
   - custom_content
   - fastqc
-  - cutadapt
+  - fastp
   - picard
   - samtools
-  - qualimap
+  - mosdepth
   - gatk
   - bcftools
   - vcftools
@@ -33,19 +33,15 @@ module_order:
       name: "FastQC (raw)"
       path_filters_exclude:
         - "*_val_*.zip"
-  - cutadapt:
-      name: "Cutadapt"
-  - fastqc:
-      name: "FastQC (trimmed)"
-      path_filters:
-        - "*_val_*.zip"
+  - fastp:
+      name: "FastP (Read preprocessing)"
   - picard:
       name: "GATK4 MarkDuplicates"
       info: " metrics generated either by GATK4 MarkDuplicates or EstimateLibraryComplexity (with --use_gatk_spark)."
   - samtools:
       name: "Samtools Flagstat"
-  - qualimap:
-      name: "Qualimap"
+  - mosdepth:
+      name: "Mosdepth"
   - gatk:
       name: "GATK4 BQSR"
   - bcftools:

conf/modules.config

@@ -193,23 +193,29 @@ process {
     }
 
 // TRIMMING
-    withName: 'TRIMGALORE' {
-        ext.args         = '--fastqc'
-        publishDir       = [
-            [
-                path: { "${params.outdir}/reports/trimgalore/${meta.id}/fastqc" },
-                mode: params.publish_dir_mode,
-                pattern: "*{html,zip}"
-            ],
+
+    withName: FASTP {
+        ext.args = [ "",
+            params.trim_fastq              ?: "--disable_adapter_trimming",
+            params.clip_r1 > 0             ? "--trim_front1 ${params.clip_r1}" : "", // Remove bp from the 5' end of read 1.
+            params.three_prime_clip_r1 > 0 ? "--trim_tail1 ${params.three_prime_clip_r1}" : "", // Remove bp from the 3' end of read 1 AFTER adapter/quality trimming has been performed.
+            params.clip_r2   > 0           ? "--trim_front2 ${params.clip_r2}" : "", //Remove bp from the 5' end of read 2.
+            params.three_prime_clip_r2 > 0 ? "--trim_tail2 ${params.three_prime_clip_r2}" : "", // Remove bp from the 3' end of read 2 AFTER adapter/quality trimming has been performed.
+            params.trim_nextseq            ? "--trim_poly_g" : "", // Apply the --nextseq=X option, to trim based on quality after removing poly-G tails.
+            params.split_fastq > 0         ? "--split_by_lines ${params.split_fastq * 4}" : "",
+
+        ].join(" ").trim()
+        publishDir = [
             [
-                path: { "${params.outdir}/trimgalore/${meta.id}" },
+                path: { "${params.outdir}/reports/fastp/${meta.sample}" },
                 mode: params.publish_dir_mode,
-                pattern: "*fq.gz"
+                pattern: "*.{html,json,log}"
             ],
             [
-                path: { "${params.outdir}/reports/trimgalore/${meta.id}" },
+                enabled: params.save_trimmed || params.save_split_fastqs,
+                path: { "${params.outdir}/preprocessing/${meta.sample}/fastp" },
                 mode: params.publish_dir_mode,
-                pattern: "*txt"
+                pattern: "*.fastp.fastq.gz"
             ]
         ]
     }
@@ -239,19 +245,6 @@ process {
         ext.prefix       = {"${meta.id}_unsorted_tagged"}
     }
 
-// SPLIT FASTQ
-
-    withName: 'SEQKIT_SPLIT2' {
-        ext.args         = { "--by-size ${params.split_fastq}" }
-        ext.when         = { params.split_fastq > 1 }
-        publishDir       = [
-            enabled: params.save_split_fastqs,
-            mode: params.publish_dir_mode,
-            path: { "${params.outdir}/preprocessing/${meta.sample}/seqkit" },
-            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
-        ]
-    }
-
 // MAPPING
     if (params.step == 'mapping') {
         withName: ".*GATK4_MAPPING:BWAMEM1_MEM" {
@@ -272,12 +265,12 @@ process {
             // However if it's skipped, reads need to be coordinate-sorted
             // Only name sort if Spark for Markduplicates + duplicate marking is not skipped
             ext.args2        = { params.use_gatk_spark && params.use_gatk_spark.contains('markduplicates') && (!params.skip_tools || (params.skip_tools && !params.skip_tools.contains('markduplicates'))) ? '-n' : '' }
-            ext.prefix       = { params.split_fastq > 1 ? "${meta.id}".concat('.').concat(reads.get(0).name.findAll(/part_([0-9]+)?/).last()) : "" }
+            ext.prefix       = { params.split_fastq > 1 ? "${meta.id}".concat('.').concat(reads.get(0).name.tokenize('.')[0]) : "" }
             publishDir       = [
                 mode: params.publish_dir_mode,
-                path: { "${params.outdir}/preprocessing/${meta.id}/" },
+                path: { "${params.outdir}/preprocessing/" },
                 pattern: "*bam",
-                saveAs: { (params.save_bam_mapped || (params.skip_tools && params.skip_tools.contains('markduplicates'))) && (meta.size * meta.numLanes == 1) ? "mapped/${it}" : null }
+                saveAs: { (params.save_bam_mapped || (params.skip_tools && params.skip_tools.contains('markduplicates'))) && (meta.size * meta.numLanes == 1) ? "${meta.id}/mapped/${it}" : null }
             ]
         }
 
@@ -479,28 +472,18 @@ process {
         ]
     }
 
-    withName: 'DEEPTOOLS_BAMCOVERAGE' {
-        ext.when         = { !(params.skip_tools && params.skip_tools.contains('deeptools')) }
-        publishDir       = [
-            mode: params.publish_dir_mode,
-            path: { "${params.outdir}/reports/deeptools/${meta.id}" },
-            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
-        ]
-    }
-
-    withName: 'QUALIMAP_BAMQCCRAM' {
-        ext.args         = '--paint-chromosome-limits --genome-gc-distr HUMAN -skip-duplicated --skip-dup-mode 0 -outformat HTML'
+    withName: 'MOSDEPTH' {
+        ext.args = { !params.wes ? "-n --fast-mode --by 500" : ""}
         ext.prefix       = { "${meta.id}.mapped" }
-        ext.when         = { !(params.skip_tools && params.skip_tools.contains('bamqc')) }
-        publishDir       = [
+        publishDir = [
             mode: params.publish_dir_mode,
-            path: { "${params.outdir}/reports/qualimap/${meta.id}" },
-            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+            path: { "${params.outdir}/reports/mosdepth/${meta.id}" },
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
         ]
     }
 
     if (params.step == 'mapping' || params.step == 'markduplicates'|| params.step == 'prepare_recalibration'|| params.step == 'recalibrate') {
-        withName: 'NFCORE_SAREK:SAREK:CRAM_QC:QUALIMAP_BAMQCCRAM' {
+        withName: 'NFCORE_SAREK:SAREK:CRAM_QC:MOSDEPTH' {
             ext.prefix       = { "${meta.id}.recal" }
         }
 
@@ -1131,7 +1114,7 @@ process{
     // SNPEFF
     if (params.tools && (params.tools.contains('snpeff') || params.tools.contains('merge'))) {
         withName: 'SNPEFF' {
-            ext.prefix       = { "${vcf.baseName.minus(".vcf")}_snpEff" }
+            ext.prefix       = {  meta.variantcaller ? "${meta.variantcaller}_${vcf.baseName.minus(".vcf")}_snpEff" : "${vcf.baseName.minus(".vcf")}_snpEff" }
             ext.args          = '-nodownload -canon -v'
             if (!params.snpeff_cache) container = { params.snpeff_genome ? "nfcore/snpeff:5.0.${params.snpeff_genome}" : "nfcore/snpeff:5.0.${params.genome}" }
             publishDir       = [
@@ -1146,7 +1129,7 @@ process{
     // VEP
     if (params.tools && (params.tools.contains('vep') || params.tools.contains('merge'))) {
         withName: 'ENSEMBLVEP' {
-            ext.prefix       = { "${vcf.baseName.minus(".vcf")}_VEP" }
+            ext.prefix       = { meta.variantcaller ? "${meta.variantcaller}_${vcf.baseName.minus(".vcf")}_VEP" : ${vcf.baseName.minus(".vcf")}_VEP }
             ext.args          = [
                 '--everything --filter_common --per_gene --total_length --offline --format vcf',
                 (params.vep_dbnsfp && params.dbnsfp)                                  ? '--plugin dbNSFP,dbNSFP.gz,rs_dbSNP,HGVSc_VEP,HGVSp_VEP,1000Gp3_EAS_AF,1000Gp3_AMR_AF,LRT_score,GERP++_RS,gnomAD_exomes_AF' : '',
@@ -1173,15 +1156,15 @@ process{
 
     if (params.tools && params.tools.contains('vep')) {
         withName: 'NFCORE_SAREK:SAREK:ANNOTATE:ANNOTATION_ENSEMBLVEP:ENSEMBLVEP' {
-            ext.prefix      = {"${meta.id}_VEP"}
+            ext.prefix      = { meta.variantcaller ? "${meta.variantcaller}_${meta.id}_VEP" : "${meta.id}_VEP" }
         }
     }
 
     // SNPEFF THEN VEP
     if (params.tools && params.tools.contains('merge')) {
         withName: ".*:ANNOTATION_MERGE:ENSEMBLVEP" {
             // Output file will have format *_snpEff_VEP.ann.vcf, *_snpEff_VEP.ann.json or *_snpEff_VEP.ann.tab
-            ext.prefix       = { "${vcf.baseName.minus(".ann.vcf")}_VEP" }
+            ext.prefix       = { meta.variantcaller ? "${meta.variantcaller}_${vcf.baseName.minus(".ann.vcf")}_VEP" : "${vcf.baseName.minus(".ann.vcf")}_VEP" }
         }
     }